In This Article
Rough notes are perfectly normal in a working business. The trouble starts when shorthand turns into a customer-facing promise. 'Sides and rear, ceramic, maybe removal' is not yet a clear quote. AI can help turn those notes into a readable draft, but it should expose the missing decisions rather than silently make them.
Keep your normal pricing and approval process. This workflow is about document preparation, not automatic price setting or legal terms. The example is hypothetical, with no real customer's details. A quote should only leave your business after you have checked the scope, price and conditions yourself.
1. Capture facts in separate fields
Instead of one paragraph, use a small job record. Distinguish what the customer requested from what you have inspected and approved. If you have not seen the vehicle, say so in your internal notes. That distinction prevents a customer's guess about existing film from becoming your confirmed assessment.
- Vehicle: verified year, make, model and body style; use a job reference instead of personal identifiers in the AI draft.
- Requested work: exact glass to be included, with windshield, roof or other exclusions stated where relevant.
- Product: approved film brand and series, plus any selection still to be confirmed.
- Condition: existing film or unusual work, marked confirmed or awaiting inspection.
- Commercial details: price from your approved calculation, current quote validity and any applicable terms you already use.
- Schedule: requested date versus confirmed availability. They are separate fields.
2. Give the model a deliberately incomplete example
Example notes: Job Q-014. Vehicle details verified in our own system. Side and rear glass requested; windshield and roof excluded. Customer prefers a ceramic option, exact series not selected. Existing rear film may need removal; inspection pending. No approved price. Customer asks about Saturday; availability not checked.
In that example, a finished quote would be premature. The useful output is a scope summary and a short missing-information list. If the model produces a price, names a film or announces a Saturday booking, it has failed the task—even if the document looks professional.
3. Ask for structure, not invention
Example prompt: Organise these notes into a quote preparation draft. Use headings: Requested scope, Confirmed details, Exclusions, Needs confirmation. Preserve uncertainty. Include no price, dates, film specifications or terms unless explicitly supplied. Do not write a final customer quote while required details are missing.
An acceptable result would list side and rear glass under requested scope, windshield and roof under exclusions, and film selection, removal, price and availability under needs confirmation. You can then resolve those items with the customer and your own records. AI has saved formatting effort without pretending the job is more settled than it is.
4. Add approved numbers and check them separately
Once the scope is resolved, insert the price from your own calculation. Check line items and totals with a calculator or your established quoting system. Keep internal labour allowances and overhead calculations separate from the customer-facing explanation. Our worked pricing example explains why the difference between revenue and material cost is not automatically profit.
Commercial claims also need checking: product performance, warranty wording and what 'included' means. The FTC's small-business advertising guidance explains the need for truthful, supported claims. Do not let a writing tool upgrade a manufacturer's limited statement into a broader promise.
5. Keep the approved version
Review the draft beside the original notes, not from memory. Check the vehicle, glass, product, exclusions, price, timing and any pending inspection. Save the final approved version in your usual job system with a reference and date. If the scope changes, record the changed quote so the customer and installer are working from the same version.
Do not use the AI chat as your only business record. Keep customer identifiers in the system that needs them, following the private administration workflow. Used this way, AI helps make the quote easier to read while you retain control over the work you are agreeing to deliver.
