You're busy delivering work that's quietly losing you money. Here's the AI audit that shows which service lines to keep, reprice, or kill.
by Ayush Gupta's AI
The problem
Most agency founders know their total revenue and rough overall margin, but have no clear picture of which individual service lines are actually profitable. Some services look active and significant while quietly draining team time, billing below cost, and pulling margins down — and nobody catches it until it's been a pattern for years.
The fix
Run a structured AI profitability audit across every service line — effective hourly rate, actual gross margin, and team cost concentration — so founders can make grounded decisions about what to scale, reprice, or cut.
The Playbook
Pull revenue, hours, and team cost by service line
Before Claude can help, you need the inputs. Export billed revenue per service line for the last 90 days, hours tracked by service type, and your blended team cost-per-hour (total salary and overhead divided by total billable hours). If you don't track time granularly, use honest best estimates per service — a rough number is still more useful than a blank.
Run the service profitability analysis prompt
Paste your service data into Claude and run a structured analysis. The goal is to calculate effective hourly rate, actual gross margin, and a clear profitability rank across all service lines.
You are helping me audit the profitability of each service line in my agency.
I'll paste my service line data below. For each service:
1. Calculate the effective hourly rate (revenue ÷ hours delivered)
2. Calculate actual gross margin (revenue - team cost) ÷ revenue × 100
3. Flag any service where effective hourly rate falls below [TARGET RATE $/hr]
4. Flag any service where margin falls below [MARGIN FLOOR %]
5. Rank services from highest to lowest margin
6. Identify the top margin-drain culprits and explain the likely root cause
Format the output as:
- A ranked profitability table
- A short analysis of what the pattern reveals
- Three specific action recommendations: what to scale, what to reprice, and what to cut or bundle
My service line data:
[PASTE — service name, revenue (last 90 days), hours delivered, estimated team cost]Distinguish the margin leaders from the margin drains
The problem is usually not the services with the smallest revenue numbers. It's the ones with high volume but poor effective hourly rates. A $5k/month retainer delivered in 80 hours at $62/hr is a margin disaster if your blended team cost is $70/hr. A $3k project done in 10 hours is what you want more of. The audit makes this contrast impossible to ignore.
Model repricing options for every underperforming service
For any service below your target margin, Claude can model the fix options clearly: raise the rate, reduce the scope, eliminate the service standalone, or bundle it into a higher-ticket package where the economics rebalance.
This service line is underperforming on margin:
Service name: [NAME]
Current price: [PRICE/month or project]
Hours typically delivered: [HOURS]
Blended team cost: [COST]
Current margin: [MARGIN %]
Target margin: [TARGET %]
Give me four options for fixing the economics:
1. Reprice — what price hits target margin at current scope and hours?
2. Descope — what scope reductions make the current price viable?
3. Repackage — how could this be bundled into a higher-ticket offer instead of sold standalone?
4. Cut — if the math doesn't work at any realistic price point, explain why this service should be retired or used only as a loss-leader component.
For each option, give me one sentence on the real tradeoff.Commit to a specific action per underperforming service line
The audit is only worth running if it ends in a real decision — not 'we should probably revisit this.' For each underperformer: raise the price in the next proposal, stop selling it as a standalone offer, or announce a price change to existing clients with a 60-day transition. Document the decision so it doesn't silently revert under sales pressure.
What changes
A clear, data-backed view of which service lines are profitable and which are bleeding margin. Founders stop guessing about their service mix and start making active decisions about what to scale, reprice, or cut — with specific numbers behind each call.
Most agency founders know their revenue number.
Very few actually know which services are profitable.
That gap is where margin disappears.
The service line blind spot
It's easy to feel busy and assume the business is healthy.
High utilization feels like success.
A full delivery calendar feels like winning.
But busy and profitable are not the same thing.
Some service lines look active and important — recurring clients, ongoing work, real team hours allocated. And they are quietly one of the worst economic decisions the agency is making.
What agencies usually don't track
Most agencies track:
- Total monthly revenue
- Overall client count
- Rough utilization rate
Most agencies do not track:
- Effective hourly rate per service line
- Actual gross margin by service type
- Hours delivered vs. hours billed per service
- Where team cost concentration actually sits
That missing layer is what lets underperforming services survive for years without anyone flagging them.
The math that makes it visible
The core calculation is simple.
Effective hourly rate = Revenue ÷ Actual hours delivered
If your content retainer bills $2,500/month and takes 35 hours to deliver, you're earning $71/hr.
If your blended team cost is $78/hr, you're losing money on every delivery.
That is not a pricing problem in theory.
It's a margin problem hiding in plain sight — on a service the agency has been actively selling and renewing.
Running this calculation across every service — SEO, content, paid ads, web builds, audits, strategy, automation, reporting — reveals the actual economics of the business instead of the assumed ones.
What the AI audit surfaces
When you run structured profitability data through Claude, it does something founders rarely do manually: it ranks services clearly, flags the ones below margin targets, and identifies exactly where the time-cost mismatch is worst.
The results almost always fall into three categories:
Scale these — services where effective rates and margins consistently exceed targets. These deserve more sales focus, tighter packaging, and potentially a price increase.
Reprice or descope these — services that are structurally viable but mispriced or scoped too loosely. The economics can be fixed with one targeted change.
Kill or bundle these — services that only work as upsell components or entry-point loss leaders, not as standalone products. Selling them as primary offers is a margin drain that better delivery won't fix.
What founders usually discover
When agencies run this audit honestly, the patterns are predictable.
Content retainers often have poor hourly rates because revision rounds were never priced into the original scope.
Web projects frequently look like wins on invoice but show thin margins when development hours are tracked accurately.
Strategy and advisory work almost always has the best margins — and is chronically undersold because it feels less tangible than execution work.
Reporting absorbs significant hours that were never explicitly priced into the retainer.
None of this is obvious from a revenue dashboard.
It only shows up when you run the actual numbers.
The decision that matters
The audit is only worth running if it produces a real decision.
Not "we should probably look at repricing that at some point."
An actual commitment: what price, what scope change, what offer restructure, by what date.
One service line repriced correctly — even a 20% rate increase on a service that was genuinely underpriced — often recovers more margin than adding a full new client at the old economics would.
Bottom line
The AI audit takes a few hours.
What it reveals is usually a meaningful margin improvement — sometimes immediately, sometimes through repricing conversations that happen in the next proposal cycle.
That is a high-leverage use of an afternoon that most agency founders keep putting off.