AI opportunity audit
This is a worksheet, not an article. In 45–60 minutes it takes you from “we should probably use AI somewhere” to a scored shortlist of the workflows most worth automating — which is exactly what a productive discovery call starts from.
Fill it in with the people who actually do the work. Owners consistently misjudge how long routine tasks take; the person doing the task on Tuesday afternoons does not.
Step 1: inventory your workflows
List every repetitive process you can think of — anything your team does the same way more than once a week. Don’t filter yet; scoring comes next. If you’re stuck for candidates, Identifying opportunities walks through where automation potential usually hides.
| # | Workflow | Who does it | Tools involved | How often | Hours/week (all people combined) |
|---|---|---|---|---|---|
| Ex. | Supplier invoice entry | Office manager | Email, QuickBooks | ~15/week | 6 |
| 1 | |||||
| 2 | |||||
| 3 | |||||
| 4 | |||||
| 5 | |||||
| 6 | |||||
| 7 | |||||
| 8 |
Step 2: score each workflow
Score every workflow from 1 to 5 on four factors. Note that judgment is scored in reverse: a workflow that needs expert judgment on every case scores low, because rule-based work automates best.
| Score | Frequency | Time consumed | Error cost | Judgment required |
|---|---|---|---|---|
| 5 | Many times a day | 15+ hrs/week | A mistake is expensive — client-visible, financial, or compliance impact | Almost none — fixed rules, anyone with the checklist gets the same result |
| 4 | Daily | 8–15 hrs/week | Mistakes cause real rework or delay | Mostly rules with rare exceptions |
| 3 | A few times a week | 3–8 hrs/week | Mistakes are annoying but contained | Rules cover most cases; some need a human call |
| 2 | Weekly | 1–3 hrs/week | Mistakes are minor and easily caught | Frequent case-by-case decisions |
| 1 | Monthly or less | Under 1 hr/week | Mistakes barely matter | Expert judgment on every case |
Total = Frequency + Time + Error cost + Judgment (minimum 4, maximum 20).
Copy the rubric to share with your team:
AI opportunity audit — scoring rubric (score each factor 1–5)
FREQUENCY — how often does this workflow run?
5 = many times a day 4 = daily 3 = a few times a week
2 = weekly 1 = monthly or less
TIME — hours per week, everyone who touches it combined
5 = 15+ 4 = 8–15 3 = 3–8 2 = 1–3 1 = under 1
ERROR COST — what does a mistake cost when it happens?
5 = expensive (client-visible, financial, or compliance impact)
4 = real rework or delay
3 = annoying but contained
2 = minor, easily caught
1 = barely matters
JUDGMENT — how much expert judgment does it need? (reverse-scored)
5 = almost none, fixed rules
4 = mostly rules, rare exceptions
3 = rules cover most cases
2 = frequent case-by-case decisions
1 = expert judgment every time
TOTAL = sum of the four scores (4–20)
16–20 = automate first
11–15 = strong candidate (likely needs a human review gate)
6–10 = partial automation at best — revisit later
4–5 = keep it humanStep 3: rank and shortlist
Carry your scores into this table and sort by total:
| Workflow | Frequency | Time | Error cost | Judgment | Total /20 |
|---|---|---|---|---|---|
| Ex. Supplier invoice entry | 4 | 3 | 4 | 4 | 15 |
How to read the bands:
- 16–20 — automate first. High volume, rule-based, costly to get wrong. These are first-sprint candidates.
- 11–15 — strong candidates. Usually worth automating with a human review gate on the judgment-heavy cases. Map one with the workflow design worksheet to see where the gate goes.
- 6–10 — partial automation at best. Often one step inside the workflow is automatable even if the whole thing isn’t. Park these for a later sprint.
- 4–5 — keep it human. Low volume or judgment all the way through. Automating these costs more than it saves.
Step 4: put dollars on the shortlist
For your top three workflows, run the hours-per-week numbers through the ROI calculator. That turns “6 hours a week” into a monthly cost and a projected payback — the language a budget conversation actually happens in.
Bring this to your discovery call
A completed audit changes what a discovery call can accomplish. Instead of spending the call building your workflow list from scratch, we start from your scored shortlist, pressure-test the top candidates, and leave you with specific recommendations — and a fixed price in writing if you choose to go further.
If you’re already an Advizr client, bring the audit to your next strategy review or share it with your FDE in Chat — it’s the fastest way to seed the next proposal.
Book a Discovery CallNext steps
- Workflow design worksheet — map your top candidate in build-ready detail
- ROI calculator — turn hours into dollars
- Opportunity mapping — the leadership view of the same exercise