What being Client Zero is teaching us:
- Access isn’t adoption, and adoption isn’t transformation. Giving employees tools is a starting point. Building fluency and changing how work gets done is where value begins.
- Activity isn’t the same as impact. Agent counts, usage metrics, and proofs of concept are useful indicators. But the real measure is business outcomes.
- The people closest to the work are often closest to the opportunity. Transformation scales when innovation becomes part of the organization’s culture, not just a centrally managed initiative.
- AI needs business context, not just data. Governed definitions, trusted information, and enterprise intelligence are foundational to making AI useful at scale.
- Economics and governance matter from the beginning. Responsible scaling requires intentional choices about cost, architecture, security, and oversight.
- Technology alone doesn’t transform a company. People, processes, decisions, accountability, and operating models have to evolve with it.
Giving 4,000 employees access to AI is a milestone. Getting 4,000 employees to fundamentally change how they work? That’s transformation.
At Presidio, we’ve learned that those are two very different challenges.
Like many companies, we started with ambitious goals for AI. But we’ve quickly discovered that some of the hardest questions aren’t even about the models or the tools. They’re about economics, trust, data, people, and ultimately, how you redesign work.
We’re helping clients navigate AI transformation by first doing the work ourselves, to learn what it takes to move from experimentation to meaningful, enterprise-wide change.
And while we’ve made tremendous progress, we’re also learning that transformation isn’t something you can buy off the shelf.
AI access is a milestone – fluency is the goal.
One of our early goals was to give 100% of our workforce access to enterprise-issued AI tools. We’ve achieved that. Today, Presidio employees have access to a combination of our own AI Studio platform, frontier models from Anthropic and OpenAI, and Microsoft Copilot Premium.
That’s an important foundation. But access is the easy part. The tougher challenge is what happens next.
Do employees understand which tools to use and when? Can they identify opportunities to improve their work? Are they comfortable experimenting? Do they understand the guardrails? And most importantly, is AI changing how they operate?
We’ve learned that you can’t scale AI adoption through a central team alone. It has to become part of the way people work, learn, and solve problems together.
That’s why we created our AI Leaders program, establishing advocates within workgroups across the company.
These leaders help their colleagues build practical skills, share what works, understand responsible-use guidelines, and identify opportunities where AI can make a difference.
Our Transformation Office can establish the framework, provide the tools, and create the guardrails. But the best ideas often come from the people closest to the work. For instance, no one understands the friction in a customer quoting process quite like the sales operations team working through it every day.
Our technical teams have now reached 100% adoption of AI in key workflows and achieved an independently assessed advanced stage of adoption.
That’s meaningful progress. But the distinction we’re making is important: having AI embedded in a software application isn’t the same as having employees who know how to use AI to rethink and improve the work itself.
That’s the kind of fluency we’re trying to build.
Innovation doesn’t have to start at the top.
One of the more exciting parts of our journey has been watching employees move from using AI tools to building solutions of their own.
We’ve established hackathon-style Build and Innovation Days across engineering and corporate functions, giving teams opportunities to turn ideas into working proofs of concept. The strongest concepts are recognized and shared through Exchange, our internal forum for showcasing innovation and spreading successful ideas across the company.
The approach is intentionally practical: find a problem, build something, test it in the flow of work, measure the impact, and share what works. Then do it again.
I believe one of the most important roles of a Transformation Office is to create the conditions for innovation to scale. That requires giving people room to experiment while also establishing a repeatable way to identify which ideas deserve broader investment.
Because innovation is valuable, but innovation that can be operationalized, measured, and scaled is where transformation begins.
3,200 agents is impressive, but it’s not the finish line.
Today, more than 1,700 Presidio employees have built agentic workflows, with approximately 3,200 active agents across the company.
Those numbers are exciting. They demonstrate that employees are moving beyond consuming AI functionality and beginning to build workflows that combine information, reasoning, and action.
But I’d argue that counting agents is a little like counting software licenses; an indication of activity, not necessarily impact.
The better questions: What work is changing? What friction are we removing? What decisions are getting better? And where are we creating measurable value?
That’s where our focus is shifting. We’re working to understand where agents are being used, which workflows they support, what outcomes they produce, and where additional controls or enablement are needed.
The value of an agent is in the work it improves, the decision it accelerates, or the friction it removes. And as agentic AI becomes more capable, the opportunity goes well beyond individual productivity. It starts to change how work moves across an organization.
AI economics get complicated fast.
One of our early lessons was that AI economics get complicated quickly.
What starts as a relatively straightforward conversation about licenses becomes a much more nuanced discussion about consumption, model selection, infrastructure, security, and the cost of scaling.
You then must ask yourself: Are we really getting value from all of this activity?
We determined that more usage isn’t automatically better usage. Our approach is built around a simple principle: right job, right tool, responsible use.
We don’t believe every problem requires the most sophisticated model, nor do we believe a single-model strategy is necessarily the right answer. Different tasks require different levels of capability, speed, cost, privacy, and control.
Presidio’s AI Studio provides flexibility to operate in an on-premises open-source environment while also accessing frontier models when the work calls for them. Shifting the right workloads to AI Studio has significantly decreased our token usage.
That flexibility allows us to make more intentional decisions about where AI should augment work, where it can automate processes, and where human judgment remains essential.
Developing an operating discipline around AI economics early in the process positions you to scale responsibly.
AI can’t transform a business it doesn’t understand.
One of the less glamorous, but arguably most important, parts of our journey is data. We learned early on that giving an AI model access to enterprise information isn’t the same as giving it an understanding of the business.
What does a booking mean? How do we define a customer? Which metrics matter? How do different systems interpret the same information?
These sound like simple questions until you try to answer them consistently across a large enterprise.
For years, we taught people how to understand the business. Now we have to teach technology how to understand the business. That’s a pretty profound shift.
It’s why I believe AI and business intelligence are converging. Historically, organizations have treated data management, BI, analytics, and AI as separate disciplines. But as AI becomes more deeply embedded in business decisions and workflows, those boundaries begin to disappear.
AI needs trusted data, governed business definitions, and semantic context to reason effectively about how a company operates. At Presidio, we’re bringing AI, BI, and master data closer together as parts of the same enterprise intelligence architecture.
For AI to help make better decisions, we need to give it the right context. That challenge is as much of an organizational one as it is a technical one.
Make the safe path the fast path.
Another lesson we’ve learned is that governance and adoption can’t be treated as competing priorities.
When employees have access to powerful AI tools, they’re going to experiment, which is exactly what we want them to do. But we also need to make sure they’re doing it responsibly.
That means clear guidance around data, security, model selection, oversight, and the circumstances under which AI systems can take action.
Our philosophy is simple: make the safe path the fast path.
If the approved tools and processes are too difficult to use, employees will naturally look for alternatives. So we had to build an environment where employees can move quickly and confidently within appropriate guardrails.
That becomes even more important as we move from AI that generates information to AI that can execute actions across business systems. Bolting on trust after deployment won’t work; it must be built into how enterprise AI operates.
The opportunity to redesign work.
Much of the first wave of enterprise AI has focused on individual productivity: helping people write faster, summarize information, analyze data, or complete tasks more efficiently.
The next level will change how work happens via a reimagining of workflows that cross teams, systems, and organizational boundaries.
At Presidio, our Transformation Office is increasingly focused on those broader opportunities, combining AI with process redesign to simplify handoffs, connect information, improve decisions, and remove unnecessary friction.
We’re looking at how workflows across the enterprise can be improved, not just how individual tasks can be made faster. Automating a broken process doesn’t fix the process. And making one department more efficient doesn’t necessarily improve the end-to-end customer experience.
What we can do now is connect the work in ways that weren’t previously possible. It also requires us to rethink traditional operating models.
How should decisions get made? Where should accountability sit? Which processes still require human intervention? How do we measure value when work spans multiple functions?
These transformation questions are central to our goal of helping Presidio grow faster, operate smarter, and scale as one company, with our clients and partners at the center of that work.
The final word.
We’re still early in what I believe will be a much larger evolution in how companies operate, but the lessons we’ve already learned are invaluable.
We’re proud of the progress we’ve made, but we’re also clear-eyed about how much work remains. Being Client Zero means we’re experiencing firsthand the technical challenges, the organizational hurdles, the economics, and the opportunities.
That experience is making us a better partner to our clients.
