SB StartupBasket
All ideas
67 /100 VALIDATE Medium complexity

Creditback — hire-bill auditor for small UK contractors

Films hired plant at drop-off and collection, reads the invoice, and drafts credit-note requests for every line you can disprove.

TradeTech SaaS UK ConstructionSMBSolo-builderAI-agentBuy-sideVideo-evidence
— views
Evaluation Scores
67/100

VALIDATE

Overall Score

14
Problem
11
Demand
11
Build
9
Distrib.
10
Revenue
7
Time
5
Defense

Creditback

1. One-liner

Films hired plant at drop-off and collection, reads the invoice, and drafts credit-note requests for every line you can disprove.

2. Trend signal — why now?

Three things line up in the UK plant and tool hire market this quarter.

The complaints are dated, public and all the same shape. Trustpilot pages for the three biggest UK hire brands are a running log of the same five charges: hire days running past the off-hire, failed collections billed to the hirer, damage recharges arriving a week later with no evidence, refuelling at marked-up rates, and damage waiver charged then not honoured. From HSS ProService’s page: “Got an email on week after collection asking me to pay for damage. Was not damaged. Spent two hours on a call trying to sort out” (July 2026); “Accused me of damage to an unused transformer… with no evidence or photos” (July 2026); “I took out damage waver and they have sent me an invoice for an extra £177 to pay for accidental damage” (July 2025). From Speedy Hire’s page: “Didn’t bother to book the scaffolding back in so was charged twice in error” (June 2026); “I was then charged for 3 weeks hire, when I only used the equipment for 1 day” (May 2026); “Charged 6 weeks hire with no communication to say the scaffolding was still there after a failed collection?” (May 2025). From Travis Perkins: “hired an excavator for a weekend and got charged for a week” (May 2026); “hit with an unjustified late return charge” and “they tack on a collection charge after the item is picked up” (August 2026).

The trade press now tells contractors to build the evidence file by hand. A June 2026 trade2base guide on managing hire spend says the quiet part out loud: “hire charges do not stop when you finish using the kit. They stop when you off-hire it”; “Always get an off-hire reference number. This is non-negotiable… Without it, it’s your word against theirs, and you’ll lose”; “A timestamped photo of clean, undamaged plant going back is the cheapest insurance there is against a disputed damage charge.” It names damage waiver at “around 10–15% on top of the hire rate” and says waivers “almost always exclude theft, loss, operator negligence and damage from misuse.” An April 2026 broker piece (Townsend McCormack, via London Loves Business) adds that hire agreements carry a clause that “transfers the full financial burden for damage, loss, or theft directly to the hirer” and that “most contractors sign hire agreements without reading the fine print.” When the advice column is “keep a photo diary and write down reference numbers”, the product is the diary that reads the invoice.

The hire companies moved off-hire into their apps, and stopped exactly where the money is. Speedy’s app page promises “Off-hire in a few simple clicks and receive a time-stamped confirmation, reducing queries”, plus delivery and collection tracking and bespoke pricing. It does not mention photographs or video of equipment condition at delivery or collection, damage evidence, fuel level, or any invoice check. That is the platform stating its own exclusion. The time-stamp protects you from one of the five charges, only with one supplier, and only if you remember to press the button.

The unlock is a week old. Google released Gemini 3.8 Flash on 2 September 2026 at $0.75 per million input tokens with video input, the second Flash-tier release in three weeks after 3.7 Flash halved the previous price on 13 August. A 30-second walkthrough of a 1.5-tonne digger at delivery and again at collection, compared frame-to-frame for new damage and fuel-gauge position, now costs pennies. That is the same capability ClearReturn scored a GO on in June for the rental yard. Nobody has pointed it at the hirer.

Provenance:

3. The opportunity

Tooling follows the fee-payer. Point of Rental, inspHire, HireHop, myRMS, HiveSuite and HireCheck all sell to the hire company: they generate the invoice, log the off-hire, photograph the kit for the yard’s file. On the other side of the counter, Equipbase sells off-hire tracking to civils contractors with a plant manager and a finance team, priced by demo. Between the yard’s software and the Tier-2 contractor’s software sits the small contractor who hires a mini digger, two dumpers and a breaker most weeks, gets a consolidated invoice from three suppliers at month end, and has no record of what was actually on site, in what condition, until when.

The CPA Model Conditions that govern most UK plant hire make the hirer’s position worse than they think. The Hire Period runs until the plant is received back at the Owner’s depot, not until the phone call. The hirer’s loss-and-damage responsibility continues if the plant isn’t made available for collection. Damage is “chargeable to the Hirer’s account – fair wear and tear excepted”, and one regional yard’s terms charge missing items at full replacement cost with hire continuing at two-thirds rate until settled. Every one of those clauses is a place a charge can appear on the invoice that the hirer cannot disprove without a dated record.

Creditback is that record, plus the reader. The contractor’s site lead films the kit for 30 seconds when it lands and 30 seconds when it leaves, and forwards the off-hire confirmation. The office forwards the invoice. Creditback matches every line to the record and flags the ones that don’t fit: three hire days after the off-hire reference, a damage recharge where the collection video shows the same panel unmarked, a refuel line where the gauge read full, a waiver premium charged alongside a damage line the waiver should have absorbed, a collection charge on an item the video shows already gone. Then it drafts the credit-note request in the language the hire desk expects, with the frames attached.

The 10× is not in the video. It is that no one has ever given a small contractor a per-line verdict on a hire invoice. Today the choice is “phone the hire desk and argue from memory” or “pay it”. Hire companies’ own apps will never audit their own invoices. Yard-side software has no reason to hand the hirer evidence. This layer can only be built from the buyer’s side.

4. Target market

  • Primary customer: Owner or office manager of a UK groundworks, landscaping, drainage, or small building contractor. 3–25 staff, £500K–£8M turnover, hiring £3K–£40K a month of plant and tools on trade credit accounts across two to six suppliers (Speedy, HSS, Sunbelt, Travis Perkins Hire, GAP, and a couple of regional yards). The person who opens the monthly hire statement and sighs.
  • Why they buy: “Charged 6 weeks hire with no communication to say the scaffolding was still there after a failed collection.” “Got an email a week after collection asking me to pay for damage. Was not damaged.” “Without [the off-hire reference], it’s your word against theirs, and you’ll lose.” They are already paying 10–15% damage waiver and a hired-in plant insurance premium to manage the same exposure, and still eating the charges neither instrument covers.
  • Rough TAM reasoning: Great Britain had 370,770 VAT/PAYE-registered construction firms in Q3 2024 (ONS), 99% of them small. Speedy alone reports around 44,000 customers spanning Tier 1s to tradespeople. Firmbase lists 3,391 active plant hire companies in the UK. Assume 40,000–60,000 contractors hire regularly enough on account to feel a monthly statement. At £79 a month, 1,000 of them is roughly £950K ARR; 5,000 is roughly £4.7M. Ireland and Australia (HireCheck’s home market, same CPA-style conditions) are the second and third markets on the same product.
  • Why now for them: Hire companies are under margin pressure (Speedy posted a £32m loss and a restructure in June 2026) and have moved off-hire to self-serve apps that time-stamp the notice but hand over no condition evidence. The contractor now carries the whole burden of proof, and the trade press is telling them to build the file by hand.

5. Product sketch (MVP)

  • Delivery and collection walkthrough on WhatsApp or the app. Site lead films 30 seconds of each item at drop-off and pick-up. Creditback stamps time, location, and the item’s asset number from the label, and asks for the fuel gauge if the video missed it.
  • Off-hire ledger. Forward the supplier’s off-hire confirmation email or type the reference. Every hired item gets an expected-last-charge date.
  • Invoice reader. Forward or upload the monthly statement or per-hire invoice from any supplier. Creditback extracts every line: hire days, transport, refuel, damage recharge, waiver, cleaning, consumables.
  • Per-line verdict. Each line is Matched, Query, or Dispute with the reason in one sentence: “Hire billed to 14 Aug; off-hire ref SP-771 logged 11 Aug.” “Damage recharge £340 for boom scratch; collection video 11 Aug 16:02 shows boom unmarked.”
  • Credit-note request drafter. One click produces the email to the hire desk, quoting the supplier’s own conditions, with the frames and reference attached. Tracks the reply and the credit note.
  • Overrun alarm. Warns when an item’s expected off-hire is approaching and nothing has been logged, and when a collection is overdue against the off-hire date, so the failed-collection charge never happens.
  • Recovered-money dashboard. Credits won, charges avoided, per supplier and per site, so the owner sees the subscription pay for itself.

6. AI angle — what’s load-bearing

Two jobs no rule engine does. First, comparing two shaky handheld videos of the same excavator eleven days apart and stating whether the gouge on the boom is new, whether the fuel gauge moved, whether the safety pin is present. Second, reading hire invoices from six suppliers in six layouts, mapping each line to a hired asset and charge type, and matching it against the ledger. Both are vision-language tasks that became cheap enough for a £79-a-month product with Gemini 3.8 Flash’s pricing. Remove the model and you are left with a photo folder and a spreadsheet, which is what contractors have now and don’t use.

7. Localization angle (if any)

UK-first by design: the CPA and HAE model conditions, the named national hire brands, GBP invoicing and the WhatsApp-on-site habit are the product. Second markets share the shape: Ireland uses the same conditions and suppliers; Australia has the same hire structure (HireCheck proves the yard-side video habit exists there) and a HRIA conditions equivalent. Each new market needs a conditions library and a supplier-invoice template set, nothing structural. India and the US are not the first move: US independent yards are ClearReturn’s territory, and Indian plant hire is mostly operated hire without itemised invoices.

8. Business model — path to $1M–$5M ARR

  • Pricing: £49/month for up to £10K monthly hire spend, £99/month to £40K, £199/month above, unlimited users and sites. Optional 10% success fee on credits recovered above £500 a month, for contractors who would rather pay on results.
  • ACV: roughly £950 (blend around £79/month).
  • Rough math to $1M ARR: ~1,050 contractors × £79 × 12 ≈ £1.0M ≈ $1.3M.
  • Rough math to $5M ARR: 3,000 UK contractors at a £99 blend (£3.6M) plus Ireland and Australia at 1,000 combined, plus a bookkeeper channel at per-client pricing. Needs a partner motion through construction accountants and one hired-in plant insurer bundling it.
  • Expansion path: Owned-plant module (service records, HAV compliance evidence is already in the hire data), materials-invoice audit for the same merchants (Travis Perkins, Jewson sell both), and a hire-spend benchmark across customers that tells a contractor their day rate is 18% above the regional median.

9. Go-to-market wedge — first 100 customers

  • Free hire-bill health check in the groundworkers’ Facebook groups. UK Groundworkers, Landscapers UK, and the mini-digger owner groups each run to tens of thousands of members. Post the offer: forward your last three months of Speedy/HSS/TP invoices, get a flagged list back free. The flagged list is the demo. Target 150 uploads in six weeks; at 25% conversion that is ~38 paying accounts.
  • Construction-specialist bookkeepers and accountants. Firms marketing CIS accounting handle the hire invoices for dozens of contractors each and already pay Dext-style per-client fees. Scrape the 400 or so UK practices advertising construction specialism, send a two-minute recording of a real invoice being flagged, offer a practice tier. 5% reply, 20 practices × three clients each in month one is ~60.
  • Hired-in plant insurance brokers. Townsend McCormack, Livingstones, insurance4plant and Premier write the cover that sits beside this problem. Fewer disputed damage claims and better evidence when one is made is in their interest. Two broker referral deals in the first quarter is the target; each broker’s book is hundreds of small contractors.
  • Xero App Store listing tagged construction, plus a comparison page against Dext for hire invoices. Slow burn, but the buyer is already there looking for invoice tooling.
  • Reply to the complainers. Trustpilot review authors who name their firm can be reached through their business pages. Low volume, very high intent, and their stories become the case studies.

10. Build complexity — justification

Medium. Video capture via WhatsApp Business API or a thin PWA, Gemini 3.8 Flash for walkthrough diffing and invoice extraction, a conditions library for CPA, HAE and the big three suppliers’ own terms, and an email drafter are all off-the-shelf. The custom work is the matching logic between invoice lines, asset numbers and the off-hire ledger, plus a per-supplier invoice template set that needs real invoices to train on. A pair ships v1 in 10–12 weeks; a solo builder in about four months.

11. Gating checklist

GatePass?Note
Legal in target market✅Hirer documenting its own hire and querying its own invoices. No third-party data, no scraping of supplier systems.
Ethical — no harm / dark patterns✅Audits the bill, doesn’t dodge it. Genuine damage stays chargeable; the product only contests what the record contradicts.
Market exists (evidence above)✅Dated complaints across three national brands, trade-press advice to build the file manually, an enterprise incumbent on the hirer side, and a waiver plus insurance market priced against the same exposure.
1–5 person team can build this✅Pair, 10–12 weeks.
Launchable with <$50K / ₹40L✅Model inference, WhatsApp API, hosting. Under £6K to first paying customer.

12. Feasibility score

AxisWeightScoreNotes
Problem intensity2014/20Real money, monthly, and the contractor knows it. But they have workarounds (phone the desk, write it off) and a single wrongful charge is tens to hundreds of pounds, not thousands. Not hair-on-fire until the six-week scaffold bill arrives.
Demand evidence1511/15Twenty-plus dated complaints across three brands, a trade guide telling readers to do this by hand, insurers and waivers monetising the exposure, and Equipbase selling the enterprise version. Missing: any evidence a small contractor has paid for an audit product. Many Trustpilot voices are DIY, not trade.
Build feasibility1511/15Off-the-shelf models and channels. The invoice-template set and the matching logic need real invoices from a dozen suppliers before it is reliable.
Distribution clarity159/15Named groups, named practice list, named brokers. Conversion from a free health check is a guess; bookkeeper channels move at bookkeeper speed.
Revenue mechanics1510/15£49–£199 sits where Dext and insurance already sit. £1M ARR needs ~1,000 UK contractors, which the channels can plausibly deliver. Past that, UK-only caps the story; the success fee helps but complicates.
Time to first revenue107/10The free health check converts inside a month if the flagged list shows real money. Trial-to-paid, 4–8 weeks.
Defensibility105/10Execution moat. The accumulating library of supplier terms, invoice layouts and which charge types each desk credits without a fight compounds, but a hire company or Equipbase could ship a lite version. Hire companies won’t, structurally.
Total10067/100

13. Qualitative modifiers

Founder-fit tags

technical-heavy · sales-heavy

Technical because the walkthrough diff and multi-supplier invoice parsing have to be right on day one or the first wrongly flagged line kills trust with the hire desk. Sales because the first 100 come from Facebook groups, bookkeepers and brokers, not from a landing page.

Key assumptions to validate (3–5)

  1. Assumption: Small contractors on trade accounts eat wrongful hire charges often enough to notice. How to test: Ask 30 groundworks and landscaping owners for their last three months of hire invoices; manually audit them. Need ≥ 2 disputable lines per contractor per quarter averaging ≥ £150.
  2. Assumption: Site leads will film 30 seconds at delivery and collection. How to test: Run it with five contractors for four weeks over WhatsApp; measure capture rate. Below 60% of hires filmed, the collection-side alarm has to carry the product alone.
  3. Assumption: Hire desks credit lines when presented with dated video and the off-hire reference. How to test: Send 20 drafted credit requests from the pilot contractors; count credits issued within 30 days. Below 50% and the value proposition becomes deterrence, not recovery.
  4. Assumption: Construction bookkeepers will resell at a per-client tier. How to test: 40 emails to CIS-specialist practices with the invoice-flagging recording; need ≥ 4 to request a partner call.
  5. Assumption: Gemini 3.8 Flash reliably distinguishes new damage from existing marks on muddy plant in outdoor light. How to test: 50 paired walkthroughs from the pilot; blind-score against a human. Need ≥ 85% agreement on damage-present calls and near-zero false “new damage” claims.

Risk flags

  1. DIY-skewed evidence: The loudest public complaints come from weekend hirers, not contractors on account. Trade customers may already get credits through an account manager, shrinking the pain. The 30-invoice audit in assumption 1 is the test.
  2. Platform dependency on model pricing: The unit economics assume Flash-tier video pricing holds. Google’s introductory price runs to 31 December 2026; a reversion would still leave video cheap, but margins need checking.
  3. Supplier hostility: Hire desks could refuse video evidence or change conditions to require disputes within seven days. The product’s alarm side survives either move; the recovery side shrinks.
  4. Market cap: UK-only, tens of thousands of real buyers. £5M needs Ireland and Australia or a broker bundle, and both are second-year work.

14. Structured verdict

Score:                  67/100
Verdict:                VALIDATE
Confidence:             Medium
Best-fit builder:       Technical founder who has worked a construction site or knows someone running a groundworks firm; comfortable selling in Facebook groups and to bookkeepers
Time to revenue:        6–10 weeks (free health check → paid)
Capital to launch:      £4K–£6K ($5K–$8K)
Top 3 assumptions to validate first:
  1. Wrongful-charge frequency on trade accounts — manual audit of 30 contractors' last-quarter hire invoices; need ≥2 disputable lines averaging ≥£150 per contractor
  2. Hire desks credit against video + reference — 20 drafted requests, count credits within 30 days; need ≥50%
  3. Site leads film the kit — 4-week WhatsApp pilot with 5 contractors; need ≥60% of hires captured at both ends
Kill criteria:
  - Abandon if fewer than 6 of 30 audited contractors show a disputable line worth more than the monthly price
  - Abandon if fewer than 10 of 20 evidenced credit requests are honoured within 30 days
  - Abandon if the big three hire brands ship condition photos to the hirer in their apps before v1 ships

15. Next step — 1-week validation sprint

  • Day 1–2: Post the free hire-bill health check in three groundworks and landscaping Facebook groups and DM 20 firms named in recent Trustpilot hire complaints. Collect at least 15 contractors’ last three months of hire invoices.
  • Day 3–4: Audit them by hand against the CPA conditions and each supplier’s terms. Tally disputable lines and pounds per contractor. Draft and send five real credit-note requests on their behalf.
  • Day 5: Go if ≥ 8 of 15 contractors have ≥ £150 of disputable charges in the quarter and at least two of the five requests get a credit or a written concession within the week. Otherwise the pain is real but too small for a subscription, and the answer is a success-fee side business, not a product.

Interested in a detailed proposal?

Get a deep-dive with market research, competitive analysis, and implementation roadmap.

Contact us

info@startupbasket.ai