FTV’s Alex Malvone on the Enterprise Control Layer: Managing People and AI at Scale

Professional services firms for decades operated on a relatively simple equation: more headcount meant more capacity, more billable hours and ultimately more revenue. But AI is increasingly performing meaningful portions of project work, while also making people more productive across a growing number of tasks. In other words, the old equation is breaking down.

Headcount and hours suddenly look like poor proxies for productivity, while traditional operating metrics no longer capture what’s driving the cost and value of a project.

Enterprise companies need a new control layer, one that shows an accurate picture of how people and AI are dividing work and provides a true measure of productivity and profitability.

At FTV, we believe companies building this control layer represent an important new category of technology. We talked with FTV Partner Alex Malvone, who co-leads FTV Ascend, about the control layer and how it fits into FTV’s broader thesis around an emerging AI-era challenge: how to manage and measure a genuinely blended workforce.

Everyone seems focused on AI replacing people. Are we actually asking the wrong question?

Yes. The story is not as simple as, ‘AI agents in, people out.’ A recent Economist report found that AI has so far created around 1 million new U.S. jobs, even as automation displaces some positions. I think it’s a much more nuanced evolution and conversation about how work gets done.

There will undoubtedly be some job displacement, particularly for repetitive tasks. But the bigger question is: what should humans be doing, what should AI agents be doing and where should humans remain in the loop? Once you start thinking about work that way, you’re left with a very different challenge: figuring out how to bring people and AI together in the most effective way.

What makes measuring productivity so difficult right now?

The way AI usage happens today, it can be fragmented, inconsistent and widely distributed throughout an organization. Who’s using it most and what’s the value created during each of those tasks?

Studies have found meaningful, but highly task-dependent, productivity gains from AI (in fact, a study this year showed productivity increases between 30-50% for coding and development tasks). Most firms don’t have a good way to translate those individual gains into the economics of a project.

Imagine a client pays $50,000 for an engagement that historically required 500 hours of human work. Now the same engagement takes 350 human hours because AI is doing meaningful portions of the development, analysis or production. Traditional metrics might show lower utilization or fewer billable hours, but that doesn’t mean productivity declined. It’s the opposite. The human-AI combination produced the same, or potentially better, outcomes with significantly less human effort.

Utilization and hours alone would only tell you half the story. The key to measuring productivity is understanding throughput from the human-AI combination.

Is the solution here a purely technological one, or is it more about improving operations and processes?

It’s really both. Experimentation is where AI adoption starts, but standardization is how it scales. Early on, you want employees testing different tools and finding new ways to use AI. As those use cases become embedded in daily work, companies need to turn the most effective experiments into standardized, approved workflows.

That doesn’t mean limiting innovation. It means putting the right governance and controls around AI to ensure consistency, compliance and quality, and then understanding which workflows are generating the highest return.

How does this thesis influence the kinds of companies FTV invests in?

We’re excited about companies building the control and coordination layer for enterprises, but more broadly about businesses rethinking how to build AI into operating models.

Take one of our most recent investments, COR, a platform that helps agencies and professional services firms understand and manage project work. COR gives companies better visibility into each project, bringing measurement, control and orchestration tools together in one place. By connecting a project’s price to the human and AI resources needed to deliver it, an agency has a real-time view into where it’s making or losing margin.

This builds on a theme FTV has a long and successful track record of investing in, tech-enabled services. These companies combine human expertise with technology, data and automation to execute complex workflows more efficiently. That spans functions like procurement with LogicSource and software development with DataArt, as well as vertical-specific services businesses like Patra in insurance and Lean Solutions Group in transportation and logistics.

As AI takes on more work in all arenas, enterprises need to understand how tasks are divided between people and technology. This makes the control layer increasingly important.

Is this new workforce management layer fundamentally a professional services problem, or do you see the same dynamics in other sectors and industries?

Professional services is one of the places where this is most visible because labor represents such a large share of the cost base. Historically, headcount has been the primary constraint on growth. More people, more capacity, more revenue.

AI changes that, and it isn’t unique to professional services. Take healthcare, where AI is automating tasks and workflows inside clinical documentation, medical coding and other back-office functions. In every case, the organization has the same challenge: measuring productivity and quality across a blended human-AI workflow.

In the next two to three years, how will new companies build out the enterprise control layer?

Last December, I talked about 2026 being the year coding becomes commoditized. What I meant was that as AI lowers the cost and time required to build software, simply being able to create technology becomes less of a differentiator. The harder problem is deciding what to build, how to deploy it and how to manage what it does.

The same dynamic is now playing out in day-to-day work at many organizations. Executing certain kinds of tasks will become less differentiated as AI becomes more embedded in companies. That makes the coordination and intelligence layer much more valuable, since enterprises need to know what agents are doing, how to control them and the ROI involved.

At FTV, we think companies building infrastructure in this coordination and intelligence layer will unlock the value in a new kind of workforce.

Alex Malvone co-leads FTV Ascend I, our newest fund dedicated to investments of $20 million to $60 million. FTV Ascend applies the same thematic, sector-focused approach FTV has honed over nearly three decades to a new generation of growth companies.