Tokenmaxxing, DAM, and the AI Blind Spot in Professional Services | ConstellationTV Episode 130

Episode 130 of Constellation TV covers a lot of ground. Hosted by Liz Miller, VP and Principal Analyst at Constellation Research, this one moves fast and pulls no punches. Here's what we covered.

Tokenmaxxing: Right or Wrong?

Liz opened the debate with a term she made clear she does not enjoy, and invited Holger Mueller and Mike Ni to weigh in.

Holger framed it as a potential democratization of business automation. For the first time, individual employees can experiment with AI without filing a change request or waiting on IT. That's new, and it matters.

Mike's take centered on the hero. The employees who go deep and push the limits of what AI can do are the ones who create the cultural shift that eventually brings the rest of the organization along. Token usage as a metric is blunt, but unleashing the right people isn't.

Liz pushed back on both. The problem isn't experimentation. It's when maxing becomes the measure of success. Usage without purpose isn't a strategy, and organizations that treat it like one will end up farther behind, not further ahead. The right measure isn't how much. It's what changed because of it.

Constellation ShortList: Digital Asset Management

DAM doesn't get the attention it deserves. Liz made the case that digital asset management is the last mile of experience delivery and the first technology that actually brings experience teams together in a way that works.

She walked through the three maturity stages: brand assurance, where it's about storage and findability; brand security, where controls, versioning, and metadata come in; and brand safety, where AI becomes the autonomous force keeping assets usable, rights-compliant, and built for delivery at scale.

Smartsheet's Brandfolder stood out on the shortlist for its seamless integration with work management, its clean UI, and its ability to serve both power users and occasional visitors without making either group feel like they're climbing a hill just to find a logo. If you're still treating your DAM like a big bucket of folders, it's time to rethink it.

The AI Problem PSOs Keep Getting Wrong

R "Ray" Wang sat down with Robert Cesafsky, COO of Certinia, off the back of his recent HBR piece, and the conversation cuts to the heart of why so many AI deployments in professional services are underperforming.

Most PSOs are putting their AI investment into service delivery. Client-facing work. Research, content, project execution. That makes sense on the surface. But the services management side, estimation, quoting, resource management, billing, revenue recognition, is being almost entirely ignored. And it requires a fundamentally different approach.

Service delivery can tolerate probabilistic AI. Services management cannot. When you're doing ASC 606 revenue recognition, you need deterministic AI. Rules-bound, auditable, traceable. The mistake PSOs keep making is collapsing both sides into one strategy and wondering why it isn't working.

The other insight worth sitting with: where you insert the human in the loop is more important than the automation itself. Expertise is becoming a commodity. Experience is not. We are the last generation of managers to only manage humans. The organizations that build their AI architecture around that reality will be the ones that win.

Buzzword Bingo with Larry Dignan

Liz wrapped the episode by lobbing the worst buzzwords in enterprise tech at Larry Dignan to get his unfiltered reaction. Automagical. Agentrification. Customer obsessed. IMHO. And of course, tokenmaxxing.

Larry's take on token maxing: nothing is more painful than a 60-year-old CEO saying maxing. It sounds absurd, it proves nothing, and a donkey can eat a lot of hay and still not win the Kentucky Derby. If your barometer for AI success is how many tokens you burned, you are measuring the wrong thing.

Watch the full episode and subscribe to Constellation Research YouTube for Episode 131, hosted by Martin Schneider, in two weeks.