Enterprise AI and Digital Sovereignty, with Tom Jenkins and Major-General (Retd) David Fraser (TRN5-V75)

Transcript: Enterprise AI and Digital Sovereignty, with Tom Jenkins and Major-General (Retd) David Fraser

[00:00:01 Text appears on screen: "Welcome. Bienvenue."]

[00:00:07 The screen fades to Taki Sarantakis standing at a podium.]

Taki Sarantakis (Canada School of Public Service): Welcome. My name is Taki Sarantakis. We have yet another treat for you today at the Canada School of Public Service. Today, we have something that everybody is talking about right now, and no, it's not Taylor Swift and it's not… who's the other guy? Who just did the Super Bowl?

Crowd: Bad Bunny.

Taki Sarantakis: It's not the angry rabbit. It is A.I., and A.I., as you all know, is something that not only is kind of dominating the discourse right now, it is also going to dominate a lot of other things. So, we have two wonderful speakers. I will introduce them very, very briefly. The first is Tom Jenkins. Tom is the Chair of OpenText, and OpenText is a wonderful Canadian success story. Tom is one of the very, very few individuals in Canada who has taken something and scaled it to a global level and then kept it there for over 30 years. It is very, very, very rare for anybody to play on the global scale for three decades, and it's even more rare for that to be Canadian. He is also one of the… it's going to sound stupid but he's actually one of the inventors of the internet. So, you can look that up. His company is the reflection of Google. When you go search something, you're searching the internet. Tom, a long, long time ago, realized that you also needed to search within your company. So, Tom is like the Google of the intranet. The second is Retired Major-General David Fraser.

[00:02:03 Taki Sarantaki holds up a copy of the book Operation Medusa by David Fraser.]

Operation Medusa, if you know, you know, if you don't, Google it, and there's also a copy for those joining us in person if you want to go forward.

So, this is the second time we are having these gentlemen here.

[00:02:20 Taki Sarantakis holds up a copy of the book The Anticipant Organization by David Fraser and Tom Jenkins.]

The first time, they came and talked to us about The Anticipant Organization. And by that, what they meant is machines are now so quick and so fast that if you're reacting, you've already lost. You need to be an organization that anticipates things rather than reacts to things.

[00:02:45 Taki Sarantakis holds up a copy of the book Enterprise Artificial Intelligence by Shannon Bell, David Fraser, and Tom Jenkins.]

And today, we're going to be talking about their latest book which is A.I. enterprise and cloud. And one thing before I turn it over to them to start their presentation, if you use Chat GPT and you use Anthropic and you use OpenAI and you don't have a subscription, you've never used Enterprise A.I. So, just keep that in mind too.

Tom, over to you guys.

[00:03:15 Taki Sarantaki walks offstage and Tom Jenkins walks up to the podium.]

Tom Jenkins (Chair of the Board, OpenText): I'll start, and then my partner in crime here will jump in. A pleasure to see so many familiar faces. Nice to see you all. We're here to talk about something that many of you are curious about, and we're going to try our best to take you down inside how this stuff works but to keep it at a high level. As my partner in crime here, General Fraser, he says, look, Tom, if you can get an army general to understand it, anybody can understand it. So, that is his humble way of saying that we've tried to make it for non-technical people. But of course, when General Fraser talks to you, you'll see what unbelievable implications it has because he speaks as the user to all of this. So, let me get the slides and we'll get right into it.

[00:04:13 A slide is shown with the title: "Today's Topic: How to Build an Agentic Bot" and the cover of the book Enterprise Artificial Intelligence next to the text:
"Chapter One – The Evolution of Enterprise Data"
"Chapter Two – The Rise of Enterprise Artificial Intelligence"
"Chapter Three – The Intersection of Data and Artificial Intelligence"
"Chapter Four – Making It Secure – The Importance of Cybersecurity"
"Chapter Five – Data Governance – The Foundation of Trusted Enterprise AI"
"Chapter Six – The Governance of EAI"
"Chapter Seven – The Architecture of Sovereign EAI Implementations"
"Chapter Eight – Putting Agentic A.I. to Work"
"Chapter Nine – The Management of EAI Applications"
"Chapter Ten – The Creation of AGI from Agentic AI"
"Chapter Eleven – The Future of EAI and Operations Management".]

So, we're going to talk about something called agentic bots. So, we're past the world of just saying A.I. You're not allowed to say just A.I. anymore. We're now starting to go into the granularity of A.I. and we're going to try and tease that out with the book, and all the books are at the back and these books are also available electronically. And you'll see, one of the reasons why we wrote the book is that everyone's in the same boat as you, that they are grappling with, okay, beyond ChatGPT and what Taki had referred to, what do we do about it? And that's when it gets really complicated. What do you do about it?

[00:04:50 A slide is shown with the title: "About the Authors" above pictures of the authors Shannon Bell, David Fraser, and Tom Jenkins with brief biographies listed for each of them.]

Our authors are myself and Dave Fraser and Shannon Bell, a wonderful rising star in Canada. She was the CIO of Rogers. She was the CIO of Amdocs in New York. She's from Stratford, Ontario and she's the CIO and CDO of OpenText. So, she unfortunately right now is with Minister McGuinty in Sydney, Australia, doing the defence consortium there, and we can talk about that later.

[00:05:19 A slide is shown with the title: "Thought Leadership for Agentic A.I. Implementations" along with the cover of the book Enterprise Artificial Intelligence under the heading "How to build Agentic AI" and the cover of the books The Agentic A.I. Genome and Managing Agentic A.I. under the heading "What Agentic A.I. to build".]

Visually, what you'll see there on the slide is really a trilogy, and this book is really the how-to, if you read the subtitle. The next book, we're going to tease it a little bit, is the next part, but this is sort of like a Star Wars trilogy where we started in the middle and now we're going back to the beginning, that kind of thing. And then, the last book is actually the first book. So, if you're tracking all that, you're doing very well. The anticipant book is being rewritten to take advantage of ChatGPT and some of the new terms we have, because when we wrote Anticipant, we were anticipating, no pun intended, that this was going to happen. Now, we're going to upgrade the nomenclature.

But before that, as Taki was referring, we thought we would not assume knowledge, and a lot of what we're going to say is pretty dramatic. As we get to the end of this slide deck, your heads are going to be quite full. We thought we should start with the journey of what made these books possible, because as Canadians, we don't tell our own stories. So, we want to tell you the story of how we got here.

[00:06:35 A slide is shown with the text: "OpenText has delivered private and public cloud solutions in data centres for 30+ years. One of OpenText's first commercial products enabled Yahoo Search in internet cafes around the world, running in our first data centre built in our Waterloo, Ontario headquarters. Today it has the largest archives in the world." Below the text are images of Tom Jenkins and Jerry Yang in front of a computer, the OpenText web index logo, the Yahoo logo, and OpenText headquarters.]

And of course, it begins on the campus of the University of Waterloo, and this is me and Jerry Yang launching Yahoo! So, this is 30 years ago, and we started doing search engines. And believe it or not, that group of machines that are there, that's Canada's first ever internet data centre. So, let me repeat that again. The first ever internet data centre was on… the reason why it was on the campus of the University of Waterloo, back then, UUNET was the only thing that had decent bandwidth. And so, we actually located on the campus because we could get decent bandwidth.

[00:07:13 A slide is shown with the title: "OpenText was founded at University of Waterloo" Arrows point from logos of The University of Waterloo and 'Acquired software divisions of industry stalwarts' Dell, HP and AT&T, towards a central logo of OpenText.]

Now, what's different about the journey of OpenText is instead of waiting to be acquired by, fill in the blank, IBM, Microsoft, etc., OpenText is one of those unique situations where we actually acquired all the software division of Dell, we acquired the entire software division of Hewlett-Packard, and we acquired the entire business network of AT&T. So, that's sort of where the journey starts to become maybe a little different.

[00:07:44 A slide is shown with the title: "35 Years of Revenue Growth" and a graph showing OpenText revenue from 1990-2025.]

What that led to was global scale. So, the story we're going to tell you about is a story that's the private sector but from outside of Canada, because the vast majority of what we do is outside of Canada, and you can see the growth profile. So, if you're somebody in ISED, you're very happy with this chart, right?

[00:08:05 A slide is shown with numerous logos for government and municipal public sector entities as well as numerous logos for retail, education, manufacturing, automotive, and legal critical infrastructure entities.
Text atop the slide reads:
"98% of $7B revenue from international markets"
"50% Europe (including 2800+ Germany, 1,500+ France, 4,000+ UK customers) and 15% Asia".]

Because it's rare for us to create global champions, forget about national champions. Here's another interesting thing, almost everything we're going to talk about is not from Canada. Okay, let me repeat that again. Almost everything we're going to talk about is not from Canada. 98% of all the activity that went into these books is actually from outside of Canada, and more than 50% is actually from outside of North America. So, that's another sort of head-scratcher a little bit.

[00:08:36 A slide is shown with numerous logos for federal, provincial, and municipal public sector entities in Canada as well as numerous logos for retail, energy, manufacturing, tech, and education critical infrastructure entities in Canada. Text on the slide reads "Over 2,000 Canadian customers".]

But we're still here in Canada. This is where we began. It's our origin, and over 2,000… and many in the public sector, of course, at all our levels but also within all the critical infrastructure of Canada. So, it's not like we started in Waterloo and we forgot about it, of course not. We have thousands of people that work here but we have tens of thousands outside of Canada.

[00:08:59 A slide is shown with the logos for Scotiabank, CIBC, BMO, RBC, TD, SNCF, Danone, L'oreal, Nestle, SAP, and Societe Generale atop a globe.
Text on the slide reads:
"OpenText business networks trusted to manage global supply chains; $15 trillion in commerce at 5.5x Canadian GDP"
"OpenText business networks for global supply chains"
"Secure, cloud-native infrastructure operating secure, global data centres"
"Global B2B integration at scale with electronic document interchange and cloud-based data exchange"
"AI-powered visibility into supply chains with real-time operational insights".]

And remarkably, the company now does 15 trillion in internet commerce, so about five times our GDP. So, this is very much a story about the world and what the world is driving us to do, whether it's running Nestlé's entire global supply chain or General Motors, etc.

So, that's what we do and that informs what you're about to hear because we're telling you a global story, not a Canadian story, but hopefully with a Canadian twist, and I'll save that for the last slide.

[00:09:38 A slide is shown with the text:
"Ch 1 Table Stakes: Your organization MUST be digital"
"AI Must Follow Data's Rules"
"Unlike traditional software, A.I. remembers what it sees. That makes its memory part of your governance landscape. Treat A.I. learning as a life cycle: decide what models can learn, what they should retain, what they must forget, and how you'll audit them."]

So, what we thought we would do is just walk you through the chapters that are in here. To give you an idea, we get called… and this is all over the world. I have been in offices where someone says, I want to use A.I., and the first thing we say to them, you have to get off paper because A.I. can't read the paper. So, you have to become digital. I know, post-COVID, you'd be surprised, but the reality is there are a lot of papers still out there. You cannot do this if you're not digital. And of course, in Canada, we still are not digital, and we're a laggard in this regard, but understand, so is the rest of the world too. The rest the world has to become digital first.

[00:10:21 A slide is shown with the text:
"Enterprise content enables agents the same as applications"
"ECM supported business outcomes: Reduced risk and exposure, Trusted A.I. responses, Operational efficiency, Faster time-to-value, Enterprise-ready A.I. adoption."]

Now, here's the interesting thing, and this is a hot topic on Wall Street right now. Over the last three weeks, we've had $1 trillion change hands, with a T, not a B, one trillion, because Wall Street is trying to figure this slide out. And what this slide says, on the left-hand side, if you have enterprise content, which is what Taki was talking about, building search engines and what have you inside the firewall, the top arc is what we would always do. This is what you know as Microsoft, SAP, Oracle, all these applications that work inside an organization. On the lower chart, you have companies like Anthropic, you have companies, say, OpenAI, that kind of thing. They're starting to do everything that you could do on the top chart. The dilemma is, not only are they replacing all that software, that's what Wall Street's been going all about for the last three, four weeks, they're also replacing us humans. And so, this is a real dilemma. But from the positioning of us in this book and the things that you may wish to learn today, is to go do everything on that far right side. You want to get something done, you want to be efficient in your operation, whether you're a human or a robot, you still need the data. So, robots are not born as large language models with all this knowledge. They have to be trained.

And that's why we need nuclear reactors and all that stuff, because the amount of power that's required to train them. So, where OpenText comes in is after three decades, we have the largest amount of content in the world. So, let me repeat that again. A little company from Waterloo, we have the largest amount of content in the world, and the reason why is because Google or Facebook or social media, that's less than 3% of the information. For a long time, people don't realize, Boeing had more information than the entire internet for about the first ten years of the internet, and it's because when you were publishing technical manuals and every component in a Boeing aircraft was different, that was an enormous amount of data. So, that's why enterprise data has always been so far ahead. And then, as the machines started talking to each other, machines today produce 100 times the amount of information today, just during the day today, than the humans do, and that breakover point happened about 15 years ago.

[00:13:00 A slide is shown with the text:
"Ch 2 Cognitive Era Begins: ChatGPT takes world by storm"
"Data quality and governance are critical"
"The effectiveness and trustworthiness of A.I. depend on high-quality, well-governed data. Data is the "fuel" for A.I. innovation; without reliable, secure, and well-managed data, A.I. systems are prone to error, bias, and operational risk."]

So, the cognitive era begins, so ChatGPT takes the world by storm, and the thing to realize is that all of its brain comes from somewhere. Well, it came from us. It came from the content that we wrote.

[00:13:15 A slide is shown with the title: "Why Do Enterprises Need To Look At Their Own Data?" above the text:
"~90% of the world's information expected to live inside organizations (e-mails, documents, records, workflows, transactions communications)"
"Public LLMs are trained on open internet, not enough for enterprise-grade regulated insights"
"Enterprises want to own and govern their content – compliance, security, competitive advantage"
"Hyperscalers don't always solve for data sovereignty, privacy, or industry specifically"
"Agentic A.I. only delivers value when powered by trusted, proprietary enterprise data".]

Now, this is what I was referring to. This usually surprises people, but in chapter two and chapter three, we try and give you a grounding in all this content, all this data. Where is it? Well, the reality is that most of it is like an iceberg. It's actually behind the firewall. So, this has tremendous implications for anything you do and anything that society does. Most of the information is not inside ChatGPT. It's not in Gemini. It's not in (inaudible), because they can't get behind the firewall. It's only when you give permission to go inside the firewall. Now, that leads to all kinds of discussions like sovereign cloud, sovereign data, and all that sort of thing. We're not going to try and solve that today here, but understand that the majority of the data is actually not visible.

[00:14:08 A slide is shown with title "Types of Content" and the text:

• There are three types of content that must be managed by any organization
• Content that is: Public (web site), Shared (supply chain), Secure (behind firewall)
• The orchestration of the content and how it is used to train A.I. goes to the heart of maintain the three different spheres that the data must live within.]

And of course, it's PII, its health records, all the things that you know you don't want to be available. Now, there are three types of content, and this is where it starts to get very complicated, because take Canada as a country, but you could equally take General Motors as an automotive manufacturer. You have information that's in the public, it's on your website, you have information that is proprietary, you don't want anyone to have it. And then, you have information that you have to share. So, if you're General Motors, you have tier one and tier two suppliers. You have to share with them in what's called an extra net. For us, we have alliances, we have trade agreements, we share information with the Americans, with CEDA, with the Europeans, NATO, etc., etc. So, we have that sort of triple sources and uses of information.

[00:15:06 A slide is shown with an image of a big pipe next to an diagram of a typical house plumbing system. Text on the slide reads:
"A Data Lake is NOT an enterprise content management"
"A Data Lake is a big pipe. This is mass scale with no valves."
"An ECM is small pipes with valves. This is fine detail with valves (permissions)."]

Now, I have to show this because I'd be remiss, and there's a whole chapter on this. There's a thing called a data lake. A data lake can provide the content into a ChatGPT, just like a Google. It's all the public information. It's in one big lake, as you think. But everything we were just talking about, all the subtleties of sovereign cloud and things like that, those are like pipes in your house, and each pipe has a valve, and the valve is permission. And so, you have to have lots of pipes and lots of different valves, because sometimes you're allowed to know something exists and sometimes you're not allowed, and sometimes you're allowed to know something exists and you can see it but you can't change it, and I could go on and on and on. So, all of that structure, OpenText created all that 30 years ago in Waterloo. So, that whole permissions framework that we use today, you would know it in security circles and things like that, corporations do the same thing. We do individual permissions and group permissions. And trust me, when you get to Walmart and you have four and a half million people, it gets very complicated. Who is allowed to see whose salary? That's complicated.

[00:16:25 A slide is shown with the text:
"Ch 4 A.I. brain has the crown jewels – security matters"
"Secure the entire data life cycle with integrated controls"
"Mandate that all data – across collection, storage, transmission, processing, and disposal – is protected with layered security measures. This includes encryption at rest and in transit, strong access controls, immutable storage, and strict compliance with regulations like GDPR and ISO/IEC 27001:2022. Any gaps in one stage can compromise the entire system."]

And then, of course, within the service, we have all kinds of different levels we have to worry. And of course, that's where security matters. There's a whole chapter on security. Now, all of you, you respect security and you hate security, as we all do. You hate two-factor authentication, etc., and all that stuff. We have to do it. We have no choice. The trick though is how do we train these A.I. brains but still get our jobs done? That's the tricky part. We, General Fraser and I, were with the Royal Canadian Air Force this morning. There were 200 vendors and the senior cadre of the RCAF talking exactly about that. How do we get things done like Ukraine really quickly, but at the same time be secure? That is a real head-scratcher, and there's 200 people right now over at the Air Force Museum trying to figure that out right now.

[00:17:17 A slide is shown with the text:
"OpenText operates 83 cloud and data centres globally"
"+300PB of data under management"
"8,000 single tenant and 14M multi-tenant cloud customers"
"99.99% cloud up time"
"Defends against more than 13.9B cyber attacks per day globally"
"Cybersecurity for 90% of police agencies worldwide"
"Over three decades, OpenText has evolved its 83 data centre and cloud models to securely support customers".
The 83 data centres are labelled on an image of a globe.]

Now, what does that mean though, for our country and also for the world? Well, we run 83 data centres globally, because if you're running General Motors and you're running Nestlé and what have you, you're all over the world running, Canada, we could run it on five, we run 83 of them, but it's the thing that's circled there that you should see, just to give you an idea about security. I won't do a guess for the audience. Some of you may know the scale of this, but since the Ukraine incursion, this has gone off the charts. We do 13 billion cyber attacks per day, okay? That's what we have to worry about, 13 billion, with a B.

Taki Sarantakis: How many times does the Government of Canada get attacked a day?

Tom Jenkins: No, no, no, no, no, it's commensurate.

Taki Sarantakis: The public figure is between seven and nine billion times a day. And after Ukraine, after what's going on in the Middle East, you can increase that number with a lot of rapidity.

Tom Jenkins: Let's just say, I can't say the name of the corporation, but there's a major financial institution in Germany that on the day… it actually started about four days before we knew something was up, but let's just say the spike in activity that occurred for FIs in Europe went off the charts and still has been at a very high level. So, now, we have to worry in our country about PIPEDA. Europe has worry about GDPR.

[00:19:06 A slide is shown with title "The Role of GDPR and PIPEDA in AI" and the text:

• GDPR is the cornerstone legislation that governs private content in EU countries, PIPEDA in Canada and USDPR in USA.
• Since A.I. is trained on content, it is a logical conclusion that A.I. will follow GDPR legislation.
• ECM is used in regulated industries to comply with GDR legislation.
• It is logical to assume that ECM will be used by corporations to implement A.I. legislation.

"The Seven Principles of the GDPR: Lawfulness, fairness, and transparency, Purpose limitation, Data minimisation, Accuracy, Storage limitation, Integrity and confidentiality, Accountability".]

The Americans worry about USDPR. Here's the thing. If you use this information to train A.I., that A.I. has to follow the same governance rules as the data in the first place. Almost no one in the world understands that. I have sat in offices, in boardrooms to tell the CEO who went with a huge amount of investment and time with his various general counsel and governance and what have you, with this board, to tell him, that A.I. you just trained that has access to the world, guess what? It has personal information in it because you trained it with it. And then, you see, you have to realize an A.I. that's trained on something through prompt engineering, you can get at it. So, you have to realize there's a big thing here that data governance must follow A.I. governance. You can't separate them.

[00:20:07 A slide is shown with title "Danger for users educating any LLM AI: NYT" along with an image of an article titled "The New York Times sues OpenAI and Microsoft for copyright infringement" and the text:

• This will be an important case for AI.
• This case will determine the use of content without permission to train an AI.
• Corporations will be reluctant to needlessly share "the keys to the castle" with anyone.
• Those that did may already be in violation of regulatory requirements on GDPR content.]

The other issue that we all have is that once you educate an LLM, it cannot forget. You literally have to tear the LLM down. That's something a lot of people don't appreciate. There's a lawsuit going on right now where The New York Times… and different vendors will take fair use, Creative Commons, if you're a lawyer, you'll know all these things. There's a big debate going on right now, and of course, the vendors are taking the position that, well, we're like Airbnb and what have you and we're disruptors and we're just going to do it for the good of society, whereas the legal world sits there and says, well, maybe not so fast. This has moved so quickly. We have not had a lot of court cases to decide what is the rights of an individual versus the rights… so, there are a lot of debates here. That extends right into A.I., is the point of this.

[00:21:06 A slide is shown with the text:
"Ch 7: The Architecture of Sovereign EAI Implementations"
"Implement a dual-zone, hybrid A.I. architecture"
"Separate sensitive data and workloads onto secure, domestically operated infrastructure. Leverage public cloud platforms only for non-sensitive, scalable applications to balance innovation with security."]

Now, this is where it gets interesting. Chapter seven talks about what we propose, the only way to skin the cat of having things that are in an A.I. that are both meant to go outside as well as stay inside, you have to take a hybrid approach, and that whole chapter talks about, just as you would with data, how you create firewalls, you have to create firewalls with A.I. So, say you have something like Cohere. So, the Government of Canada works with Cohere, Anthropic. We do a lot of work with both of them all over the world. They can make a container of an A.I. that you can control inside your organization. That's the only way you can really achieve a hybrid. Because if an A.I. calls home, wherever home is, it just left the country, it left your organization. You don't have control of it. And if anybody tells you, we encrypt it for you, that's a joke because almost all encryption that we have today will be decrypted. Just go on the internet and you'll see, some people estimate two years from now, five years from now, who cares? The point is, don't think that because your data is encrypted, that's fine. It's not.

[00:22:07 A slide is shown of an example of "Agent & Handoff Workflow | Canadian Passport" explaining how data passes through classified and declassified zone during its validation process.]

So, I put as an example, and we have it in the book, I didn't call it a Canadian passport, I just said passport, but that's a classic example of, we have to live in three different spheres at the same time to issue someone a passport.

[00:22:38 A slide is shown with the title: "Canada's Sovereign Cloud & A.I. Platform to Power National Innovation and Global Trade" and the text:
"Sovereign Cloud Platform Architecture"
"Applications & A.I. Services (e.g. Cohere, OpenText) <--> Platforms & Operations (OpenText) <--> Infrastructure Fabric (e.g. Bell, Telus, CGI)"
"Managed Service: Monitoring, Observability, and Cybersecurity."]

And then, of course, in the Canadian aspect, I didn't put this slide in the book because the book's really meant for the rest of the world, but we have all the stack to be able to do this. We're one of the few countries in the world outside of the two superpowers that actually has the ability to do this.

I don't have a slide on this but I was trying to explain to the Air Force officer, general officer cadre, Canada's a superpower in software. Let me repeat that again, because we don't say this to ourselves. Canada is a superpower in software. Do you know in the Toronto to Waterloo corridor, we have more tech employees than Silicon Valley. And by the way, that's not this past year. That's been for almost six years now. We crossed over. Do you know when they're head-hunting in the valley, they go to Toronto. You have to realize that environment, that ecosystem is bigger than anything in New York, in Austin, Texas, etc. It's a great crown jewel that we have.

[00:23:43 A slide is shown with the text:
"Ch 8: Putting Agentic A.I. to Work"
"Deploy agentic A.I. to drive enterprise value"
"Accelerate productivity and adaptability by adopting agentic A.I. applications that autonomously perceive, plan, decide, and act – enabling automation of complex workflows and reducing dependency on manual intervention."]

Now, the most important thing about agentic A.I. is that it's productivity, and I'm going to talk about what's the difference between a ChatGPT, a superintelligence, and agentic A.I. in a minute, but agentic A.I. is an agent. It's something that can be a co-pilot to a human being or replace a human being but it's making a decision on one thing. So, in some ways, it's brilliant, and then in the other ways, it's really dumb, because it can only do one thing. We as humans, generally, we're in the middle of book two right now on the genome, we think on average, a knowledge worker is actually five agents, on average, but I'll come back to that in a minute.

[00:24:27 A slide is shown with the title: "The Agentic A.I. Migration of the Enterprise" and the text:
"Org chart drives agent selection and training"
"Workflows drive orchestrators".]

So, what's going to happen? What are you going to end up doing as leaders? What is going on in private sector right now, not in the future, going on right now, is we're going through org charts and we're replacing them with bots. We're right now improving productivity and, in some cases, replacing. Things that you call a workflow is now going to be known as an orchestration, okay? And an orchestrator will be able to do workflows in ways that we as human beings that have to eat, drink, sleep, go home, visit with our family, get educated, etc., the org chart for a bot organization which is (inaudible), that consumes electricity and excretes heat, totally different world.

[00:25:20 A slide is shown with the title: "Agentic Genome for the Enterprise" and the text: "The Agentic Genome for any organization can be mapped from the HR org chart and the workflow of the organization. This represents a tremendous opportunity to discuss new possibilities with our users".]

So, that brings us to a genome.

[00:25:24 A slide is shown with the text: "A Canadian Agentic Genome" and the cover of the book The Agentic A.I. Genome.]

And that's the part that we wanted to talk about was if we could engage in and we challenged the Royal Canadian Air Force today, why wouldn't we leapfrog all the other countries? So, think about that for a minute. As Canadians, we emulate the Americans. We're actually small enough that all of us can be in one room and have a discussion, just like the Air Force, but we're big enough that we can do global scale. That's a rare thing for a middle power and yet we're able to do this.

[00:25:58 A slide is shown with the title: "Canada as the First Machine Driven Country" and an image of a map with robots placed along it.
Text below reads: "Participants:

• Federal, Provincial, Municipal Governments.
• Regulated Industries.
• Allies and Trading Partners.]

So, imagine Canada as the first machine-driven country. Now, you may want to say, I don't want to live in that country because I'm a human being, I don't want to be in a country with bots. The reality is society has to solve this. We have to solve this, and I'm not here as a social engineer, etc., but I'm trying to give you an idea. Technically, if we want to keep ahead and at least keep our position from a quality of life and standard of living, we better learn how to do this. Because I can tell you, and that's why I did the slides at the beginning, I'm a first-person witness to everything going on in the world right now, they're starting to figure this out. Think of Latvia, Estonia. They're very small but they're very digitally savvy.

[00:26:43 A slide is shown with the title: "The Agentic Genome of Canada" and the text: "Mapping the roles and the workflows leads to the simulation of the Agentic Genome for the Government of Canada" above an agentic A.I. map of the Government of Canada.]

Okay, so, why couldn't we build an agentic genome for Canada? Now, I took this all from the internet and the staff that are working on it, there's no proprietary information here.

[00:26:55 A slide is shown with the text:
"Top 10 Ministries
1. Treasury Board Secretariat (TBS)
2. Employment and Social Development Canada (ESDC)
3. Health Canada
4. Public Services and Procurement Canada (PSPC)
5. Canada Revenue Agency (CRA)
6. Immigration, Refugees, and Citizenship Canada (IRCC)
7. Innovation, Science, and Economic Development (ISED)
8. Transport Canada
9. Environment and Climate Change Canada (ECCC)
10. Indigenous Services Canada (ISC)"
"Each department operates:

• 100 functional agents
• 5-8 domain orchestrators
• Human-in-command control points".]

But we then had fun. We said, why don't we take our top ten ministries and postulate how we would do agents to accelerate everything we do in the federal government?

[00:27:06 A slide is shown with the title: "1,000 Agents in Genome" and the text:

• Starting with 100 agents per major ministry then orchestrated by 10 major workflows between these industries.
• After considering existing org chart of human resources, the following lists of agents are a first step on the Agentic A.I. journey for this organization
• Human-in-command control points

1. Treasury Board Secretariat (TBS)

• 100 Explicit Agents
• Policy & Authority Agents
• Expenditure Management Framework Agent
• Treasury Board Policy Interpretation Agent
• Program Authority Validation Agent
• Statutory Spending Authority Agent
• Vote Structure Interpretation Agent
• Digital Policy Interpretation Agent
• AI & Automation Policy Agent
• GC Enterprise Architecture Policy Agent]

And there you go. I'll just very quickly… but you can decompose Treasury Board, and many of you have done your tours at TBS. Expenditure Management Framework Agent, Treasury Board Policy Interpretation Agent, Program Authority Validation Agent, these are all positions within TBS. A lot of this is based on an algorithm that you all learned, but we can take all the data that's in GCdocs and train an agent to do that.

[00:27:38 A slide is shown with an agentic genome diagram for the Government of Ontario.]

Taki Sarantakis: There's no data in GCdocs because nobody uses it.

Tom Jenkins: Well, you have to actually.

Taki Sarantakis: I'm sorry, I'm sorry.

Tom Jenkins: Okay, but the reason why no one uses it… that's an interesting aside. Do you know the oldest piece of software used in OpenText across the world, and we're talking 1.2 billion users? It's GCdocs. Why it has not been upgraded, it's not for me to say. You're using something we designed 25 years ago, but it's critical for records management and for governance, etc. But anyway, that's a whole other story, Taki. The Government of Ontario, there's no reason why a provincial government which is heavy in education, health care, etc., it's the same thing.

[00:28:22 A slide is shown with the title "Map All of Canada" and the text:
"The flow of this map starts with the governments and profiles the following agentic maps:

• Federal Government of Canada and the 10 largest departments by size
• Provincial Government of Ontario and the 10 largest departments by size
• The crown corporations and the agencies of both government for utilities and transportation
• The Municipal Government of Metropolitan Toronto

Then the map continues with the most important regulated industries:

• Royal Bank of Canada
• Air Canada
• Bell Canada
• McCain Foods
• Canadian National Railway

The map then adds alliances and trade partnerships:

• Alliances: NATO, NORAD
• Partnerships: USMCA, CETA, CPTPP".]

And there's no reason why we couldn't do a map of all of Canada.

Because remember, we run the content of Air Canada, of Bell Canada, of McCain Foods. There's no reason why we couldn't do this.

[00:28:36 A slide is shown with an agentic genome diagram for Bell Canada.]

And you do the same thing, there's Bell's agentic genome.

[00:28:40 A slide is shown with an agentic genome diagram for Royal Bank of Canada.]

There's Royal Bank's agentic genome.

We had this morning, the RCAF and Boeing and Lockheed and MDA and Bombardier, and there's no reason why an agent that's sitting in a sensor on a fighter jet that has metal fatigue can't just talk to the agent at Boeing or whoever is doing the in-service support, right? That's the kind of unbelievable operational productivity leaps that you can make if you have the imagination to get your head around this.

[00:29:14 A slide is shown with a chart of agent training by function and content source.]

And these are all the training and the functions and where the content sources are. This is not science fiction. We're doing this right now with some of the largest corporations in the world inside their supply chain. I just made it Canadian so you could relate to it. But understand, let me repeat this again, major corporations that have millions of employees are doing this right now. So, you think that Canada has productivity lag? Just wait. We don't do this, we better catch up.

[00:29:46 A slide is shown with the text:
"Ch 9: The Management of EAI"
"Integrate digital agents into workforce planning and management"
"Treat A.I. as an extension of your workforce. Define clear roles and expectations for digital agents, adopting HR practices like job description and performance objectives to ensure alignment between human and A.I. team members, and to foster organizational understanding and acceptance."]

So, the management of EAI is where I'm going to hand over to General Fraser to talk to you.

[00:29:51 A slide is shown with the covers of the books Managing Agentic A.I. and The Anticipant Organization.]

If your head's not spinning now, wait until he talks. He's going to talk to you about the reality of having robots in your organization.

General.

[00:30:03 David Fraser walks up to the podium and Tom Jenkins walks offstage.]

Major-General (Ret.) David Fraser (Canadian Armed Forces, Member of the Board, OpenText): Tom, thanks very much. Can everyone hear me? So, why is an old general standing up and talking to you?

[00:30:10 Taki Sarantakis points to the chair beside him.]

You got to go up there.

Taki Sarantakis: Tom, just come sit over here.

[00:30:13 Tom Jenkins walks over and sits down next to Taki Sarantakis.]

Major-General (Ret.) David Fraser: You got to say words twice to an Air Force officer.

This started, actually, 15 years ago, Tom and I, from two different paradigms. I came from your paradigm, talking to a private sector guy in the paradigm, and we ended up coming to the same conclusions from two different paradigms, and that's the book we talked about called The Anticipant Organization. And so, we started talking, and now, fast forward today, I stand in front of you as a hybrid leader. I'm not a general. I'm not a civilian. I wake up most days, I don't know who the hell I am, but I am who I am, and the function is that I'll never get away with… I will always be known as that.

[00:30:56 David Fraser holds up a copy of his book Operation Medusa.]

I work for an FI company today and I can't get away from my generalship, okay? They still call me General, whatnot, but we all talk about the same thing now and it's about winning, and how do we win in an environment where there's a whole bunch of things that are changing, where we can't change that, it is what it is? And that's what I'm going to talk to you about in a couple of minutes here.

[00:31:18 A slide is shown with an image of a drone swarm and the text:
"A drone swarm: Manage your machines before they manage you!"
"The domain in which machines talk to each other by means of electronic or photonic pulses moving at or close to the speed of light."]

So, this is a swarm. And ten years ago, this was something that for me, I was thinking about the possibility, how do I deal with a swarm? Four years ago and a month ago, Ukraine was invaded by Russia, okay? And I was talking at midnight to Lisa LaFlamme, and four years later, I'm still talking to CTV, but we're on our fifth war in that period of time, but we're still talking about this and this is about a swarm, and a swarm now is not a theoretical thing. It's a reality. This is now about machines talking to other machines in nanoseconds, and you and I are not part of the conversation. And if you look at what's going on in Russia and in Ukraine, they're attacking each other and trying to overwhelm each other with cheap equipment, cheap equipment going after multi-million or billion dollars pieces of equipment, and that's the new reality, and Task Force Rubicon where the Russians are doing it, and the other thing the Ukrainians have done with this technology about swarm is changing the entire supply chain. It is happening and changing so rapidly. It's just the new reality.

[00:32:43 A slide is shown with the title "4 Dimensions" and the text:
"Unit of productive time: nanosecond"
"Speed of transaction: speed-of-light (c)"
"Scale of input: unlimited"
"Domain of interaction: suprahuman".]

So, what's (inaudible)? Here are the four things that actually have changed and that are affecting us every day. I work for an FI community right now, a bank. Most people in my regiment don't realize I work for the bank because I wrote three books since then. There are pictures in the book so they do know it was Dave Fraser, so that's a good thing, but here are the four things that have happened. Time compression, and that's what Tom and I started talking about. I used to have days and weeks to make a decision. Now, I don't have any time to make a decision. I have to anticipate because the decisions are happening so rapidly.

Do we have any programmers in this room?

Unidentified Speaker: (inaudible)

Major-General (Ret.) David Fraser: Okay, that's important though, because I used to talk to people. Now, I have to talk to programmers, and the programmers who are writing the ones and zeros, you have to do the same thing that I do with talking to humans, and you've got to make a machine do the same thing that a human does. Inside a BMO, where I work, we actually have machines that we give names to because they are service providers. So, when we go to a meeting, there's a machine there with us too, because we're actually talking to that machine because it's now operating at speeds and complexity and with no mistakes than a human being does. That's the new reality. The next thing we talk about is how fast is this happening, speed of light. The scale is unlimited and it's now the supra human, and this is just the new reality. So, we're now leading people and machines. So, here's the bottom line up front. It's still all about people, okay? Programmers are still really important people. Now, they've got to come out of the darkness and actually talk to people, and that's normally when they're starting to squirm and whatnot, but they're really important. So, I'm glad to see you out here.

[00:34:34 A slide is shown with the text: "Domain: suprahuman" and an image of the UBS trading floor in New York in 2015 next to an image of the UBS trading floor in New York in 2016.]

But here's the realities of what those four things are doing, and this is old, you look on the left-hand side, this was the trading floor, and on the right-hand side, that's what happened in a year based on technology. And A.I. is now working so fast, it is now going to fundamentally change, engineering, lawyers, doctors. It's going to change it. And right now, A.I. can actually take and is replacing it five for one, five engineers for one bot, and it's just accelerating. The impact on organizations, off the scale, and this is just the reality that we're dealing with, and that's something that we can talk about if you want.

[00:35:18 A slide is shown with the text:
"When you cannot participate, you must anticipate"
"Anticipant roadmap – Transform your organization in six steps:
A) Follow Anticipant Principles – Learn, Adapt, Embed
B) Prepare Yourself – Identify, Assess, Admit
C) Control Your Data – Capture, Curate, Link
D) Build Anticipant Teams – Engage, Guide, Deploy
E) Remap Your Organization – Prepare, Shorten, Build
F) Rehearse To Survive – Identify, Prescribe, Rehearse".]

But it's not all lost, okay? You can control this. And when we wrote this book, Tom and I went, well, we just can't just say that the sky is falling in and what are we going to do? Here's how we can get yourself and your organization, how do you come up to speed with it? And here's just a pathway of how to do it so in fact you are managing your data, you are controlling the situation, and you're getting the desired effects you want and you are actually starting to anticipate as opposed to react.

[00:35:52 A slide titled "Anticipant Organization is shown with a chart listing features of an Anticipant Organization. Two columns one labelled YES, and one labelled NO are shown. Under the YES is the text; approach: proactive, org chart: elastic, processes: collapsable, operations: adaptive, plans: crowdsourced, work: iterative. Under NO is the text; approach: reactive, org chart: flat, processes: lean, operations: automated, plans: sampled, work: perfect.]

And how do you know you've got an anticipatory organization? Well, here are some simple things to take a look at, yes and no, and I just want to comment on a couple of them right now. Elastic, quick war story. I had an organization overseas, 20,000 people, and it was a hierarchy. You had a person at the top, a CEO, me, and then all sorts of people. Well, one night we had an incident where something bad was happening. My communicator came over and said, 200 kilometres away, an American sergeant was in the fight for his life and I collapsed the entire organization over radio. I called him and I said, this is the boss, you are my number one priority, I'm going to get all this stuff done for you, I'm going to start stacking stuff up above you, we're going to save you tonight, and we collapsed the whole organization. That's elastic. And he never met me before, I've never seen him since then, but we can take organizations like governments and businesses, and you can collapse them when you have a crisis and then blow them back up. That's what we have to do in order to survive and to prosper. So, that's one of the things.

And the other thing is it has to be iterative, and that's the most important thing, but iterative means we have to trust your people, okay? And I used to go into an organization like this and I would say, you have my unqualified trust, I in turn will earn your trust, and I'm probably being the meanest person ever because who wants to disappoint the boss? When something happened, I never had to say a word. They showed up. They would be so upset that they disappointed the boss, and that's just you have to trust your people and you have to give them the authority that they can go off and do things, and how did we do this and how did we make an organization like that? We started to war game it, we started talking about scenarios, and I started empowering my people to say, don't come back to me and asking me for the decisions, you make the decision based on what we understand of each other. So, this is how you can start from where we are to go to where we have to be, and it has to be inclusive of your machines.

[00:38:01 A slide is shown with the text:
"Ch 10: The Creation of AGI from Agentic AI"
"Complete paths to AGI"
"The transition to AGI is shaped by two dominant hypotheses: the Scaling Hypothesis (AGI emerges from scaling current models) and the Discontinuity Hypothesis (AGI requires fundamentally new architectures and reasoning). Executives should monitor both trajectories for strategic planning and investment."]

And with that, I'm going to turn back over to Tom.

[00:38:05 Tom Jenkins walks up to the podium and David Fraser walks offstage.]

Taki Sarantakis: See, he's from the Army, so he knows. He just went and sat down, so he didn't have to be invited.

Major-General (Ret.) David Fraser: I can follow instructions.

Tom Jenkins: We could get going on the difference between the Air Force and the Army.

So, a couple of things to just bring us home. There's something else, and I'm going to go back to science in a minute here, just to give you an idea to how to think through all this A.I. stuff, but there's another reason why we want to pay attention to agentic A.I., because we may be able to build superintelligence from the agentic. Okay, now, right now, the Americans have a horse race of the hyperscalers, so Elon Musk and all the others, each one of them going off and doing their own Manhattan Project to build the superintelligence, and they're taking an approach which ultimately will probably yield what they desire, which is the super brain, but they will need nuclear reactors, they will need a lot of heavy lifting. We may be able to create it in our country because we're small and we may have something that's an advantage. So, let me unpack this slide for you.

[00:39:27 A slide titled "Use data required to train EAI for AGI by Orchestration" is shown. A bar graph chart showing A.I. versus Data and the factors for secure information management is displayed.]

So, this will be the most complicated slide. We're already almost done. So, you're almost there, you're almost there, but I want to unpack it because I want you to see, on the horizontal dimension, it says parameters. So, parameters are little pieces of information. Let's just say it like that, okay? And on the vertical dimension, what do you want to do? Do you want to do a single task, do you want to discuss something, or do you actually want to solve a wicked problem, something really hard, solve cancer, solve whatever? If you take that grid, I've circled red because that's what we're talking about, we're taking only a small number. So, you guys would have by now heard about large language models. What we've been talking about here are small language models. There are only a small number of parameters because you're asking it to only do one thing. Even though it's brilliant, it's also really dumb. It only can do one thing. It doesn't have the intelligence of a human.

What you know, what sort of landed on everyone was ChatGPT, generative A.I. That's halfway. You can talk to it and it needs about a billion parameters. What the press is talking about is the top right, solve anything, frontier model, superintelligence, or AGI. There is a debate among scientists. What's being done right now with quantum computing, brute force, generally, everyone agrees that is a way to do it, but just like Geoff Hinton at U of T and Nigel Shadbolt at Oxford, there is a contrarian group that says, hey, maybe you don't need to go do these really big quantum-based computers, but rather you do a massive network of little agents. This is an open debate right now in the community. If we as a country went after that genome thing, it gives us a chance actually to build something quite special.

[00:41:42 A slide is shown with the text:
"Ch 11: The Future of EAI and Operations Management"
"Drive the shift to automotive operations"
"Lead your organization to move beyond reactive monitoring by investing in AI-powered, self-healing systems that proactively prevent incidents and optimize performance."]

So, that chapter is on all of that, so you can read that. That walks you through the possibility of it, and we are now right at science fiction. No one knows.

So, it's your turn. Go build your own agentic bot now. We've given you the how to, this is how you can do it. You have a country that's blessed with the capacity to do it. The methodology of large language models was built at U of T. At the Air Force this morning… yeah, I brought them. We brought a whole bunch of books for you, not just that book but we were over at the Air Force, so we thought, okay, we'll bring some of the books over. So, we stole some books out of the museum. They're in the back.

[00:42:33 Tom Jenkins holds up a copy of the book Pathway to the Stars by him and Michael Hood as well as a copy of the French edition titled La Voie vers les Étoiles.]

So, in here, I got too many things in my hands here but here is a book that I wrote with Mike Hood, our former Commander of the Air Force. Do you know, in this book, U of T and NRC did something amazing? They designed the G-suit, okay? That was done in this country. We were the first ones to figure out how to do it. We have done amazing things before. We should know our history. So, when I sit and say to you, why don't we build an agentic genome? Why don't we build a different way to do superintelligence? We've done this before. We can do it again. We just have to have the imagination. The other thing, and I promised Taki I wouldn't do too much. You didn't come to a bookstore today, but maybe.

[00:43:23 Tom Jenkins holds up a copy of the book Aviation Nation by him and Michael Hood.]

We also wrote this for grade six science. So, this is called Aviation Nation.

[00:43:33 Tom Jenkins turns to the first page of Aviation Nation and holds it up.]

And if you pick up a copy, we integrated it with all the grade six science curriculum, lesson plans, podcasts all over the country so that a 12-year-old would learn, because that's the year they become an air cadet. So, a 12-year-old would learn. And so, we've been running this now for two years. They are completely full at the Air Cadets now.

[00:43:55 Tom Jenkins holds up a copy of the book Ingénieux and a copy of the book Ingénieux Junior, both by him and David Johnston.]

And then, just for fun, we also brought an oldie but goldie. This is for the centennial. I got to do this with the Governor-General, that was a lot of fun, but it was 150 of our greatest inventions, and we did the children's book as well. So, all those books are in the back. Please get them. Because when someone in your staff says, that's too big, the Americans should do it, or fill in the blank, somebody else should it, no, we should do it. Why not?

[00:44:25 A slide is shown with an article from The A.I. Mag titled "Could a National Public 'CanGPT' Be Canada's Response to ChatGPT?" next to the text:
"What about a CanGPT? We could call it MAPLE.AI"
"These Canadian organizations could team together to deliver this: CBC, Cohere, OpenText".]

So, here's my last slide. Why don't we do this? Why don't we do our own version of ChatGPT? Now, I'm being somewhat entrepreneurial here. Why don't we just call it Maple? And why don't we get OpenText and Cohere and CBC to build it? We have all the technology to do it. It's our content. So, let's do it. Why would we wait for, fill in the blank, somebody else to do it? Why don't we do it? We could build our own ChatGPT for our own country.

So, there's your innovation library that we brought and the educating that we're doing.

And by the way, just to give you an idea about what I tried to start at the beginning, these books are available in 10 languages because we have way more customers that speak, take your pick, Japanese, German, Arabic, way more than Canada.

There you go.

Taki Sarantakis: Thank you.

(Applause)

[00:45:20 Tom Jenkins walks offstage and sits down next to David Fraser and Taki Sarantakis.]

So, we're going to go for about 10, 15 minutes, the three of us. Then, we're going to say goodbye to our virtual audience and we will continue here with questions for the audience for those of you that made the trek in cold, cold Ottawa, not as cold as yesterday. So, we're going to try to keep this in kind of a governmental context for the people watching this, and I'm going to pick on a word that each of you brought forward, because it seems to me those are the two words that historically the government and governments across the world have used to either do things or not do things. And today, that's flipped, and the private sector uses these two words differently than governments do. So, I'm going to start with the Air Force. The first one, the first of the two words, is speed, okay? We don't understand speed in government, and I don't just mean the Government of Canada, I don't just mean provinces. Speed in the real world, outside of government, as you said, it's measured in nanoseconds or in seconds or in minutes or maybe 10 minutes, maybe an hour. How do we measure speed? We'll get back to you next week, we'll get back to you next month.

Tom Jenkins: If the derivatives trading floor at Citigroup goes down, anybody know how much they lose per minute? $10 million a minute, okay? Speed is real in the private sector.

Taki Sarantakis: So, talk to us a little bit about speed as you've seen it within the government, and not just at the Air Force and at CAF but also in terms of how you see how decisions are made, how procurements are done, how RFPs are drafted. Talk to us a little bit now about your experience in the private sector.

Major-General (Ret.) David Fraser: Well, I'll give you my experience both in the private and the public sector. In the public sector, it's old, so bear with me. I am old. We had something called whole-of-government in Canada and the forcing function was called Afghanistan. It was probably the best time I can say I was ever Canadian. I stood up, I watched Ottawa lead the way because of a forcing function and it was called the life of a man and woman wearing a uniform, and this country came to the fight and gave me everything to the point where I had to push it back, okay? And we were able to take a procurement system that is as convoluted and as clumpy, and I can say words that only soldiers can say, Air Force, they don't understand these words, but I can say these words, and we were able to get it done to the point where the rest of the world, to Tom's point went, wow, how did you guys do it?

And let me give you an example. I went down to Australia, simple thing, and I was talking to the number two of the Australian Defence Forces and he went, "Fraser, sit down. How the hell did you guys in Canada do this? Because we wrote you guys off years ago. You've come on the world stage, you have done something no other nation has done in the world. We all talk about it, but you guys are doing it. I've got my whole staff down there and you go down there and tell my staff how Canada was able to actually break through the logjams and actually start delivering stuff on time in a timely way because of the forcing function called a war." We can do it.

Taki Sarantakis: Can you define a forcing function for us?

Major-General (Ret.) David Fraser: It was time, speed, compression. When I showed up in Afghanistan, we were driving around in G-wagons. When I left there, we were driving around in something called a LAV that had come from a LAV 3 to a LAV 4 to a LAV 5. It was happening as fast as what was going on in Ukraine today. We can do this.

Taki Sarantakis: So, I'll pause for a second. We're good in crisis, I think is what you're saying. Keep going. What happens when we're not in crisis?

Major-General (Ret.) David Fraser: Well, you can answer that question better than I. And in the private sector, well, here's the thing, the private sector is in crisis every day. Ask yourself, this is the fundamental question, who is responding to whom? I used to ask my intelligence staff and my operator staff in the private sector and the public sector, because I used to run manufacturing firms, I now am with BMO and I still ask, in BMO, this question, if we're responding to our competition, we're losing, how do we get to the point where they are responding to us? And that is the fundamental question. It doesn't matter what paradigm you're in, and this is what Tom and I came to. If you're responding, you're losing. So, what are you going to do about it? And if you've got to start breaking some walls and changing some authorities, you're compelled to do that, otherwise you become redundant and you disappear. And in the public sector, we are planning geniuses. We don't execute very well. And in the private sector, they are executing geniuses and they don't plan very well, okay? We're not perfect on either side of the paradigm.

Don't let facts get in the way of a good idea or an opinion. That's my problem in the private sector right now. We're executing, we're running into the sound of the guns, and you get yourself killed. That's not smart. In my former world, we didn't move fast enough until I got into Afghanistan and I almost ran out of ammunition one day. That's a really bad day when I have to call back home to my boss and say send ammunition, and they send a plane down to New Zealand to get it because we don't produce it here in this country. And when you need ammunition or you need a part or a widget, probably the same time that Canada needs it, everyone else in the world is using it at the same time, okay? That's a really bad situation. And both paradigms, we can do it. Here's the thing, it's about people. It's about leadership. It's about, you know what's wrong? Fix it.

Taki Sarantakis: So, that speed, and think back, those of you that can because some of you are probably too young for this, think back to before the internet, think back to before Google. Government had a monopoly, okay? On a lot of things, including, in a lot of cases, information, including on kind of whether or not you get unemployment insurance, whether or not you get the Canada child tax benefit, etc. That kind of monopoly on services, that kind of creates a monopoly mindset which is kind of like, yeah, you want same day service from department X, we'll get to you when we get to you, okay? Monopolies are known for a lot of things. They're not known for speed, okay? And we have to break that mindset because speed, if you're slow today, you're toast, okay? A lot of deputies, your deputies, your bosses, are actually now saying the big used to eat the small, today, the fast eat the slow, okay? And think about that in your jobs, because when you say, yeah, we'll get that back to you next week or next month or in a couple of days, the world isn't waiting for you, okay? Everybody now works at a different level of speed than the public sector works.

So, that's word number one, speed. Word number two is scale. You talked about Walmart a moment ago during your presentation, and I think you mentioned four million employees, if I remember correctly. That's another interesting word in a governmental context. In a governmental context, whenever we couldn't do anything, again, think back to before the internet, think back, and again, I know it's very difficult for people to do that, we would always say, well, we can't, we can't do that because unlike the private sector, we serve everybody, we have to serve 25 million Canadians, we have to serve 30 million Canadians, we don't just serve our customers, okay? But now, people serve millions and millions and millions and millions, if not billions, of people all the time. And then, we also would say, well, we can't do that because we're very slow and because we have to do two official languages. How many languages does BMO use? How many languages does OpenText use? So, the second one is scale. Talk to us a little bit about scale, because it seems to me that little kids now can serve, with apps and things like that, they can serve 100 million people today and that's more than twice the size of the Canadian population.

Tom Jenkins: Well, there's no question that almost anything we do in tech, we start with planetary scale. We have 1.3 billion users, 1.3 billion users. So, we do not have the luxury of subdividing that or piloting something. We have to go live. That's why there are so many cyber incursions. Languages, 30. I think there's at least 30. We'll do 10 generally in marketing for major areas of marketplaces. But yeah, scale, you have to be planetary or you will not survive, period. There are no borders in the digital world, and this is going to be a big challenge for us when we think about sovereign and what have you and we think about creating things. Think of the structures that we've built over the years to retain our Canadian identity, CRTC, that kind of thing. In this world of the internet, good luck, right? There are virtually no borders.

Taki Sarantakis: So, speed and scale have fundamentally changed in the real world. I don't think they fundamentally changed in the public sector. And again, this isn't of the Government of Canada. This is almost all public sector entities. We have not adapted yet to the speed and scale of the new world. Then, the last thing I want to talk about before we let our virtual audience go, talked a little bit about agents, talked a little bit about organizational structures, I think a lot of people now are kind of going, am I going to lose my job? What's going to happen to me? What's going to happen to my kids? And that's very natural because there are very, very few things that should kind of make you go, ugh, than kind of your livelihood. Major-General, you talked about, at the end of the day, it's still going to be about people. So, reconcile those two things for us, those two statements.

Major-General (Ret.) David Fraser: A private sector example, and it was a company that did, a highly-skilled workforce that did, a machining of very high tolerances and they brought in robots, and the workforce were absolutely against it, we're going to lose all our jobs to the robot. They didn't lose one person because they had to actually program the machines and then supervise the robots to do what they had to do. They expanded the capacity of the company but they took the people and they gave them a higher level of education and training so they could actually do what they need to do. And in our first book, we talk about the Agrarian Revolution, the Industrial Revolution. And every time we've done it, we've allowed the humans to do something different. And early on, it was to do things less dangerous and allow us to use our brains in more creative ways. Well, we're now at a stage where we're not invited to the party because it's machine to machine, but who's writing the code? Who's supervising those?

Because right now, if I look at all this technology and whatnot, who's got kids in the room? Okay, I got two grown-up boys, okay? When they were three, they needed maximum dad supervision. My kids are 37 now. They still need maximum dad supervision, all right? They are always going to need dad supervision, all right? The only thing now is that when they call me, it's a lot more expensive than when they were young, okay? We still need people to supervise the machines, okay? They do not operate by themselves. They have incomplete judgement. Look at Elon Musk and what he's doing. I mean, he's done amazing things but he still has to have engineers and still has to have people to kind of test it, supervise it. And when things go wrong, well, who's going to fix it? So, there is this need for us still.

Tom Jenkins: So, there's a huge point here that all of you in the service must remember. To use Dave's analogy, you have to stop thinking about human in the loop but human on the loop, okay? And our book talks about that a little bit, and we talked about the Air Force today, but there's something more profound that all of you need to realize. If you wish to govern something, so say this is the country of Canada, or you want to govern a city or you want to govern a corporation.

Taki Sarantakis: Or the natural resource sector or the transportation sector.

Tom Jenkins: Whatever you want to govern. If you can't speak the language, you can't govern. If the service itself is not literate in this, they will not listen to you and they will either go elsewhere or they'll ignore you. If the service wants to govern, it must be at the cutting edge, because if you're not, you're the slow. So, what I was trying to interject earlier, and I'll preface this by saying my wife sent me an Instagram this morning, and it was a mother talking to her daughter, trying to explain about Robert Redford dying, and she said to her daughter, "Well, Robert Redford was my mother's Brad Pitt," and the daughter said, "Who's Brad Pitt?" Okay, so, I had to preface it this way because Richard Dicerni is someone who taught me so much about governing, and I would sit at dinner with Richard and I'd talk about a policy thing we were working on. And I would say, "Richard, we're not working on this, we're not advancing this." And he said, "It's okay. When the risk of not doing something becomes greater than the risk of A, B, or C, we will do something." The service is now at that point. Not doing something is more risky. The service must do, because if it doesn't, the alternative is you made a decision, you decided to do nothing. That's way more risky.

Taki Sarantakis: Exactly. So, I'm going to ask you each one more question, but the way I would summarize this part of the conversation and the way I summarize to a lot of people is humans who use A.I. will replace humans who don't use A.I.

Tom Jenkins: Hundred percent.

Taki Sarantakis: And if you internalize that, you have learned a lot. Humans that use A.I. will replace humans that don't use A.I. Now, one last thing because we're talking about it a lot today, we've been talking about it here. Sorry, not today but we talk about it in Canada constantly now. We never did, except at the School, we've talked about it for many, many years, sovereignty, Canadian, being kind of a master of your own domain, so to speak. You each work at great Canadian entities. Tell us why it's important that those are Canadian companies, and I'm going to start with you. What does that mean to us in the room? What does that mean to Canada? The BMO is headquartered here. The BMO's management team is here.

Major-General (Ret.) David Fraser: Next to your health, your financial well-being is probably the next most important thing. And from a governance point of view, from a protection point of view, from a privacy point of view, we're Canadians. It's got to be kept in Canada. We do not outsource that and make you a digital citizen of another country or a digital customer of another country. It's about being Canadian. And working for BMO, we take that seriously and we will do everything we can to make sure that those systems and whatnot, and there's an awful lot of compliance that we have to do, government compliance we have to do, but it's about the customer and protecting the customer's privacy and your financial well-being so that you can actually do what you need to do, and I deal with it every day. I've got a segment inside a BMO of a million market share that I'm responsible for, and that's what we talk about every day to make sure we get it done right, but it's a hybrid approach. There's some stuff, going back to what Tom said, it's public but there's a lot of stuff that's about you, it's private.

Taki Sarantakis: But I want to push you just a little bit. When I do my banking, I'm doing debits and credits. Why is it important that that company is Canadian rather than that company, whether I'm getting my debits and my credits from a company in Europe?

Major-General (Ret.) David Fraser: I think it goes back to trust. Do you trust the organization that you're dealing with to take care of you, that you don't go there one day and there's nothing there in the account? It's about the trust, and do you have a mechanism to go back to your government that can actually go back to that, that company, that Canadian company, that can regulate it and actually take care of yourself because you're one person of 35 million? And so, I look at that as a trust factor, as a credibility factor, that I don't have to worry about it at night when I go to sleep, or my children or my wife. I don't have to worry about it.

Taki Sarantakis: So, Tom, why is it important that OpenText is Canadian? You told us that almost all your customers are non-Canadian.

Tom Jenkins: OpenText would not exist today if it wasn't Canadian, and let me explain why. Trust, he just said it. And by the way, just on sovereign cloud and all that sort of complexity, the Munk School has been doing work on this, and I think you will see a white paper come out from Janice Stein and her crew. It's a complex topic, so just as a thing to watch pop up in the next few weeks. Trust. I said this publicly this morning to the Royal Canadian Air Force. Do you know that OpenText runs the content for 40, 4-0, 40 militaries, 40. We're a middle power. Know your history of Canada, because I didn't appreciate this, because I was surprised, 40 different countries' defence departments. That's your core, right? They would rather have a Canadian have the keys to the castle. We are in many ways reliving our past history. In those books that are at the back here Ingénieux, etc., do you know after World War II, Canada ran all the currency. The Canadian Mint across the street printed the money. This is an era of counterfeits and all that stuff. Canadian Mint printed all the money. Canada Post taught everyone how to run a post office. Bell Canada had Bell International set up and ran telephone exchanges.

This country has had this role before. And in the world we're moving into with superpower competition, this country has a tremendous role to play as a middle power but we have to grasp it and understand why, but we cannot play that role if we are not at the very cutting edge, if we are not the best, and that starts with the service all the way down. We have to be the best if we want to play that role, but that's hard because that means for all of us, we have to demand much more of each other. I can tell you today at the RCAF, you have a whole bunch of uniformed officers that are talking to vendors, and do you know who wasn't in the room? You. You know who they talked about? You. So, we have to realize all of you are part of you, and it's important that we seize the moment, because war is a terrible word to say but let's say we are in a new competitive paradigm. Maybe that's a nicer way to say it.

Taki Sarantakis: That's a nice euphemism.

Tom Jenkins: Nice euphemism. But the reality is, it's all about, this is a real-world problem now and we have to rise to the occasion. We've done it before but this is right now. You are the generation that has do it, not somebody else. It's your generation.

Taki Sarantakis: Last word to you, Major-General.

Major-General (Ret.) David Fraser: I talked about forcing functions. And in Afghanistan, well, the decade we were there, we had a forcing function, protect men and women and do the right thing for a nation that was at risk. This is not peace anymore. There's a new forcing function, and I see it everywhere I go, travelling internationally and whatnot. And going back to Tom's point, a lot of people are coming back to Canada. They're coming back to Canada and they're saying, you don't have baggage that a lot of nations do, we need you to tell us and to share with us, how can we do things? And we have a value prop that we probably don't quite understand as a nation of just how important our value prop of that Canadian flag, what it means today and the opportunities that are now being presented to us, but there's a forcing function behind it. It's very complex. Tom talked about it out there. There's a whole bunch of factors coming at us right now. It's a tsunami wave. You don't have a choice, all right? We've got to become the leading edge to do what is right for our nation. But at the same time, it's opening up an awful lot of opportunities internationally for this nation. I cannot believe the number of people talking to us now.

Taki Sarantakis: Please join me in thanking Major-General David Fraser and Mr. Tom Jenkins.

(Applause)

[01:09:33 The CSPS logo appears on screen.]

[01:09:38 The Government of Canada logo appears on screen.]