Dear colleagues,

A promising AI experiment can quickly become routine, taking old workarounds and unnecessary reporting with it. Before making it standard practice, Finance needs to ask whether the work still serves the business.

Our 2026 FP&A Trends Survey shows why this deserves attention: even AI users spend 42% of their time collecting and checking data, compared with 49% among non-users. There is still good reason to examine how this work is done and what could be simplified.

In this Digest, the featured guide, Claude Skills for Finance, offers a practical way to package recurring tasks. The other articles in this edition examine the business context AI needs, whether the work adds value and how to review the result.

I hope the time we save brings us closer to the business and gives us more opportunity to help shape its decisions.

Best regards,

Larysa Melnychuk

CEO and Founder, FP&A Trends Group

Most finance teams still use AI one prompt at a time. You get a great answer one month, then spend the next trying to recreate it.

Claude skills fix that. Define the workflow once, then run the same process every cycle. Anthropic’s own finance team has built a library of 70+.

This guide shows you how to do the same, and includes six ready-to-run .skill files, including:

• Budget Pressure-Tester: reviews a draft budget before the board does
• Model Sanity-Checker: catches broken formulas, buried hardcodes, and totals that don’t tie
• Board Pack Prep: surfaces likely board questions and drafts evidence-backed answers

You’ll also learn which workflows are worth turning into skills, and where automation should stop and finance judgment takes over.

Download the guide →

The spreadsheet that corrects the official revenue report may be familiar to Finance and invisible to AI. Oreoluwa Onabowale explains why these undocumented adjustments constrain AI returns, with five checks for the Enterprise Context a workflow needs before it can be trusted.

Check the missing context →

If a report disappeared, which decision or control would suffer? Andreas Stoller’s value-complexity matrix helps Finance distinguish work to eliminate, standardise, automate or elevate. It also exposes a less visible failure: headquarters can save time while pushing more work into subsidiaries.

Test what deserves to stay →

A polished variance explanation can arrive without the investigation that once stood behind it. Daniel Brandt distinguishes trust in the output, the process and the judgement. His test for review is practical: can the person approving the work trace the evidence, challenge the assumptions and defend the conclusion?

Read what review must establish →

Our webinars are complimentary and educational. Register to attend live or receive the recording afterwards.

The upcoming sessions extend this discussion beyond the workflow itself: how Finance uses the capacity AI releases to contribute more to business decisions, and how it assesses whether new capabilities are delivering value.

Time released from reporting creates an opportunity for Business Partnering, but influence depends on how that time is used. This session examines how FP&A enters decisions earlier, challenges assumptions and helps leaders weigh alternatives.

Register now →

Catalyst Acoustics Group provides a practitioner perspective on applying AI to finance workflows. Alongside a demonstration of reusable skills, the session examines use-case selection, ROI and integration with existing finance platforms: the work needed to turn a faster task into measurable value.

Register now →