“Technology Doesn’t Matter As Much as the Outcome,” Says Amol Shah, Microsoft

July 21, 2026

Amol Shah, General Manager, Specialist Sales at Microsoft Canada, shares how enterprises can move beyond AI experimentation to real business transformation by getting the fundamentals right: unified data, governance by design, and an intelligence layer that makes agents as trusted and productive as employees. Drawing on nearly two decades at Microsoft, he explains why data has shifted overnight from an IT concern to a C‑suite strategic asset, why companies playing it safe with AI will spend tomorrow catching up, and how the role of partners is evolving from plumbing and migration toward industry IP, adoption, and change management. 

The episode will answer: 

• How do you move past endless AI pilots and build the unified data foundation, governance, and intelligence layer that lets agents and employees work from the same trusted context? 
• What changes when data stops being an IT project and becomes a C‑suite strategic asset, and how do leaders create the urgency to avoid falling behind on AI the way some fell behind on cloud? 
• How should companies rethink their partner ecosystem as value shifts from migration and plumbing toward industry IP, adoption, and change management at scale? 

Watch the interview: 

  

Read the podcast as an interview: 

(The interview was shortened and edited using ChatGPT) 

Mark Kohout: Hello and welcome to this Adastra podcast. My name is Mark Kohout, and I lead the AI and Data Governance practice at Adastra North America. Adastra is a global AI, data, and cloud systems integrator, and we're coming to you today from our Toronto studio. 

Joining us is Amol Shah, General Manager, Specialist Sales at Microsoft. Amol brings nearly two decades of experience at Microsoft, with leadership roles spanning Azure Data and AI, go-to-market strategy, cloud enablement, and specialist sales. Most recently, he served as General Manager for Americas Data and AI Go-to-Market, where he was responsible for leading go-to-market strategy for Azure Data and AI across the entire region. Today, Amol leads Specialist Sales at Microsoft, bringing a strong executive perspective on how organizations can connect cloud, data, and AI investments and turn them into winning business outcomes. 

We're delighted to have you here. Welcome to the podcast. How are you? 

Amol Shah: I'm great. It's great to be here. 

Mark Kohout: We've got a great Toronto day for it, nice and warm. I hope that's inspired some thoughts you can share with our audience. Let's start with your background. You've had a long career at Microsoft across sales, go-to-market, and cloud leadership. Tell us a bit about that journey and how it's shaped the way you think about data and AI today. 

Amol Shah: It's a great place to start. I've been at Microsoft for approaching 20 years across various functions. I actually started in finance and then moved through marketing, product marketing, and partner sales. As you mentioned, my previous role was leading go-to-market for Data and AI in the Americas. Most recently, I've taken on leading our Enterprise Solution Sales team in Canada, managing about 320 customers across the country. 

I've been fortunate to see things from many different angles, from the pure numbers perspective in finance all the way to getting out of the single deal and figuring out how to build a motion and an engine through go-to-market, leveraging our partner ecosystem. 

Three things I've taken away. First, the technology doesn't matter as much as the outcome it's serving. 

Mark Kohout: Right, and sometimes more than we think it matters. 

Amol Shah: Exactly. Early on, you get fixated on products, products, products, but really, it's about the outcome. Second, go-to-market is a team sport. 

Mark Kohout: It is. 

Amol Shah: It's bigger than I ever thought. When you're selling to thousands of customers worldwide, you need to leverage a partner ecosystem. 

Mark Kohout: To get that multiplier effect. 

Amol Shah: Exactly. Our partners have niche skills, industry knowledge, and expertise we need to scale. But there's also enablement. Your field is only as good as how well they're enabled. So the idea of "one team, one go-to-market" is critical. 

The last thing: I've gone through multiple shifts at Microsoft. I was there when we launched cloud, and AI is going to be no different. The customers that play it safe are the ones catching up. I saw it with cloud, and I'm starting to see it with AI. So when I talk to customers, it's about how to drive urgency, scale, and intentionality. 

Mark Kohout: Very interesting. Let's drill down on that Microsoft perspective. What you made me think about is the underlying DNA of Microsoft. The technology changes. We've seen the same at Adastra, from data warehousing through distributed computing, and then the world changed with cloud. You've had a front-row seat to the evolution of Microsoft's cloud, data, and AI perspective. What would you say has changed most in how enterprises think about data and AI today compared to even two or three years ago? 

Amol Shah: A few things. First, it started as a side project. "What can this technology do?" 

Mark Kohout: Skunkworks. 

Amol Shah: Yes. Now it's, "How do I think about my business differently?" We've moved away from experiments and made it real. 

Second, the world of data has changed. It used to be siloed within IT, which managed the guardrails. Now it's at the CEO level, at the line-of-business level. Almost overnight, data has become a strategic asset. 

Mark Kohout: A differentiator. 

Amol Shah: A differentiator. Everyone cares about it now. Something as simple as semantic models didn't matter before. Now it matters, because data is the asset. 

Mark Kohout: Context is key. 

Amol Shah: Exactly. From an agent perspective, that's been a big change over the last few quarters. IDC, Gartner, and Capgemini all cite stats like 1.3 billion agents by 2028; 82% of C-level executives expect to integrate agents in the next couple of years; 40% of companies expect agents integrated into task-specific workflows. Again, it's not an IT-focused play. It's line-of-business. Everyone needs to care. 

Mark Kohout: So a multidisciplinary, cross-functional perspective needs to be brought to the table. 

Amol Shah: Yes. The last thing I'd say is the change in speed of execution. Two or three years ago, it was, "Okay, I'll run this, and hopefully in a year, maybe six months, I can get it into production." That's collapsed considerably. The question now is, "How do I go from demo to production really quickly?" That year has shrunk to months, weeks, sometimes days. 

Mark Kohout: And is Microsoft's technology driving that? 

Amol Shah: It is. Whether you're building through Foundry, leveraging Copilot and agents, or using Fabric as your unified data layer… 

Mark Kohout: The agentic control plane as well, Agent 365. 

Amol Shah: Exactly. We're building our products and platforms to address the challenges customers have with scale. You mentioned Agent 365. Observability is probably the biggest issue right now in scaling agents. We've built that into everything we're doing. We can talk more about how you build those foundational layers of intelligence, trust, and so on. 

Mark Kohout: Let me be a fly on the wall for a second. You said data increasingly has a seat in the C-suite, being elevated on the corporate agenda, seen as a strategic asset. But that's a mindset shift for companies that have seen data and AI as technology initiatives. How do you catalyze their thinking about data and AI as a business transformation agenda? 

Amol Shah: Great question. Very top of mind, and an area Microsoft is investing heavily in. It's an exciting time for data. 

Mark Kohout: Mind-boggling, too. 

Amol Shah: Mind-boggling. The amount of innovation is amazing, but it's a double-edged sword. Customers end up confused and looking for clarity on how to manage data and leverage their estate for AI. 

Mark Kohout: It's a recurring theme in this podcast series. We've started talking about AI fatigue. 

Amol Shah: Exactly. I talk to customers about two things. First, most customers have 5, 10, 20 different data sources they're trying to stitch together in a secure, governed way. It's expensive and hard, so customers stall. "How do I clean my data?" becomes a two-year project, and you can't scale your AI ambition. 

But data is the fuel that transforms AI. The "garbage in, garbage out" analogy is still true. Your AI quality is always capped by your data quality. So when customers are stalling, I tell them: start at a foundational level by unifying your data platform. At Microsoft, we do that through Fabric. It's one of the most important areas we've invested in, a unified data platform delivered in a secure, governed, and cost-effective way. Cost matters, because data costs can spiral out of control. 

The second piece is agents. We talked about billions of agents being integrated into the workforce. That's a massive structural change. How does a customer make their agents as trusted and productive as their employees? 

Mark Kohout: You need to treat them almost as complexly as you'd treat employees. 

Amol Shah: Exactly. But agents need context to do that. So we talk about the intelligence layer: how do you give employees and agents organizational context, business context, and institutional knowledge to make decisions? In our world, we call that Microsoft IQ. Work IQ is the organizational context, Fabric IQ is the business data context, and Foundry IQ is the institutional knowledge. 

This matters because agents, employees, and AI aren't working in silos. They're a workflow within a company. Large retailers, oil and gas companies, financial institutions. All these areas are connected, and the data and people need to be connected too, in a secure, governed way. Whether you call it Customer IQ or something else, that intelligence layer makes a huge difference in scaling and getting value from data. 

Mark Kohout: As a data governance guy, it feels like an extension of metadata management. Building a common language around the organization that impacts people as well as agents. 

Amol, take me back to something you said earlier about shortening time to market, collapsing that six-month production cycle for AI models and agents. You talked about the technology foundations: IQ, Agent 365, unification of data through Fabric. For organizations listening today who want to accelerate value from their data, what's the single most important foundational step they need to get right first? 

Amol Shah: Great question. "Get your data foundation right" is easier said than done. "Fix your data" is easier said than done. It's almost impossible to do all at once, which is why you run parallel motions. 

I always start with trust. You mentioned governance earlier. It's the one area that stalls everything. Build governance and security early. 

Mark Kohout: Right. 

Amol Shah: We've all experienced files sent to the wrong person, or an employee having access to something they shouldn't. As people build more agents, this will become more of an issue. Governance, observability, security, permissions, all of that needs to be embedded early in the cycle, not later. 

Mark Kohout: It's a bit of a cliché, but we can think of it as governance by design. 

Amol Shah: Exactly. A lot of the challenges we saw early on came from doing it later. 

Mark Kohout: And that's what stalls the POC-to-production transition. Suddenly the organization is dealing with regulatory, compliance, privacy, and security checklists all at once. 

Amol Shah: Anything regulated becomes non-negotiable. And you're spending a lot of money for this stuff. You don't want to buy something and then turn off functionality because of security issues. You want full value out of what you're paying for, and if it's not governed correctly, that will stall it. 

Mark Kohout: And training models on inappropriate or inaccurate data will cost you so many tokens it could be a career-ending move. 

Amol Shah: We're moving to a consumption world, so this is a real consideration. The only way to handle it is to get the foundational elements right. 

Mark Kohout: What I'm taking away is parallel efforts, building governance as you go, focusing on subsets of data because otherwise time to market is too slow. 

For leaders who have adopted AI at scale (Copilots, custom AI solutions, agentic AI), what advice are you giving customers trying to drive meaningful business impact? 

Amol Shah: At Microsoft, we're fortunate to be Customer Zero on many things, so we've seen this play out at home. 

Mark Kohout: You've tried it. 

Amol Shah: Exactly. A few lessons. First, we're past the "let's do 20 pilots" stage. Pick one or two things and make them real, with numbers behind them and clear outcomes. 

Second, design for adoption and change management from the start. That's often missed, and then customers wonder why employees aren't using it. You have to build that into the design from the beginning. 

Third, care about the boring stuff: measurement, governance, ROI, the feedback loop from customers and employees. It all matters. 

Mark Kohout: The devil's in the details. 

Amol Shah: Exactly. It has to be part of the design of any initiative you launch. 

Mark Kohout: So companies that could get by with lip service in the past won't be able to in this new paradigm. Have you seen patterns in how AI is being adopted? Is it something people do once and have it, or is it more of a ladder? 

Amol Shah: It's a bit of a roller coaster. From a maturity perspective, I'd break it into a few stages. First is experimentation and POCs, right-sizing to see how far you can go. Stage two is realizing the value and starting to build a concise business case, along with the capabilities. Stage three is scaling. That's where governance really kicks in, and I see many customers stuck there. 

What I don't see enough of is stage four: real transformation. How is your business actually changing because of what you've invested in? That's the area we need to work on globally. Are you really changing business processes? Are you innovating your products and services? 

Mark Kohout: Are you rethinking your value chain? 

Amol Shah: Are you rethinking everything end to end? When I gave that earlier example about data, it was: how is a brand manager connected to supply chain? Have you really thought about how those two interact? In most cases, we're not there yet. We're making progress, but we're not fully there. That's what we want to help customers accelerate. 

Mark Kohout: So you're really rethinking upstream and downstream dependencies, how one part of the business influences another, to find opportunities for innovation. It sounds like you really need to understand your business processes in depth. 

Amol Shah: Exactly. That's why the intelligence layer is so important. For example, a supply chain agent understanding the institutional knowledge of contracts, and a marketing agent taking that context to make a decision. It's all connected. It's hard to put together, but you need to think about it early. That's why intelligence and trust are the most foundational elements customers need to consider. When you get that right, everything scales more quickly. 

Mark Kohout: We're almost out of time, but before we close, I'd like to draw you out on the partner ecosystem. As companies rethink and scale the deployment of the agentic workforce, how do you see the role of partners evolving? Where can partners create the most value right now? 

Amol Shah: It's funny, Mark. I often hear that partners will play less of a role going forward. I don't buy that at all. I actually think they're more important than ever. The value has just shifted. 

For the last decade or so, it's been about the plumbing: migrate and modernize, stand up the platform, and go. Where we need partners to go now is more around IP, strategy, and adoption. We need partners with deeper industry-level knowledge. How are you building IP across oil and gas, financial services, public sector? Working with customers to identify how they can run their business differently, we need partners to help scale that. I'm glad we get to work with Adastra every day, because you've done a fantastic job with that. 

The second piece is adoption and change management. We need partners to help there. You can't just build something and say, "You're on your own." It has to be by design. Partners need to ensure what they're building is not only built well but actually used by customers and employees. Those are the two areas where we'll get a ton of value from partners going forward. The ecosystem is as important as ever. It's just shifted. Don't get me wrong, the plumbing and migration work is still important. 

Mark Kohout: Still important. 

Amol Shah: There's a massive addressable market and huge opportunity there, and it still needs to be done. 

Mark Kohout: But it's also about envisioning the new target state, what the business is going to look like. 

Amol Shah: Exactly. 

Mark Kohout: Amol, a fascinating and insightful conversation. What I'm taking away is that successful AI adoption isn't just about technology. It's about strong foundations, intentionality in design, a cross-functional and C-suite perspective on how AI transformation reaches throughout the business, and the need to get the right foundations in place without stalling. Thank you very much for this discussion, and for taking the time out of a busy agenda. We've been trying to get this in the books for a few months. 

Amol Shah: I appreciate it. This was great. Thank you again. 

Mark Kohout: Thank you. And to our audience, if you've enjoyed today's discussion, please like and subscribe to this podcast series for more insights on data and AI. So long for now, from Toronto.