Writing
Notes on finance, systems, and AI.
I write about finance transformation, EPM systems, planning architecture, and AI-native finance.
Fix the accounting layer first: notes from the Pigment Catalyst panel
Over 50% of EPM programs experience delays, rework, or reduced adoption because organizations start before governance and data standards are in place. The tool is rarely the problem.
Read the full article →
When someone on your finance team leaves, where does the knowledge go?
Good documentation captures the what. Almost none of it captures the why. And in FP&A, the why is the whole job.
Read the full article →
Variance analysis is not a data problem. It is a memory problem.
AI can spot the anomaly. But it cannot tell you why the number moved, because the why was never written down anywhere it can reach.
Read the full article →
A conversation with Pigment's co-CEO about EPM and frontier AI
The relationship between embedded AI in EPM and standalone frontier models is evolving fast. The vendors who listen to practitioners will lead the next chapter.
Read the full article →
I said AI could not replace EPM platforms. Then I built one in 30 minutes.
A working FP&A planning platform from a conversation. Rolling forecast, variance analysis, scenario planning, and governance. No code, no vendor, no implementation timeline.
Read the full article →
Pigment vs Anaplan: a modern approach to EPM architecture
Pigment and Anaplan solve the same planning problem with very different model engines. Here is how they compare on structural flexibility, implementation speed, and data integration.
Read the full article →
AI-native EPM will not just be better AI inside old platforms
The next generation of planning systems will not be Anaplan or Pigment with a chat box bolted on. It will be planning surfaces designed around model generation, dynamic UIs, and workflow context from day one.
Read the full article →
The problem with mixing bottom-up forecasts and CFO overlays
Bottom-up forecasts and CFO overlays solve different problems. When you blend them in the same column, you lose the ability to defend either one.
Read the full article →
Finance workflows are institutional memory
Most of what a senior FP&A analyst knows is not in the model. It is in the steps they take to build it. That is the layer worth capturing.
Read the full article →
Why AI agents need clean finance data
Finance data is not just messy. It is structurally ambiguous. Agents that can reason over it require a semantic layer the org actually agrees on.
Read the full article →
Planning granularity is not just a model design choice
Granularity decisions cascade into reporting cadence, accountability, and how operators actually use the plan. Pick the wrong level and the model becomes shelfware.
Read the full article →
The hidden architecture behind good FP&A
Good FP&A teams look fast because the structural work was done long before the deadline. Here is the architecture underneath.
Read the full article →