Sample report — what £300 buys

This is a sample AI Opportunity Assessment — the exact document you get for £300, in the format you’d get it, at roughly the length you’d get it.

The company below is fictional. The format, the arithmetic, and the confidence notes are not. Every number is worked from evidence the way we’d work from yours — and where the evidence runs out, the report says so instead of guessing.

AI Opportunity Assessment

Prepared for: Broadland Hose & Fittings Ltd (a fictional company)
Prepared by: Trippnology
Date: 2026-11-09
Fee: £300, credited against the build if you proceed. Includes a Project Assessment once you choose which opportunities to pursue.

1. What this document is

This is an assessment, not a quote. It records what the job looks like today, then ranks the three tasks worth automating, in the order we would build them, with the time and money each saves. You keep this document whoever builds the work.

It is based on the completed AI Tools Questionnaire, the call on 2026-11-05, and the evidence you provided on 2026-11-06: your sent-quotes export, the master price spreadsheet, three supplier price-list PDFs, and a timed sample of ten real quotes.

One limitation: we saw one week’s timing and eight weeks of quote volume. That’s enough to rank the work; it’s not enough to promise the figures to the pound — see the risks in section 10.

2. What the job looks like today

Broadland Hose & Fittings sells industrial hose, valves, and fittings from a trade counter near Acle, to farms, boatyards, and workshops. Six staff. Customers ring or email; the website is a brochure with a PDF price list.

The job in question belongs to Jo, the office manager, and it works like this: an enquiry arrives by email or phone, Jo finds the part in the supplier PDF price list (Ctrl+F), checks stock in the master spreadsheet, then writes a reply with the price and lead time. If the quote isn’t accepted in a week or so, it either gets chased by hand or quietly dies.

Work type Per week Time each Hours per week
Writing quotes 28 ~15 min 7.0
Chasing quotes (calls + “did you get my quote?” emails) ~11 ~16 min 3.0
Re-keying supplier price-list updates 5 hrs per quarter — 0.4
Combined — — ~10.4

What this costs today:

  • Quotes lapse. Dave (owner) estimates two to three orders a month are lost to quotes that were never chased. Nobody counts them, which is the point.
  • Jo gets interrupted. Quote work happens between phone calls, in the afternoon slump, and takes longer than it should because of it.
  • The evidence exposed a data problem. Eleven lines in the master spreadsheet disagree with the supplier PDFs — two are selling at last year’s price. Re-keying does that, eventually.

Added up, that table is about £14,000 a year of Jo’s time at her loaded cost — before counting a single lapsed order. Nobody invoices for it, so it feels free. It isn’t.

3. The three opportunities, ranked

Ranked by payback: how quickly the build pays for itself against the hours it saves. The arithmetic assumptions are in the appendix; change them and the figures move.

Rank Task Hours saved/week £ saved/year Payback on build
1 Quote drafting 6.0 £8,460 3–4 months
2 Quote chasing 2.25 £3,175 ~4 months
3 Supplier price-list updates 0.3 £480 — see note

On top of the table: one recovered order a month — the least the quote chasing should do — is worth roughly £900 a year of gross margin that today lapses. It stays out of the figures above on purpose; it has to earn its way in through the quote log.

Rank 1 — Quote drafting

What it is: when an enquiry arrives, the tool reads it, looks the parts up in the master spreadsheet, and drafts the reply — prices, lead times, your wording — into a review queue. Jo reads it, adjusts anything that needs adjusting, clicks send. The tool never sends anything; Jo approves every quote.

Volume: 28 quotes per week, currently ~15 minutes each (7 hours/week). Counted from your sent folder: 214 quotes over eight weeks, 27–30 most weeks. Jo’s timed sample of ten real quotes came out at 14–16 minutes, which matches.

After: Jo reviews a drafted reply — about 2–3 minutes each, roughly an hour a week.

Time saved: 6 hours per week ≈ 282 hours per year.
Money saved: ≈ £8,460 per year, counting Jo’s time at her loaded cost of £30/hr (6 hrs × 47 weeks × £30).

Confidence: high — the quotes follow a fixed shape, and two independent sources (sent folder, timed sample) agree on the volume.

Depends on: the price data being machine-readable. Today it’s a PDF, so step one of any build is a one-time clean-up of the master spreadsheet (section 5).

Rank 2 — Quote chasing

What it is: about 40% of quotes need a nudge. Instead of Jo remembering to chase, the tool reads the quote log, drafts the follow-up — “just checking you saw our quote for the camlock fittings, still happy to help” — and Jo ticks and sends. Phone calls stay human, for the big ones.

Volume: ~11 quotes per week worth chasing, currently 3 hours/week of mixed calls and re-sent emails.

After: tick-and-send on drafted chases, ~45 minutes a week.

Time saved: 2.25 hours per week ≈ 106 hours per year.
Money saved: ≈ £3,175 per year at £30/hr loaded cost (2.25 × 47 × £30). Separately — and not counted in that figure — the chases that currently never happen are where your two-to-three lost orders a month live. One recovered order a month is roughly £900 of gross margin a year (average order £220 at ~35% trade margin); if the tool does the least it should, the margin alone pays for version 2’s build in year one. We don’t count it because we can’t prove it yet; the quote log will.

Confidence: medium — the hours saved are certain; the recovered-orders effect is real but unmeasured until the log exists.

Depends on: the quote log from rank 1. This is why chasing is version 2, not version 1.

Rank 3 — Supplier price-list updates

What it is: when the quarterly supplier PDFs arrive, the tool reads them, diffs them against the master spreadsheet, and proposes a change list — new prices, discontinued lines, mistakes. Jo approves, and the website PDF and trade-counter sheet regenerate from the one clean master. One source of truth, three outputs.

Volume: 5 hours per quarter currently; roughly an hour of review after. Saves 4 hours per quarter ≈ 16 hours per year.
Money saved: ≈ £480 per year at £30/hr loaded cost (4 hrs × 4 quarters × £30).

Confidence: high on the mechanics, modest on the payoff — it’s ranked third because the hours are small. But the eleven wrong lines we found are the real argument: the current process eventually sells something at the wrong price, and the fix costs almost nothing once rank 1’s clean master exists. We’d build this third, or fold its data work into rank 1’s build.

Depends on: the same clean master spreadsheet as rank 1. Build the master once, serve both.

Combined: about 8.5 hours a week, or £12,000 a year at Jo’s loaded cost — plus the orders that currently lapse, worth conservatively another £900 a year at one recovered order a month. Call it thirteen, counting only the things we can defend. Still not counted: Dave’s time, the interruption tax on Jo’s afternoons, and the wrong-price exposure from the eleven stale lines — because we can’t put honest numbers on any of them yet.

4. What we recommend building first

The shape

The tool watches the enquiries mailbox. When an enquiry looks like a quote request, it drafts the reply into a review queue — a simple list where Jo opens each draft, edits if needed, and sends from your normal email. There is no chatbot, no auto-send, and nothing your customers would notice except faster, more consistent quotes. It runs in the background of what you already do, not instead of it.

Why this shape and not a bigger one: the failure mode that matters here is a wrong price going out to a farmer who then tells everyone at the mart. Keeping a human on every send costs Jo an hour a week and buys you zero wrong-price incidents. The bigger shape — auto-sending, answering the phone — is where that risk lives. Not version 1. Possibly not ever; see section 11.

Scope of what comes first

Version 1: quote drafting for the current top part ranges, with the quote log you asked for — every quote issued, what was quoted, whether it was won. Around three to four weeks from sign-off.

Version 2, after measured use: the chase emails from the quote log, still tick-and-send. Started once V1’s log shows a month of real quotes to chase.

Version 3: the supplier price-list diff and regeneration of the website PDF and counter sheet. Quarterly job, so it can wait its turn.

Keeping these separate is deliberate. Each has a different failure mode, and a problem with one should never be able to break another.

5. Fix these before or during the build

Issue Source Fix
Master price data lives in a PDF workflow Suppliers send quarterly PDFs; lines get re-keyed One-time clean-up of the master spreadsheet (a day of Jo’s time, with the eleven wrong lines from our evidence on her desk), then the master is the source and the PDFs become inputs
Eleven stale/wrong price lines Our diff of your master vs supplier PDFs, 2026-11-06 Included in the clean-up above; the tool must flag — not silently trust — any price older than the quarter
Stock levels partially in heads and whiteboard Trade counter Closing stock goes into the spreadsheet at day’s end (2 minutes); until then the tool must say “check stock” on affected parts and Jo confirms — built in regardless

Until these are done, the tool must refuse to act on the affected cases and pass them to a person — which we will build in regardless.

6. Proposed rules (for your approval)

Rule Examples Target
Quotes over £2,000 flag to Dave before sending Hydraulic press kits, bulk valve orders Dave eyeballs the big ones without doing all of them
Stock shows under 2 → draft says “check stock” 1″ camlock couplings Jo confirms availability; the tool never promises stock it can’t see
Nothing auto-sends. Ever. Every quote, every chase Jo clicks send on 100% of customer-facing mail

Owners: Jo approves drafts day to day; Dave is the fallback approver for holiday cover. If neither is around, the queue holds and the quotes wait — which is exactly what happens now, minus the drafting.

7. How we will measure success

Baseline from the evidence, to be confirmed on the first monthly report:

Metric Baseline 30-day target 90-day target
Time spent quoting + chasing ~10 hrs/week ≤5 hrs/week ≤2.5 hrs/week
Quote turnaround (enquiry → sent) Same day, typically 2–4 hrs Under 1 hour Under 15 min
Stale price lines in the master 11 found this quarter ≤2 0

Jo and Dave confirm these numbers before the build starts. They become the monthly report we send during the support period — including the one number nobody tracks today: quotes sent vs quotes won.

8. If it stops working

The old way still works and stays in place. The master spreadsheet, the PDF price lists, and Jo’s fifteen-minutes-per-quote process are untouched by the build — the tool reads them, it doesn’t replace them. If the automation fails, Jo quotes by hand the same afternoon and rings us. The fallback runs indefinitely; nothing expires and nothing holds your data hostage.

9. Indicative costs

Item Indicative
Version 1 — quote drafting + quote log £2,000–£2,800
Version 2 — quote chaser +£800–£1,200
Version 3 — supplier price-list diff + regeneration +£800–£1,200
Running costs £10–£25 per month
Support and improvements £60–£120 per month

Included in the build: mailbox connection, the review queue, the quote log, the refusal-to-guess behaviour from section 5, and the monthly report during the support period. These are indicative ranges so you can plan; the Project Assessment — included in your £300 — fixes scope, timeline, and firm prices before any building starts. At the middle of the range, version 1 pays for itself in about three months — and that’s counting only the hours, not the lapsed orders.

10. Risks and dependencies

  • Mailbox access. The tool needs read access to the enquiries mailbox — a shared mailbox is cleanest. You provide it; nothing starts without it.
  • Data quality. The eleven wrong lines must be fixed before go-live. Until then the tool flags old prices rather than quoting them — the safety net is built in, but the clean-up is on your side of the fence.
  • Adoption. The savings assume Jo works from the review queue. If she quietly goes back to the old way after a fortnight, the hours come back — the first monthly report will show it, and we’d rather know than not.
  • Evidence limitation. One timed week and eight weeks of volume. We’d be surprised if the true figures moved more than ±25% either way, but a different quarter — beet season, boat season — could move them. The 30-day report replaces this section’s guesswork with your numbers.

11. Out of scope

  • Stock reordering optimisation — needs a year of clean sales data first; the quote log starts collecting it as a side effect
  • Accounts integration — nothing in this assessment touches your accounts package
  • The website — it’s a brochure, it works, it stays one

What we’d tell you not to do

Someone, somewhere, will happily sell you all three of these. This report would tell you not to. For the record:

  • A website chatbot. Your site gets about 150 visits a month. A robot receptionist there is a pet, not a tool.
  • A bespoke AI model. Your problem is admin, not data science. Everything above works with tools that already exist; custom training would be paying more to know less.
  • The printed trade catalogue. You’re retiring it next spring. Automating its production would be polishing something with a launch date for the skip — the report says wait, and this is us saying it in writing so you can hold us to it.

12. Next steps

We suggest a 30-minute call to walk through this assessment. Three outcomes, all fine:

  1. Proceed with us — choose the opportunity or opportunities you want to pursue. The Project Assessment is included, fixing scope, timeline, and firm costs; the fee is credited against the build.
  2. Proceed elsewhere — this document is yours to hand to another supplier, and we are happy to answer their questions.
  3. Park it — no follow-up sequence. The report will still be here when the job next becomes the problem of the week.

Appendix — evidence inspected

  • AI Tools Questionnaire, completed by Dave Carter, received 2026-11-02
  • Session at the trade counter, 2026-11-05 (60 minutes) — current-state process capture written up the same day
  • Sent-quotes export: 214 quotes over eight weeks, received 2026-11-06
  • Master price spreadsheet: 520 lines, received 2026-11-06
  • Supplier price-list PDFs: three suppliers, current quarter, received 2026-11-06
  • Timed sample: ten real quotes, logged by Jo, week commencing 2026-11-02

Assumptions used in the arithmetic: 47 working weeks per year (you close between Christmas and New Year, plus a week in summer), Jo’s time at £30/hr — her loaded cost: wage, employer’s NI, pension, and the holidays she’s paid for. Not her payslip rate; what an hour of Jo actually costs you. For the lapsed-order estimate only: average order £220 at ~35% trade margin. Dave’s time and trade-counter time are deliberately not counted — any figure that only works by valuing the owner’s hour is a figure that fails quietly. Quote volume comes from your sent folder, not from memory. All of these are easy to change — tell us your numbers and the figures move.


This company isn’t real. Yours is. The £300 assessment is an hour of watching how you actually work and a report like this one, built from your numbers. Book yours — or see how it works first.