Sample report
Read the whole deliverable
before you commission one.
This is a complete Workflow Audit Report, the Phase 1 deliverable from an AI Ops Review. Ranked pain points, root-cause analysis, data readiness, one recommended first build, and the work we tell clients not to do. The client is fictional so we could publish it in full.
- Pages
- 11
- Engagement
- A$4,950
PDF · 11 pages · 130 KB. No email required.
What's inside
Nine sections, including one that tells you what not to build.
- 01
Executive summary
Where the time goes and why, in three paragraphs. Written so a partner who reads nothing else still knows what to do next.
- 02
Workflows inspected
Every workflow we looked at, with frequency, headcount, current time cost, and a pain rating. That includes the ones we assessed and dropped.
- 03
Technical constraints
The systems you run, where your data lives, the compliance rules you work inside, and the fixes you already tried that failed.
- 04
Ranked pain points
Each pain scored on impact, tractability, and data readiness, then ranked. A list tells you nothing about what to do first.
- 05
Root cause analysis
For each pain: the symptom, the immediate cause, the systemic cause, what you tried, and why it didn't work.
- 06
Data readiness
Whether the data behind each pain can support a build today, or what you have to clean up first. This is usually the constraint, not the model.
- 07
The recommended first build
One opportunity. Expected impact in hours and dollars, an effort estimate, and the risks that could sink it.
- 08
What we recommend against
The builds that look obvious and aren't, named, with the reasoning. Most audits skip this. It's the section that saves money.
- 09
Next steps and appendix
Three follow-ups, the interview log, methodology notes, and the conditions that would make the report stale.
From the report
This is what a finding looks like.
Three excerpts from the PDF. The subject is a 40-person accounting firm losing senior capacity to monthly management accounts.
Ranked pain points
| # | Pain | Category | Cost | Impact | Tractability | Data readiness |
|---|---|---|---|---|---|---|
| 1 | Management accounts production | Reporting drudgery | 9–10 h/wk | 5 | 4 | Ready |
| 2 | Dual-system WIP aggregation | Manual reconciliation | 2–2.5 h/wk | 4 | 3 | Needs cleanup |
| 3 | Advisory proposal assembly | Document parsing | 3–5 h/proposal | 3 | 4 | Needs cleanup |
The recommended first build
Automated management accounts commentary via Xero API
Pull current-period financial data from each client's Xero org, run it through a rules-based commentary generator that identifies the firm's standard observation types, and produce a draft narrative the senior accountant reviews and sends. The accountant stops writing and starts editing.
- Impact
- 190–260 hours recovered per year. At A$180/hr, that is A$34,000–A$47,000 of capacity, or two new retainer clients without a hire.
- Effort
- Medium. 6–10 weeks to build. Evaluate Fathom and Spotlight Reporting first, because the build may already exist.
- Risk
- The review step cannot be dropped on speed grounds. Client-specific narrative voice needs calibration over the first two to three months.
What we recommend against
“Sending AI-generated management accounts commentary to clients without accountant review: the commentary contains financial observations that carry professional liability if incorrect, and several clients table these packs as board papers for ASIC-registered entities. The 20% of exceptional months are exactly the months where an unreviewed draft is most likely to mislead, because the generator’s rule set covers the standard observation types only. A draft-and-review workflow is the minimum required; auto-send without review is not appropriate regardless of measured generator accuracy.”
About this sample
We invented the client so we could publish the report.
Claremont Business Services does not exist. We built a fictional 40-person accounting firm because publishing a real client's report would mean publishing their numbers and their systems. No real client data appears anywhere in the PDF.
The report itself went through the pipeline that builds every paying client's Phase 1 deliverable, in the same format and at the same length. Commission an AI Ops Review and this is what you get.
PDF · 11 pages · 130 KB
Want one written about your operation?
The AI Ops Review maps your workflows, ranks what each one costs you, and names the first thing worth building.
A$4,950. Two calls, two documents.