← All use cases

Use case #9 · CPA

Firm partner with no coding background built 40+ internal apps, incl. an owner-pay tax optimiser

Mike Libbey, Partner & COO, YBL ·

The short version

What they did

Describe a tool in plain English, have AI build it, test and share; first serious build was a tax-planning optimiser for paying company owners.

Result (as claimed)

Owner-pay planning “used to take our senior accountants and CFOs eight to ten hours per client now takes about two”; firm “runs on more than forty apps I built myself, with zero developers.”

Prompt

I’m an accountant building a small owner-pay planning tool, and I don’t code. I’ll provide [jurisdiction], [entity type], [tax year], [the taxes to include], [reviewed rates and calculation rules] and [a worked spreadsheet example with made-up figures].

1. Build one HTML file that compares salary and dividend or distribution mixes using inputs I enter. Store no client data.
2. Show each mix and its total tax in a table, with the calculation steps and assumptions visible.
3. Explain how to open the file and test it against my worked example.

Use plain labels and short help text. Ask for missing rules or rates before calculating; don’t fill them from memory. List anything the tool doesn’t handle. Treat the results as planning estimates for my review, and don’t describe a mix as suitable for a particular owner without that review.

Do it yourself

These steps are my suggested way in, not Mike Libbey’s setup.

Level: Intermediate for a tool staff use. A one-file calculator is close to Beginner.

What you need

  • A chat assistant that can build and preview small web pages, such as Claude.
  • A planning calculation you already do in a spreadsheet, with a worked example you trust.
  • Made-up inputs. The tool should store no client data.

Start simple

  1. Pick one calculation you repeat, such as salary versus distribution mixes.
  2. Paste the case’s Prompt, adapted to your jurisdiction, entity type and the taxes you include.
  3. Open the file and enter your worked example. Compare every total with your spreadsheet.
  4. Tell it what is wrong and ask for one fix at a time.
  5. List the assumptions it makes on screen, and correct any you don’t agree with.

Take it further

  • Add the rates and thresholds as a visible table you update each year.
  • Keep a short list of apps you want, and build one a month.
  • Share a tool with staff only after a second person has tested it against known answers.

Check your work

  • Test edge cases: zero salary, income above wage-base limits, a loss year.
  • Check every rate and threshold against the current official source. AI tools often use last year’s numbers.
  • These tools support your advice. Review the result for each client before you rely on it.

Before you rely on it

  • This is general information about how accountants use AI, not professional guidance for your engagements.
  • You stay responsible for the work under your professional standards, including review and sign-off. An AI tool’s output is a draft, not evidence.
  • Before you send client data to an AI tool, read the tool’s data and training terms, and get any client consent you need. For tax return information, that includes IRC section 7216.

Source checked. The link, author, date and quote were confirmed against the source. The result itself has not been reproduced. Prompt adapted for this library, inspired by the case. Business size: Firm (YBL); size not stated. Results are quoted as the source claimed them and never improved. Spot an error or a broken link? Email charlie@cbarmorecpa.com.

More in Advisory/CAS

New cases go out every Saturday in Inside AI.