AI Guides › Playbooks
By Nigel Guy · 7 min read
Most freelancers who start using AI tools do the sensible-looking thing: they finish the work in a third of the time and quote a third of the price. It feels honest, and it is how you end up competing with every other freelancer on the same platform who has made the same arithmetic choice. The tool got cheaper for everyone; your fee fell with it.
You may have seen the claim that AI freelancers charge four times more than everyone else on the same platform. We could not verify that figure from any primary source, so we do not use it here. What is durable and checkable is the mechanism: a client who is buying a result will pay for the result, and a client who is buying hours will pay for hours. This guide is about moving your quote from the second kind to the first.
The rule: price the outcome the client gets, scope it tightly, and never itemise the time or the tool.
Four things, none of them expensive.
One card per deliverable. Fill it in before you write any proposal.
| Line | What goes in it |
|---|---|
| Result | What the client can do or decide afterwards that they cannot do now |
| Deliverable | The exact object: pages, format, number of options |
| Without it | What the client currently does instead, and roughly what that costs them in staff time or delay |
| Scope edge | What is included, what is not, how many revision rounds |
| Turnaround | A date, not a number of hours |
| Price | One figure per package, chosen against the "Without it" line, not your hours |
If you tell a client "this took me ninety minutes", you have handed them an hourly rate to divide by. If you tell them "you will have a decision-ready report by Thursday", the comparison becomes the cost of not having it. You are allowed to be fast. You are just not obliged to invoice for speed as if it were a discount.
Do not hide the tool, either. If the client asks, say you use AI tools and that you review everything. Selling the result does not mean being vague about method; it means not pricing the method.
Take your current opening line and run it through three swaps.
Imagine a small letting agency in Leeds that currently has a staff member spend an afternoon each month copying enquiry data into a spreadsheet and writing a summary for the owner. This scenario and every figure in it are invented for illustration.
Your card might read: Result, the owner knows which listings generate enquiries and which stall. Deliverable, a two-page report plus a one-line recommendation per listing. Without it, a half day of staff time and a summary nobody trusts. Scope edge, one dataset export per month, one revision round. Turnaround, three working days from receiving the export. Price, three tiers, perhaps £120, £200 and £320 for the basic, standard and call-included versions, all illustrative.
The verification slot is the biggest block on purpose. It is what the client is really paying for.
You are a careful analyst writing a monthly summary for a small business owner who is not a data specialist.
Context: [BUSINESS_TYPE], reporting period [PERIOD]. The data below is an export from [SOURCE_SYSTEM]. Personal details have already been removed.
Goal: a two-page report the owner can act on. A good result names the three most important patterns, says what each probably means, and ends with one recommended action per pattern.
Steps:
1. Check the data for gaps, duplicates and odd values. List them before analysing anything.
2. Calculate [KEY_MEASURES] and show the working so I can verify it.
3. Identify the three patterns that matter most for [OWNER_GOAL].
4. Write the report in plain English, no jargon.
Output format: headings for Summary, Three patterns, Recommended actions, Data problems found. Put every number in a table I can check.
Rules: use only the data supplied. Do not invent benchmarks or industry comparisons. If something is missing, ask me for it instead of guessing. Mark any inference as an inference.
Before answering, check that every figure in the report appears in your tables and that no claim goes beyond the data.
DATA:
[PASTE_DATA]
Fill in the business type, period, source system, key measures and the owner's goal, then paste the cleaned data.
Track three things per proposal: how many clients accepted, which tier they chose, and how many revision rounds they used. If nearly everyone accepts instantly, you are probably too cheap; raise the middle tier. If nobody picks the top tier, rewrite what it adds. If revisions regularly exceed your limit, the scope edge is vague.