“The technology is good enough. The real hurdle now is people, fear, and change management,” says Shannon Bell, CIO, OpenText

March 19, 2026

Shannon Bell, EVP, Chief Digital Officer and Chief Information Officer at OpenText, shares how “information first” thinking, simplicity, and agentic AI are reshaping how large enterprises work. She explains why most enterprises don’t have an AI problem but an information problem, how to go slow to go fast with AI, and how a blended workforce of humans and AI agents turns scarce skills and fragmented processes into scalable value.

Crucially, she digs into change management and the future of jobs: why fear of displacement is often higher than the reality, how to position AI as a copilot rather than a competitor, and what it means to give 22,000+ employees an AI development goal so they can actively shape how their roles evolve. She also shows where agentic AI is ready now, such as search and summarize, root cause analysis, and software delivery, and why success depends on clear roles, governed data, and using HR and SRE teams as early champions to build an “AI fabric” across the enterprise.

• What does it really take to make AI an assistant, not a threat, for your workforce?
• How can you start small on messy, real-world systems and still build toward an AI ready data estate?
• Which foundations, guardrails, and operating model let you decentralize AI innovation without losing control?

Watch the whole interview:

Read the podcast as an interview:

(The interview was shortened and edited using ChatGPT)

Mark Kohout: Most enterprises don’t have an AI problem, they have an information problem. Today, we explore how strong information governance turns AI from experiments into business outcomes, and how AI agents can become part of the workforce.

Joining me is Shannon Bell, Executive Vice President, Chief Digital Officer, and Chief Information Officer at OpenText, where she oversees IT and digital systems, data platforms, networks and communications, cloud operations, security, and compliance. With over 25 years of leading transformations across Canada, the US, and Europe, she brings deep experience in integrations, product innovation, and digital strategy, and is passionate about data‑driven innovation and business growth.

Shannon, welcome. Delighted to have you here today.

Shannon Bell: Thank you for having me.

Mark Kohout: And for braving the cold Toronto day. Let’s start with leadership. You’ve led transformation at some of Canada’s largest technology organizations. Which experience most shaped your approach to leading digital innovation and scaling it?

Shannon Bell: It wasn’t a single experience, but a series of technology evolutions. Early in my career, I worked in networks just as IP technology was being introduced and helped deploy IP VPNs in Europe. It felt very new and daunting, but I learned that even “scary” technology is still just technology. If you stay focused on the business problem and the outcomes you need to deliver, you can work through it.

That ingrained two things for me: no problem is unsolvable, and technology is an enabler of change, not the end goal. As we moved from waterfall to agile and into early digital  (self‑service, online banking, mobile apps) each wave reinforced the same lesson: don’t be afraid of new technology. Be curious, learn it, and always ask, “How do I use this to drive a real business outcome?” Projects fail when they use technology for technology’s sake.

That mindset led directly to how I work today: stay outcome‑driven, stay curious, and treat technology as a tool, not a destination.

Mark Kohout: So, a bit of “keep calm and carry on”: stay focused on the business and fundamentals. Don’t get fixated on the tech, but ask, “How can I change the business using this technology?”

Shannon Bell: Exactly. And also, stay curious, but don’t define yourself by a specific technology.

At one point, I saw myself as a pure network expert and almost passed on a move into software because I was too attached to what I knew. I’m very grateful I was pushed, because I learned that in software almost everything is solvable, it’s more forgiving than networks, and it changes your view of the value you can deliver to customers. It taught me to stay open to learning new domains so you can keep evolving with technology.

Mark Kohout: And ultimately evolve the business. Beyond curiosity and not being defined by a stack, what other principles have stayed with you across acquisitions, market changes, and large‑scale integrations? Any axioms you apply when you look at what the company needs to do, even beyond technology?

Shannon Bell: Simplicity wins the day.

There’s a tendency to overcomplicate, either because people don’t know enough or because they know too much. If you can translate between business and technology, and simplify in both directions, you build trust and get better outcomes.

In tech, we often get lost in acronyms and the power of the technology, and we lose our stakeholders. Relationships are built on clear communication: explaining what the technology can and can’t do in terms that make sense, and distilling business problems down to their essence.

In the era of AI, everyone wants to solve the most complex problems. In reality, the better strategy is to start with simple problems. They feel basic, but that’s where you build the foundation and get quick, tangible wins. For me, simplicity in how you describe technology and frame problems has been a constant pillar.

Mark Kohout: So,simplicity is elegant and something people can rally around.

Shannon Bell: Absolutely.

Mark Kohout: It makes me think of governance, which can become an acronym salad. If you can’t focus on the “why” and what you’re really solving for, you’re working uphill.

Shannon Bell: Exactly.

Mark Kohout: Let’s shift from business management to the technology powering the transformations you’re driving at OpenText: agentic AI. You’ve spoken about a blended workforce of human talent and AI agents. What does that look like in practice, and how should executives think about it?

Shannon Bell: Agentic AI is a huge opportunity, but it only works if the foundations are there. You need good business processes and strong information management. Without that, you won’t get meaningful outcomes.

Assuming those foundations exist, our approach has been “go slow to go fast.” The technology can move quickly, but if you apply AI to a poorly understood process, you’ll get poor results. We looked across our business processes and asked which ones lend themselves to agentic AI and where to start. We chose low‑hanging fruit, like our IT help desk, where we had well‑documented processes and direct ownership.

The first phase wasn’t AI at all; it was putting the right systems and tools in place to manage those processes, which reduced about 30% of L1 requests. Then we layered AI on top in a very targeted way: clear roles for agents, clear roles for humans, and agents matched to specific tasks.

Early on, we made agents too complex and ran into edge cases and failures. We pivoted to very discrete, task‑based agents. You can orchestrate those simple agents into richer flows and reuse them in different scenarios. We also map roles side by side: what the agent does and what the human does, especially as we move into more sophisticated work.

Our hardest role to hire for is Site Reliability Engineer. SREs are critical and scarce. Their leader said, “I need my SREs on the highest‑value work; some of what they do can be handled by agentic AI.” We defined exactly what agents would do and what SREs would retain. SREs still make decisions, change cloud configurations, and drive architecture, sizing, and scaling. Agents support them by handling repetitive analysis and preparation. It increased their capacity and helped address a talent bottleneck, a great example of human–agent synergy, not replacement.

Mark Kohout: We’re back to simplicity and clarity about business architecture. If you can’t articulate what’s being done, you probably don’t understand it well enough to automate it.

You also mentioned “twinning” SREs with assistants, which brings us to people. How did SREs and help desk staff feel? How does deploying an agentic workforce affect employee experience, and how do you drive culture change and motivation?

Shannon Bell: The human response isn’t uniform. People with more technology experience are generally less fearful and more open. SREs quickly saw the benefit: offloading repetitive tasks and focusing on higher‑value work.

Across the broader organization, there’s still fear. The hype about what AI can and can’t do, and which roles it might displace, creates anxiety. We frame AI as an assistant that changes how you work, not something that replaces you, but there’s still apprehension.

To address this, we gave every one of our roughly 22,000 employees an AI development goal, regardless of role. The goal is to raise the baseline of AI literacy, demystify it, and help people see practical potential. We invested in training, tools, examples, and tailored change management. Two years ago, I would have said our biggest hurdle would be technology rollout. Now I’d say the technology is “good enough” for many use cases; the primary hurdle is people and change management.

Mark Kohout: Is AI different from other technologies you’ve worked with in terms of how people react?

Shannon Bell: It’s a bit different, but not completely.

We’ve gone through waves like email, the internet, mobile, and 3G/4G/5G. There was always hesitation and hype. Coming from telecom, I remember the early promise of 5G and constant questions about the “killer app.” Very similar to AI today.

What’s different now is concern about displacement: which roles will change, how fast, and what that means for education and skills. Are we training students for the right roles? Are we adapting curricula for the future workforce? That uncertainty creates more apprehension, and we don’t yet fully understand the long‑term potential of AI, just as we didn’t with earlier technologies.

Mark Kohout: Let’s go back to “go slow to go fast.” Would you consider OpenText an early adopter of agentic AI, and what was really behind that philosophy?

Shannon Bell: There were two tracks.

First, as a technology company, we had to bring AI into our products for customers. That had clear market urgency. Second, we had our internal adoption of AI for our own employees, where we deliberately moved slower so we didn’t get ahead of our foundations.

OpenText has grown through many M&As. Harmonizing processes, systems, and tools is a constant challenge. We kept asking: do we have the right foundation and architecture to build AI on? AI on bad data and poor processes gives bad outcomes.

We balanced employee demand for AI tools with governance, security, and readiness. We learned that simply releasing tools without change management leads to low adoption and frustration. So we shifted to phased, cohort‑based rollouts with concrete examples. That’s slower at first but higher impact.

We also prioritized use cases with clear business value, especially innovation and revenue, not just cost cutting, and paid attention to organizational readiness. HR was one of our earliest and most strategic partners because they are crucial for change management. If HR understands and benefits from AI, they can champion adoption across the company.

“Go slow to go fast” means build foundations, pick the right partners, be selective about use cases, and accept some managed failures, but avoid uncontrolled, organization‑wide failures.

Mark Kohout: Based on what you’ve learned, where will agentic AI drive measurable value for large enterprises over the next 12–24 months?

Shannon Bell: There are clear areas where it already delivers strong value.

First, search and summarize: mining large volumes of content and surfacing key points. Second, data analysis: AI is excellent at heavy‑lift analysis and pattern recognition, with humans interpreting and applying results. Third, operations: in network and cybersecurity operations, AI helps with root‑cause analysis and anomaly detection, finding the “needle in the haystack.”

We’re still cautious about full autonomy. We generally keep humans in the loop for decisions and orchestration, though more autonomous flows will emerge.

Fourth, software delivery: AI helps with technical writing, more consistent user stories, analytics‑driven product planning, and auto‑generated test cases, plus code assistants for developers. That improves productivity and quality.

The biggest value comes when business owners and domain experts become champions. AI doesn’t eliminate domain expertise; it amplifies it. We support that with a hub‑and‑spoke model: a central Center of Excellence for governance, standards, and tooling, and empowered business teams where there’s maturity.

Mark Kohout: Stretching a metaphor, you’re building an “AI fabric” across the enterprise, like a data fabric.

Shannon Bell: That’s a good way to put it.

Mark Kohout: Before we close, let’s touch on foundations. How necessary is strongly managed information for good AI outcomes?

Shannon Bell: It’s fundamental.

AI runs on data. If your data isn’t well‑governed, high‑quality, and accessible, your AI outcomes will be unreliable. You need a strategy for structured and unstructured data: how it’s stored, governed, secured, and made available. Good information management lets you trust AI‑generated insights enough to use them in decision‑making. Organizations that have invested in this can move faster and with more confidence.

Mark Kohout: Many organizations have fragmented legacy systems and data. How can they move toward an AI‑ready data estate as quickly and seamlessly as possible?

Shannon Bell: The most important thing is to start.

Looking at a complex legacy landscape can be overwhelming and lead to “big bang” thinking: “We’ll build a perfect data lake, then do AI.” Those projects take years. Instead, pick a specific area: a business problem, a data domain, a use case. Understand the data, clean and structure what you need, and build AI on that subset to drive a clear outcome.

Deliver a small win, then expand. Over time you modernize and rationalize your data landscape through focused steps, not one massive program. Pragmatism and incrementalism beat perfectionism. Keep it simple, solve one real problem at a time, and let those successes fund and justify the next steps.

Mark Kohout: Meet the world as it is, not as we wish it were.

Shannon Bell: Exactly.

Mark Kohout: Lastly, how is AI transforming the role of the CIO?

Shannon Bell: It’s transforming it quite radically.

The core of the job is now partnership with the business. One of my closest partners is HR, which wouldn’t have been true years ago. I see my role as a technology partner to business leaders, helping them achieve outcomes like revenue growth, customer experience, efficiency, and risk reduction, using technology as a lever.

That requires a deep understanding of how each part of the business works. It’s less about managing a software factory and more about understanding the whole organization end‑to‑end. It’s more multidisciplinary and more fulfilling. My team understands the impact they have on the business. Earlier in my career, I knew a lot of technical detail but couldn’t always link it to business value. Today, that link is the starting point.

Mark Kohout: So the CIO is now a multidisciplinary business leader, not just a technologist.

Shannon Bell: Absolutely.

Mark Kohout: One last question. If there’s one misconception about AI you wish executives understood better, what would it be?

Shannon Bell: That speed is not the primary measure of success.

There’s a perceived “race to AI.” Boards ask, “How fast are we going? Are we ahead or behind?” That can push organizations to move too quickly, with weak foundations, and fail. I’d rather be asked about our foundations – governance, data quality, security, talent enablement, readiness – than how many AI tools we’ve deployed.

The goal isn’t speed for its own sake; it’s to move deliberately and deliver durable, high‑quality outcomes. Work smarter, not just harder or faster.

Mark Kohout: A clear invitation to work smarter, not harder. Shannon, thank you for taking the time to speak with me and for sharing your experience leading large‑scale transformations.

And to our audience: thank you for joining us. If you found this discussion compelling, please consider liking and subscribing. See you next time on the Adastra Podcast.

Shannon Bell: Thank you.

Listen to the podcast: