The AI-era fractional firm checklist

There’s a growing consensus in the fractional CFO/FP&A space that this is the most turbulent moment the industry has ever seen.

We’ve covered the broad strokes of that change across my last four posts. You have to navigate a changing playbook and an efficiency paradox that punishes you for getting faster. We talked about why change initiatives don’t just happen—they require someone to drive them after the shine wears off. Then we covered why AI adoption should work like compound interest.

If you've been following along, there's a decent chance you've been nodding while you read…and then closing the tab and getting back to your actual job.

I get it. Can’t say I haven’t done the same with similar content.

So here’s the payoff: a checklist that answers "so, what do I do about it?" that I promised at the end of the last post. One you can start working on today to position yourself and your firm for the future of fractional finance.

Phase 1: Get your positioning straight

Everything downstream is easier once you're clear on what you're actually selling. When a client asks for more work at a lower fee, the reflex is to treat it as a pricing conversation. It usually isn't. It's a sign they can't see the value in what you're already delivering, no discount fixes that.

• Write down the real reason each client hired you. Look at your SOW, and match up the work you're doing against what they committed to and what you're delivering now. Give yourself a grade and evaluate how well you are doing against the baseline.
• Run the "so what?" test on your last three deliverables. Does each one tell the client what to do differently, or does it hand them the numbers and leave the interpretation to them? Building the model has become commoditized. The recommendation attached to it is what they're paying for.
• Name where you're most exposed to the other firms. For the fee you charge, is the client getting value? Is there a reason why they might want to pursue either a more strategic resource ($) or a more junior firm (less money)? How do you combat that while you're working with the client?
• Write one sentence on why a prospect should pick you over the other three firms they're talking to. If that sentence would be true for those three firms too, it isn't a differentiator yet. The honest places to look: industry depth, the pitfalls you've already steered clients around, and the stack you've implemented across your book.

Phase 2: Focus on visibility before optimization

You can't manage what you can't see. And you definitely can't sell an efficiency you can't measure.

• Give your managers a real view of where team time goes, by client. What's still manual, what's been automated, and what's taking longer than it should. Use your Practice Management system to look at time by engagement by task. Come up with hypothesis on what workflows are taking the most amount of time, and encourage your managers to talk to their team about their workload 
• Find the work one person does in two hours that still takes everyone else two days. That gap is your best efficiency lead in the building, and it's invisible without the view above.
• Separate the profitable client from the quiet month. A repeatable process win and a client on vacation look the same on a margin report. Until you can tell them apart, you're guessing at both your real capacity and which clients are actually worth keeping.
• Decide how you'll talk about your efficiency and your tech stack before the client brings up the fee. Why you chose those tools, how they improve the speed and consistency of what you deliver, and what that means for their finance function.

Phase 3: Make your first AI deposit

AI in finance is moving at a dizzying pace. Trying to keep up with the latest models and workflows too often leads to paralysis. Start small instead: pick one workflow, get it working, write down what you learned, then go again.

• Pick one painful, repetitive workflow as an AI delegation candidate. Not the hardest problem you have. One that’s frequent enough to matter, annoying enough that the team wants it fixed, and repeatable enough that a win compounds.
• Buy something to attack it before you build anything yourself. You'll learn more from four weeks using someone else's product than from four months building your own—and right now you don't know your real requirements well enough to build against them.
• Pressure-test the solution against your real client data. Every tool demos beautifully. Demos don't have your messy exports and weird charts of accounts. Write down where it breaks; those breaks are your requirements list.
• Quantify the hours saved per client, per month. Write the number down. It's your ROI, and it's the efficiency you sell to clients instead of hiding from them.
• Pick the next workflow and run it again. If you take what you learned from the first one, this one should take around half the time to set up. These are the early returns from AI compounding.

Phase 4: Make change stick

You can nail your positioning, build real visibility, and land a genuine AI win…and still watch it all fall apart as your people slowly revert to their old ways. Lasting change needs a dogged owner and support from the top.

• Name a change leader before you announce anything. Have a bias towards the most trusted person rather than the most senior person. Good litmus test: ask your candidate the questions they'll get from the skeptics and listen to how they answer.
• Give them a mandate with numbers attached. "You own this now" is not a mandate. What does success look like at 30 days? At 60? At 90?
• Make the "why" concrete before you announce it. Not "we're rolling out a new BvA process," but what actually changes for clients, for managers, and for the person whose day it disrupts.
• Build one feedback loop. A Slack channel, a weekly working session, office hours…the format matters less than people knowing their feedback has somewhere to go. Feedback with nowhere to go turns into private complaining.
• Protect their time and back them publicly, especially during the dip. Offload some of their client work so this isn't a side task on top of a full book. Most initiatives fail because leadership is loud at launch and silent during implementation.
• Encourage wins to travel. You’ll 10X the impact of that Claude skill someone built by sharing it across the firm than keeping it siloed. Encourage your people to celebrate their wins and share their learnings.

Deciding to change was always the easy part

I don’t know a single fractional firm that has all of the above figured out. This is new to all of us and the situation changes what seems like every day with a new tool or Claude capability unlocked.

The real difference-maker at this point is just recognizing that change is needed, choosing a path, and walking down it one step at a time. The firms that feel miles ahead have gotten there by making small deposits and steadily watching them compound. They don’t have better AI tooling or a crystal ball.

Deciding to change was always the easy part. Now you've got the checklist to actually do it.

Thanks for reading along these last few months. If you want to talk through where your firm actually sits on this list, my inbox is open: david@getaleph.com.

And I still owe a few of you that BBQ ranking.