The Real Resourcing Math of APM
In a recent meeting, our sponsor told their CEO they were replacing Adaptive with Fabric.
“What’s Adaptive? Never heard of it,” the CEO said.
Our sponsor followed up: “The bigger opportunity is replacing Tableau.”
The CEO leaned in. “Tell me more.”
That exchange captures the real scope of Agentic Performance Management (APM) – and why the most common objection we hear about it, resourcing, is often framed against the wrong benchmark.
APM doesn’t just replace your CPM tool. It covers data, BI, CPM, and the new agentic layer in a single motion. The investment, impact, and value should be measured against all four, not against traditional CPM alone.
It’s completely understandable to have questions around the resourcing required for APM. Do the tools require investment and time to learn? Yes. But no more than any enterprise-grade CPM platform demands, and the return is on a different order. Here is how the investment is different:
Faster implementation, fewer people.
For one client, we replaced five years of Anaplan work with APM in under a year, at half the initial investment, with no more than three FTEs staffed. That leverage is compounding with AI. AI agents can already handle more of the design and build directly in Fabric with open APIs and DevOps tools - something no proprietary platform can match. The first time I confirmed the before and after, I found it hard to believe. Now I’m now convinced we’re only scratching the surface.
No CPM background required.
We just delivered a project with two team members who had under six months of experience and no formal CPM background. One has a solid handle on SQL and Python. General technical aptitude was enough to deliver. This is a major departure from Anaplan, where it would often be 9+ months before someone could lead a project end-to-end.
Learning investment pays off long-term.
The skills APM demands are the ones that keep their value. Time spent building agentically with SQL, Python, and React is far more transferable and durable than learning how to make a specific tool’s formulas work for your use case. And the alternative isn’t necessarily simpler. Newer CPM tools carry similar learning curves, which is unsurprising, because performance management at scale is inherently complex. Mid-market tools can start simpler, but unless they stay at QuickBooks scale, they all get harder over time. In contrast, APM gets easier as you get more proficient with agentic building and harnesses get even better.
Resourcing built for the AI era.
Before we pivoted toward APM, we were deep in the Anaplan space. Anaplan’s own recommendation was a center of excellence staffed with not only model-builders but dedicated UX designers and other specialized roles. That’s significant headcount with limited supply and I did not support this approach for most clients. APM runs on a leaner shape: general data engineering support, embedded dual-threat product owners, and AI engineering support. This is the resourcing model of the AI era, not the cloud-SaaS era.
Ownership transfers faster.
Once the initial technical barriers break, teams pick up the slack surprisingly quickly. One client built out their own integrations, models, and even a reporting MCP server inside their first year. That agency extends well beyond CPM scope into daily work. Teams get acclimated to using Claude Skills, Claude for Excel, and micro-apps for Fabric, then quickly apply those skills elsewhere, becoming major force multipliers.
More scope, better ROI.
Because APM spans those four categories at once, a data engineer isn’t just building integrations. They can own the architecture and the flows all the way through to the reporting model. Similarly, a domain expert can stretch more into the technical and product domain than they ever could before, handling changes on their own without waiting on IT tickets. With broader scope, the resourcing math gets more compelling, not less.
Where does the extra time go?
Less time spent on technical design, building, and integrating systems means more time optimizing business processes, more time building strategic business partnerships, and more time driving adoption and change. We’ve found that while nearly everything else can be accelerated, these areas still very much require human effort and time.
APM looks daunting at first, the way the first iPhone did: a phone, a music player, a TV, and a hundred apps crammed into one device that people initially struggled to categorize. AI and genuine platforms like Fabric now make enterprise software integration possible at a depth the CPM space has never had. That’s what APM represents, and why resourcing it properly delivers real value.
