Andrea Schnepf, founder of NEPF, joins host Michael Bernzweig to break down why most enterprise AI pilots fail and what the successful 5 percent do differently. She explains how tech-first, reactive decisions waste budget, stall adoption, and erode trust, then lays out three pillars for making AI a real business capability: clear purpose, leadership alignment, and a robust operating model. Andrea also shares examples from companies like Microsoft and Intuit, and walks through a practical three-phase roadmap—foundation, pilot-and-learn, and scale-and-embed—so executives can link AI to strategy, governance, and day-to-day work instead of scattered tools.
Andrea Schnepf, founder of NEPF, joins host Michael Bernzweig to break down why most enterprise AI pilots fail and what the successful 5 percent do differently. She explains how tech-first, reactive decisions waste budget, stall adoption, and erode trust, then lays out three pillars for making AI a real business capability: clear purpose, leadership alignment, and a robust operating model. Andrea also shares examples from companies like Microsoft and Intuit, and walks through a practical three-phase roadmap—foundation, pilot-and-learn, and scale-and-embed—so executives can link AI to strategy, governance, and day-to-day work instead of scattered tools.
Key Topics:
Timestamps:
00:00 - Introduction: Why most AI pilots fail and how to avoid common mistakes
02:30 - Andrea’s framework for building AI foundation centered on purpose and vision
06:45 - Signs that organizations are approaching AI as a tech first, disconnected from strategy
10:15 - The role of leadership alignment in AI success — shared narrative and trust-building
14:00 - Critical components of an effective AI operating model — decision rights, governance, metrics
19:30 - Examples of structured AI cultures at Microsoft and Justworks
24:10 - Practical roadmap: from foundation to pilot, to scaling and embedding AI
28:50 - Managing organizational change, culture, and talent in AI transformation
34:20 - How to identify topic gaps using AI models to enhance content authority
40:50 - Capacity-based pricing: managing demand and scarcity in consulting and tech services
45:00 - Final thoughts: actionable steps to embed AI into your business DNA
Michael Bernzweig (00:05.976)
Joining us today, we have Andrea Schnepf She is the founder of NEPF, a transformation and change management firm that helps executive leaders at large enterprises deliver complex initiatives, including AI digital transformations and multi-billion dollar post merger integrations. So with that, Andrea, welcome to the event.
Andrea Schnepf (00:31.671)
Thank you so much for having me, Michael. Really appreciate it. The topic I'll talk through today is top of many leaders' minds, which is about making the AI vision real and actually turning strategy into a system for impact. AI is currently creating intense excitement and pressure. Boards want progress, competitors are experimenting, and employees just want direction. Yet many organizations are stuck.
between ambition and action. A recent survey found that 94 % of executives are unsatisfied with their current AI solutions, and that's signaling that even with heavy investment, AI isn't yet delivering what leaders actually hoped for. 64 % of CEOs say that the risk of falling behind drives them to invest in some technologies before they have a clear understanding of what the value they bring to the organization.
actually is. And that's a really defensive posture. Implement something now and figure out why you're actually doing it later. And this reactive behavior is making organizations forgo linking AI to actual value. And leaders notice that this is risky. Only 37 % of leaders say that it's better to be fast and wrong than right and slow, down when it comes to technology adoption.
So there's a real tension, the pressure to move quickly and the awareness that missteps can erode trust and waste resources. That's why this discussion is about getting clear on that purpose and building the system around it so that AI becomes part of how you run the business, not just another race that you want to join. Now, AI we know is reshaping key functions and is driving a huge amount of opportunities and that urgency to act is clear.
But let's just explore quickly why so many organizations are struggling. By 2027, Gartner actually predicts that more than 40 % of AI projects will be canceled, not because the algorithms don't work, but because organizations are leading with technology first and are treating the business case as an afterthought. Let's break down how these issues can play out. First, this leads to wasted investments.
Andrea Schnepf (02:58.669)
Organizations pour money into AI tools and platforms, but without alignment, readiness and integration into workflows, those investments go underused or fail outright, eroding ROI and trust in those future initiatives. Second is a cultural drag and employee frustrations. Because employees fear AI or don't see how it helps them, they resist and actually outright avoid using it.
turning that anxiety or lack of involvement into one of the biggest reasons that AI adoption can stall. Then the lack of coordination. When departments experiment with AI independently and without a centralized purpose, organizations just end up duplicating tools, burning resources, and slowing progress, turning AI into an expensive, disjointed effort rather than a strategic advantage.
And that also then leads to loss of competitive advantage. When AI isn't guided by a clear vision and integrated across the business, decisions stay slow, innovation lags, and fragmented low-impact adoption leaves more agile competitors to capture the real advantage.
MIT recently conducted a massive multi-method study that reviewed publicly disclosed AI initiatives, interviews with several large organizations, and survey responses from senior leaders at these organizations. Now, the results from that study is staggering. It found that an astonishing 95 % of AI pilots fail.
because organizations skip the critical bridge of connecting strategy and structure. 95%. So we ask ourselves, what do those 5 % of AI pilots that succeed actually get right? We found that those pilots consider three critical pieces to bridge business strategy to execution and scale. The first key pillar is a clear purpose and vision.
Andrea Schnepf (05:11.557)
The organizations that succeed know exactly why they're using AI and what outcome it's meant to deliver. That clarity keeps teams focused and prevents the scatter of chasing every possible use case. The second key pillar is leadership alignment. Successful organizations create a shared narrative at the top where leaders are consistent in the why, the priorities and how decisions will be made.
That alignment gives the organization direction and reduces confusion downstream. And finally, it's the operating model design. They'd be these companies built the structures needed to sustain AI, decision rights, roles, workflows, and governance. In other words, they designed a system so that AI isn't just a pilot, it becomes part of how the organization runs and how work actually gets done.
Now, the takeaway here is very simple. AI success isn't about moving fast. It's about building the foundation that connects ambition to execution that makes scale possible.
So let's look at each of those three pillars. AI can easily feel like something happening to us, not something that we're part of to employees. And when that happens, most employees are left asking a simple question, what's actually in it for me? So before people can rally around AI, the leadership needs to be aligned first. That shared vision of the leadership team becomes the anchor of AI adoption.
when leaders clearly explain, first of all, why AI matters for our business, how AI will enable and support our strategy and goals, and how it will positively impact employees' work. Second, how it will integrate into everyday workflows. What will actually change in people's day-to-day work and how AI will automate routine tasks so that employees can focus on higher impact, more meaningful work.
Andrea Schnepf (07:23.929)
And then thirdly, how it will help people be more productive and effective. The concrete impact AI will have on outcomes and success, sharing key examples of what successful AI will look like in the organization. The key takeaway here is without this clarity outlined here, even the most sophisticated tools will just stall in confusion or resistance.
So let's now turn to how we actually transform that vision into genuine belief and direction. A vision only matters if people believe in it. And belief comes from when that vision is modeled by the leaders. The goal is to make AI adoption feel like a shared journey, not a top-down initiatives. Leadership at all levels of the organization play a really important part here. Starting at the very top.
The executive team needs to just really align on a story and the priorities of AI, followed by leaders and managers communicating this vision to their teams. And finally, employees understanding what the change means for them. Now I want to give you two examples. First of all, Microsoft. Their AI transformation began with a shared vision.
Empower every employee to achieve more with AI. So let's just listen to that again. A shared vision, empower every employee to achieve more with AI. Leaders were able to model this message and they did it and they framed AI as a tool to enhance human creativity and productivity. This was backed by really strong training, ethical guidelines and transparency.
And this approach led to company-wide buy-in and seamless adoptions of tools like Copilot that, of course, Microsoft is really championing. The second example here is Justworks. They took a more bottom-up approach through AI hives, which were small, cross-functional groups that are testing real AI use cases. By focusing on education and empowerment instead of mandates, AI became a company-wide movement
Andrea Schnepf (09:44.667)
practical, human-centred and aligned with that daily work, which is just so important.
Andrea Schnepf (09:53.625)
Now, vision is essential, but vision alone doesn't scale. To turn that direction into real results, real business results, organizations need a system, which is a defined operating model that makes AI executable. The operating model is ultimately the bridge between ambition and actual action. It turns the why of AI into the how of getting work done differently
across the organization.
A strong AI operating model establishes structure, governance, and decision clarity so that teams know, first of all, how decisions get made. Who is in charge of making key decisions? Two, who owns what? What parts of the AI is owned by whom? And three, how progress is measured. What does success actually look like in this implementation? Without having these systems in place, AI efforts become
isolated pilots and whilst that is better than nothing, it just becomes scattered across departments with no shared priorities, no accountability and no path to scale. Now, a mature AI, a mature operating model turns AI from a set of experiments into a really repeatable enterprise wide capability. It aligns strategy with
day-to-day execution, ensuring that AI connects across functions rather than operating in silos.
Andrea Schnepf (11:35.281)
So organizations that build AI into their operating model scale implementations three times faster, capture over four times more value and reach production in about one third of the time compared with organizations that do not integrate AI into their operating model. So how do you do that? A defined AI operating model gives structure to the transformation.
It clarifies how AI initiatives are prioritized, how decisions are made, and how success is measured against real business objectives. Now, to do that, you need to consider five key things. First is the strategic orchestration. How does AI integrate into your business priorities, budgets, and KPIs? That needs to be really clear because if it's not, you're just aimless.
Two is the decision architecture. And that means establishing a clear governance model on who decides what, when and how. And this needs to be really understood by everyone. Three, lifecycle gates. Review ideas for AI at each stage, not just at pilot and continuously assessing for risk, but also for opportunities. Role systemization. Who owns what?
understand who is responsible for different aspects of AI and how it integrates. And then last but not least, are the really important integration enablers. So making sure that there are tools in place to create a smooth adoption of AI and push to scale. This may include a particular upscaling of current employees. So some successful examples include a lot of discipline around this.
And we mentioned Microsoft earlier, and they actually embedded AI metrics into regular business reviews so that AI became part of how performance is tracked, not just a side conversation or a side project that select individuals only are part of. Another example is Intuit. They created cross-functional AI councils around product, data, and risk, and they cut approval times by 40%.
Andrea Schnepf (13:59.596)
showing how structure can actually increase speed and trust. And that's a very important point because looking at this kind of structure can often look like we're slowing down progress, but the opposite is true. We are actually with this structure in place, accelerating progress and opening up for AI to provide a huge and immense amount of business value. So redesigning your operating model
For AI, it isn't just structure. It's really how you build the capability to prioritize, adopt and scale to sustain competitive advantage.
So let's bring this all together and look at a practical-faced roadmap for sustainable AI adoption across your organization. Now, this approach is designed to give you the structure you need to scale AI in your organization whilst maintaining flexibility to respond to challenges as they arise. Now, this roadmap that we've seen over and over again work very well unfolds in three key phases.
First one is really this foundation that we've touched on in this session. And this is what we've covered on. This is where we define a really compelling vision so that everyone is aligned on why now, what will change, and what outcomes that we're driving. This is where we redesign critical operating model elements such as roles and workflows so that AI is embedded in how work runs and scales beyond pilots. We align leadership and build the case for change
and then engage stakeholders continuously. After this, we move into pilot and learn. We move into action and we build momentum, generate early wins and ensure the AI tools are genuinely delivering values. And then the third phase really is where we scale and embed. This phase ensures that AI is integrated in your organization's DNA not as a one-time project, but as a continuous scalable shift.