16 days in Texas: What 60+ conversations with consultancy leaders revealed about AI, human expertise, and the future of services

Over 16 days, I went from Dallas to Austin, meeting with more than 60 agency and consultancy leaders along the way. The trip culminated in my first SXSW — an experience that felt like standing at the epicenter of where technology, creativity, and business collide. We also collaborated with our friends at The Society of Digital Agencies (SoDA) across the weekend, from an intimate BBQ dinner with agency leaders to a curated afternoon of sessions that ended in a lively happy hour.

The program and speakers sparked a lot of thinking, but the ideas really took shape in the conversations around them — over coffee, in offices, at dinners, and in meetings. And one thing became very clear: AI is not just changing how agencies work, it’s redefining what they sell, how they price, and how they operate. 

AI is reshaping all phases of project delivery

AI is no longer experimental, it’s operational. Agencies and consultancies are rethinking delivery, whether it’s faster discovery and ideation, accelerated development cycles, or AI-assisted QA, content, and strategy. The result? Work that used to take weeks can now take days.

But that raises a critical question: if work takes less time, what exactly are you billing for?

The death of the hourly rate?

This question came up again and again. The traditional hourly model is under real pressure. Not just philosophically, but commercially.

We’re seeing a clear shift toward:

• Value-based pricing (charging for impact, not effort)
• Outcome-based models (tying fees to measurable results)
• Subscription models (ongoing access to teams, tools, and capabilities)

AI is accelerating this shift. When automation compresses delivery time, billing by the hour becomes harder to justify commercially.

And clients are starting to question it.

Client expectations have already changed

One of the insights from the trip is that clients now assume AI is part of delivery.

They expect:

• Faster turnaround times
• Lower costs (or at least not higher ones)
• More innovation and strategic input

But here’s the catch: while clients expect efficiency gains, they don’t yet fully understand the true cost of AI.

The hidden cost of AI (that no one is talking about enough)

Right now, AI tools are relatively inexpensive. But every consultancy leader I spoke to agrees with me on one thing: this won’t last long.

Costs will rise. But, more importantly:

• Significant time goes into training and refining AI tools
• Expertise is required to use them effectively
• Internal workflows must be redesigned
• Teams need to learn how to collaborate with AI agents

In other words, AI isn’t just a tool, it’s a capability. And capabilities have a cost.

Should AI be billable?

This sparked some of the most interesting debates during my trip.

If AI is part of delivery, should it be:

• Absorbed as overhead?
• Passed through as a cost?
• Packaged into pricing models?

AI creates a paradox for agencies. It reduces the effort required to deliver work, while increasing the expectation of value. If you price based on effort, you lose. If you price based on tools, you commoditize. 

My perspective, and one increasingly shared by agency leaders is that AI should be billable. But not as a line item. The sustainable position is to price based on value, outcomes, or ongoing access, and anchor that value in the expertise required to achieve, guide, and stand behind it.

Agencies need to:

• Educate clients early. It’s far easier than trying to justify it later.
• Frame AI as part of the value delivered
• Embed it into pricing models (value, outcome, or subscription-based)

The key is shifting the conversation from: “We used AI to do this faster” to “We delivered this outcome, enabled by AI and human expertise working together.”

A new operating model: humans + AI agents

This is where things get really interesting. If AI becomes a core part of delivery, then logically, should AI agents be treated as resources?

We’re already seeing early signs of this:

• AI agents contributing to project execution
• Automation handling repeatable delivery tasks
• Hybrid teams of humans + AI

The role of humans does not disappear. It changes. Human judgement, context, and accountability become more important, not less. AI can generate, accelerate, and scale, but it still lacks ownership, nuance, and the ability to make trade-offs in complex client situations. 

What’s becoming more interesting is defining the boundary between AI and human expertise. Where can AI accelerate or automate work, and where does human judgement, accountability, and context still need to step in? This isn’t just a theoretical question, it directly affects pricing, delivery models, and how agencies structure their hybrid teams.

Once you start looking at delivery this way, a different set of operational questions start to emerge, like:

• Do you “allocate” AI agents to projects?
• How do you measure their utilization?
• How do you forecast capacity when part of your workforce is non-human?

What this means for operations (and why it matters now)

This shift isn’t just about pricing or delivery, it’s about operations.

Professional services companies need to:

• Automate as much back-office work as possible
• Rethink resource management to include AI
• Adapt forecasting and planning models
• Support new pricing structures

This is where AI-powered operating platforms like Agileday become necessary. Not as another system to manage, but as the operations backbone that connects how work is sold, planned, and delivered in a hybrid workforce setup.

Decisions need to happen in real time, based on live data. And that changes the role of the platform entirely. It moves from a system of record to a system of coordination.

These next-generation operating platforms must:

• Support pricing models based on value, outcomes, and ongoing access, not just time
• Handle forecasting when delivery is faster, less predictable, and partly AI-driven
• Utilize AI as part of delivery operations, embedded in workflows, planning, and execution
• Provide visibility into how work actually gets done across both people and AI
• Account for the whole talent ecosystem, including employees, freelancers, vendors, and AI agents
• Connect operational data with external systems and AI tools in real time, using protocols like MCP
• Maintain control and trust over data as it moves beyond internal systems

Because the reality is that you can’t run a next-generation services business on last-generation operational tools.

Final thoughts

If there was one takeaway from 16 days in Texas, it’s this: we are still early, but the shift is already underway. AI is not replacing agencies and consultancies. It is redefining them.

The winners will be those who:

• Adapt their pricing models
• Educate their clients
• Embrace AI as part of their delivery DNA
• And modernize their operations and tooling, AI-first, to support it all

If you’re seeing similar shifts in your own business, I’d be genuinely curious how you’re approaching it ...let me know. These conversations are still evolving, and it’s clear no one has it fully figured out just yet.