The Human Metric: A New Measurement Framework for the AI-Era Agency
by Hannah Kreiswirth, Founder @ wirth.works
The metrics most agencies use to run their businesses were not designed for agencies. They were borrowed from manufacturing and professional services models built around billable lawyers and accountants—industries where the relationship between time spent and value created was direct and visible. Agencies inherited them because they were available and consistent, not because they were right.
For a long time, that was fine. Now it isn't.
Average billable utilization dropped to 66.4% in 2025, a four-year low, with margins following it down.¹ Ninety percent of professional services firms are using or planning to use AI—but only 38–45% report a measurable business impact or clear ROI.² The industry is moving faster than ever and measuring itself less accurately than ever. These aren't signs of failure. They're signs that the tools we use to steer our businesses were built for a different road.
This moment of disruption is also a rare moment for redesign. And the blueprint is closer than most agencies think: a measurement system grounded in conditions that support human creativity, connected to business outcomes, and built to get smarter over time.
The Measurement Problem Isn't New—AI Just Made It Urgent
Time was always an imperfect measure of knowledge work. A clock doesn't capture judgment. An hour doesn't tell you whether the work was good, whether the client relationship is healthy, or whether the team has what they need to keep doing it well. Crucially, time doesn't measure the things that make great work possible in the first place: creativity, inspiration, innovation, health, purpose. These are the upstream conditions that determine the quality of everything that follows: who someone is, what they're capable of, how they feel in a given week, how trained and engaged they are. What time offered was consistency, a common unit that could be applied across clients and projects regardless of what was actually being made. It was a control variable, but hardly a source of truth.
And AI has broken even that. When generative tooling compresses a deliverable timeline by 3–4x, an agency still billing by the hour watches its revenue fall as its "productivity" rises. About a third of agencies have already fielded explicit client requests for an "AI discount."³ The gap between what the metric measures and what the work is worth has never been wider.
What's needed isn't a better metric but a different kind of measurement architecture entirely.
The Loop That’s Missing
A real measurement system connects what you're trying to achieve to what you're actually tracking, and tells you where you are before you've arrived somewhere you didn't want to be.
Most agencies have some version of the external loop running: delivery model to pricing model to client results. The problem is it runs in one direction. You deliver, you bill, you wait to see what happened.Time accounting doesn't help here either; it tells you what happened after the fact, not whether the conditions for good work were ever in place. There's no early warning. No internal signal feeding back into how ideas, methodology, and collaboration become most valuable.
What's missing is the loop that runs inside the organization first—one that starts with people, with the human conditions that make great work possible, and uses those as leading indicators for everything that follows. For example, Google's Project Aristotle found that psychological safety was the single strongest predictor of team effectiveness, outperforming talent, resources, and strategy.⁴ Gallup's research consistently shows that teams with strong human conditions outperform disengaged ones by 23% in profitability.⁵ But, human conditions move first and fast. Contribution fades before revenue does. Creative energy drops before client satisfaction scores do. A measurement system that starts with fiscal outcomes is always reading yesterday's news. One that starts with human conditions is reading tomorrow's.
This measurement system approach includes an internal loop and an external loop, connected and continuously informing each other. Many agencies have intentionally built strong cultures and people operations. What's been missing is the internal loop itself: structural measurement and a common metric to anchor it.
What Replaces the Hour
If time is no longer a reliable organizing unit, what replaces it? The answer that makes the most sense is the human one. Not because it's idealistic, but because it's the layer closest to where the work actually originates, the layer that moves earliest, and the layer that has always driven quality whether or not anyone was measuring it. It also mirrors what agencies already know about their best client work: that human-centered outcomes require human-centered inputs.
What this might look like in practice is something like judgment density—a way of mapping not how many hours went into a project, but how much of what got made required genuine human judgment: the account lead who hears something in a client call that changes the direction of a project; the creative director who knows, immediately, that a brief has been written for the wrong audience; the producer who catches a cultural moment and adds the one line that makes it land. None of that shows up in a timesheet. All of it is what clients are actually paying for.
This applies to how agencies use AI as much as how they use people. The quality of an AI-assisted workflow—the prompts, the processes, the guardrails, the decisions about where human judgment takes over—is determined by the humans who design it. A team operating without clarity, creative energy, or psychological safety won't build effective AI systems any more than they'll produce great work without them. The conditions that make human work excellent are the same conditions that make AI implementation effective. You can't separate the two.
This isn't an argument against AI. It's about embracing powerful technology while making a clear choice about what stays human. The question isn't what AI can't do yet—it's what we want to keep human regardless. In a service industry built on understanding people, that's not a limitation but central to the industry's value. AI gives us the chance to redirect human energy toward the work only humans can do, and that humans genuinely want to do more of, and to use those tools to make that creativity more accessible and expandable.
Building data around judgment iteratively—tracking what conditions produced the best work, which decisions led to which outcomes, where the loop tightened and where it leaked—is how the human metric becomes measurable over time. Not finding the perfect KPI once, but building a feedback system that gets more accurate—and stronger and more profitable—with each turn. What might that look like in practice? A couple starting signals worth tracking:
Creative synthesis and insight — not whether someone completed a task, but whether they made a connection nobody else did: the strategist who reframes a client's problem in a way that changes the brief, the designer whose instinct about a typeface turns out to be the whole campaign
Decision quality over time — which judgment calls led to which outcomes, and what was true about the team and moment when the best ones were made
These aren't final KPIs. They're the beginning of a data trail—one that gets more useful the more honestly and intentionally you commit to it.
Finally, the Right Data
The same technology disrupting how agencies bill for work is making it possible to measure things that were previously too expensive or too qualitative to track.
Organizational health signals that once required infrequent, expensive, or highly managed surveys can now be captured continuously. Patterns in how teams communicate, where energy is high, where friction is building, how AI-human collaboration is working—this is increasingly legible data, not just felt sense. When you can connect those conditions to outcomes over time, you begin to build something the industry has never had: an evidence base for what actually produces great work, consistently, across different clients and contexts.
That's the foundation value-based pricing has always needed. Not just the assertion that your work is worth more than your hours—the demonstration of why, grounded in the conditions and decisions designed and invested in to serve the outcomes you deliver. When internal measurement becomes accessible to clients, pricing stops being a negotiation about rates and becomes a conversation connected to evidence and value. That's a more resilient, more defensible, more scalable agency business.
The Invitation
This is a rare moment of design because this kind of disruption creates the conditions and capability to build something better than what we inherited. The same instinct that produces great client work—starting with the human, understanding the system, designing toward an outcome—is exactly what this moment is asking us to turn on ourselves.
The starting point isn't a new tool or a new dashboard. It's a set of honest questions to understand the system:
What are the human conditions that make your best work possible, and are you measuring whether those conditions exist?
What would it mean to build a feedback loop that tells you that before it shows up in a P&L?
What does a resilient, human-creativity-centered agency actually look like—designed around our desired outcomes, not inherited from industries that look nothing like ours?
The agencies that answer those questions first will also carry those models into the organizations they serve. A measurement system grounded in human conditions, connected to business outcomes, and built to get smarter over time—that's not just a better way to run an agency. The agencies that develop this infrastructure won't just run better internally—they'll be able to show clients what makes their work consistently worth the investment. That's the bridge between the measurement you do for yourselves and the trust you build externally.
It’s time for us to build the infrastructure our industry has always deserved—open, resilient, human-centered, and designed for everyone inside and besides us to thrive.
Sources
1 SPI Research, Professional Services Benchmark Report, 2026. knowledgelib.io/business/industry-benchmarks/professional-services-benchmarks-2026
2 Deltek Clarity Industry Studies, 2026. prnewswire.com/news-releases/the-latest-deltek-clarity-industry-studies
3 Digital Applied, AI-Era Agency Pricing Models: A 2026 Decision Guide. digitalapplied.com/blog/ai-agency-pricing-models-2026-decision-guide
5 Google re:Work, Project Aristotle. rework.withgoogle.com/intl/en/guides/understand-team-effectiveness
6 Gallup, State of the Global Workplace Report. gallup.com/workplace/349484/state-of-the-global-workplace.aspx
Hannah Kreiswirth, wirth.works
