TheShortlist Report

Measured research on what AI recommendsLatest reading

Counts, not opinions · No paid placement · Method published

Our books are a mess and we have an audit coming. Who can help?

Books a Mess Before an Audit? Here's Who AI Says to Call

4 firms lead, appearing in 8.0–12.0% of 25 AI responses. The sample cannot separate them, so we do not pretend it can.

25ChatGPT responses
4Firms named
1Question asked
Last measured

Based on 25 ChatGPT responses to this question, collected August 8, 2026. Firms are counted once per response. How we measure.

Key findings

  1. The Alliance Group, Everyday Accounting LLC, Kranz Consulting and Preferred Small Business Solutions lead this sample, each named in 8.0–12.0% of 25 ChatGPT responses to the question “Our books are a mess and we have an audit coming. Who can help?”, collected August 8, 2026.
  2. The gaps between The Alliance Group, Everyday Accounting LLC, Kranz Consulting and Preferred Small Business Solutions are inside the margin of error at 25 responses, so The Shortlist Report groups them in one tier rather than ranking them against each other.
  3. 4 firms cleared the reporting threshold on “Our books are a mess and we have an audit coming. Who can help?” in ChatGPT as of August 8, 2026; the full table with response counts is below.
  4. No firm paid to appear in this ranking; placement cannot be bought, and the counts come from ChatGPT responses rather than from submissions.

The ranking

Share = the percentage of the 25 responses that named the firm. Bars are drawn to the same scale across the table and repeat the number in the share column. Firms within a tier are not distinguishable at this sample size and are listed in no particular order.
FirmResponsesShare
Tier 1 4 firms, 8.0–12.0%
The Alliance Groupthealliancegroup.comProvides close-cycle optimization, process redesign, and financial transformation consulting.3 of 2512.0%
Everyday Accounting LLC2 of 258.0%
Kranz ConsultingSpecializes in ERP and close automation for reconciliations, journal entries, consolidation, and manual spreadsheet work.2 of 258.0%
Preferred Small Business Solutions2 of 258.0%

Why the tiers fall where they do

Tier 1 holds The Alliance Group, Everyday Accounting LLC, Kranz Consulting and Preferred Small Business Solutions, named in 2 to 3 of the 25 ChatGPT responses. They are grouped rather than ranked because the gaps between them are inside the margin of error at this sample size — one extra mention would reorder them, so an order would be reporting noise as a finding.

The test is McNemar’s at 95%, applied against the same 25 responses for every firm. The full reasoning, and what happened when we ranked one-to-N instead.

What the sample shows

4 firms lead this question: The Alliance Group, Everyday Accounting LLC, Kranz Consulting and Preferred Small Business Solutions. They appear in 8.0–12.0% of the 25 responses. The gaps between them are inside the margin of error at this sample size, so they are listed together rather than ranked against each other.

93 firms were named at least once across the sample. 4 of them cleared the threshold of 2 responses and appear above; the rest were named once and are inside the noise.

How to read this

Firms are grouped into tiers, not ranked one to fifteen. A tier holds every firm whose 95% confidence interval overlaps the leader's, which is a long way of saying the sample cannot tell them apart. Order within a tier is arbitrary — read the counts, not the position.

Share is the percentage of responses that named the firm — not a quality score. A firm scoring low here may be excellent and simply under-written about online, which is a finding about its visibility rather than its work.

The same question was asked repeatedly because AI assistants do not answer identically twice. A single response would tell you almost nothing.

What this measurement does not tell you

  • It is not a quality ranking. Share reflects how much has been written about a firm online, which is what an AI assistant reads. A capable firm nobody writes about scores low here.
  • It is a reading from a date range, not a permanent standing. This sample was collected August 8, 2026; models change and so do their answers.
  • It covers ChatGPT only. Other assistants answer differently, and we do not extrapolate across engines we have not measured.
  • Order within a tier is arbitrary. Read the counts, not the position. Why we use tiers.

Cite this report

Danielle Trettson. (2026). Books a Mess Before an Audit? Here's Who AI Says to Call. The Shortlist Report. https://www.shortlistreport.com/article/books-are-a-mess

Underlying counts: CSV · JSON · reusable under CC BY 4.0 with attribution.

Disclosure

No firm paid to appear on this page, and no link here is an affiliate link. Rankings come from measured responses; firms cannot submit, sponsor, or buy placement.

Something wrong on this page? How corrections work. This report is one of 30 covering fractional CFO and outsourced accounting firms see the full set.