Get Started

An example report, in full

This is exactly what a £29 Verified Idea Report looks like — the assessment of the idea that was sent in, the evidence behind each judgement, sources you can check, and the further ideas we found around it.

Request your own report

The report, in full

This is a real report, reproduced in full. It was produced on 28 July 2026 by exactly the process that writes every £29 report. Nothing has been added, removed or reworded.

The idea it assesses is a worked example we submitted ourselves, not a customer's idea. We would not publish a customer's report without their agreement.

This report is from idea.uk. Here is the idea you sent us: We run pre-tenancy and end-of-tenancy inventory inspections for residential letting agents across the UK, and over twelve years we have built up time-stamped photographic condition records on roughly 40,000 rental properties, each linked to the deposit adjudicator's decision where the tenancy ended in a dispute.

It comes in two parts. First we assess the idea you sent: the problem it addresses, the evidence that people have it, who else is already out there, where it is defensible, where it is exposed, and a specific next step. Then we look for further ideas around it — checking each against what already exists and whether people would actually pay — and set out the ones worth pursuing, along with the ones we set aside and why.

We use AI to research and draft this report, with live web searches for the checking, and a person reads it before it is sent. Where a finding rests on something we read, the source is listed underneath so you can check it yourself.

Your idea, assessed

(You did not submit a single product idea. You described a business and an unusually good asset: twelve years of dated inspection photos on about 40,000 rental homes, with roughly 6,000 of those linked to how the deposit adjudicator actually ruled. So we have assessed the proposition your business already implies, which is: we are the inspection firm whose reports hold up when a deposit claim is challenged, because we can see how thousands of real disputes were decided.)

The problem, and the evidence people have it

The problem is that letting agents and landlords lose money on deposit claims because their evidence does not stand up, and they have no way of knowing that before they submit. This is real and documented, not guesswork. The Tenancy Deposit Scheme's own figures for the year to March 2025 (reported by the National Residential Landlords Association) show about 4.7 million deposits protected in England and Wales, an average deposit of £1,175, and roughly 1% of deposits, about 46,950 cases, going to formal adjudication. The scheme's published analysis says cleaning is the single biggest cause and appears in over half of its dispute cases, with damage to fixtures and fittings second. Its chief executive has said publicly that this points to the importance of good check-in and check-out reports. There are also named, published examples of claims failing on evidence. Property Industry Eye reported an adjudication where the agent's check-in report was described as vague and did not contain enough to help the landlord, and noted that adjudicators tend to give less weight to reports not written by an independent inventory clerk. The Deposit Protection Service published a case, covered by Estate Agent Today, where conflicting check-in paperwork and no supporting evidence left the adjudicator no choice but to decide for the tenant. One honest caveat, and it matters: 99% of tenancies end with no dispute at all. For an agent managing 1,000 tenancies that is a handful of contested cases a month, not a daily fire. A problem that is real but occasional puts a low ceiling on what people will pay to fix it.

Signs of real demand

The strongest demand evidence in this whole submission is your own: 300 agency clients on rolling annual contracts, paying repeatedly for inspections. That is money already moving, which most ideas cannot show. Second, software firms are already selling artificial intelligence report writing to exactly your audience, which means they believe agents will pay for it. Peach Software advertises that its tool turns photos and notes into report text and claims reports up to 70% faster. Inspecto sells a video walkthrough that becomes a report. The Property AI advertises detection of items, condition and damage from photos and describes its output as ready for use as deposit dispute evidence. InventoryBase, a long-standing name here, is now advertising artificial intelligence features too. On the trade side, the Association of Independent Inventory Clerks says it has over 1,350 members, so this is a real occupation with real spend around it. Now the honest part. We found no product anywhere that people pay for which predicts how an adjudicator will rule, or scores a claim's chance of success. No pricing page for it, no case studies, no agents complaining that it is missing. So the specific thing your archive makes possible is unproven demand, not a known market waiting to be served. You would be testing whether anyone wants it, not filling a queue.

Who else is addressing it, and how

Three groups. First, and most important, the deposit schemes themselves. The Tenancy Deposit Scheme already puts a free Disputes Chatbot on its own website, described as a way to find out how successful your claim would be and what you need to do to present it properly. It also publishes an Adjudication Digest that takes a real decision and explains the reasoning, offers free real-life case studies, and sells adjudication workshops through its training arm. They hold every decision they have ever made, tens of thousands a year, plus the reasons. Second, the software firms: InventoryBase, Inventory Hive, Peach Software, Inspecto and The Property AI. They are all adding photo-to-report writing and some claim damage detection. InventoryBase also runs a network for farming inspections out to freelance clerks, which is close to your own model. Third, other inspection firms competing on speed and accreditation, such as InventoryFlex in London, which advertises accredited clerks and next-day report delivery.

What people would use instead today

An agent who wants to know if a claim will stand up does four things today, all cheap or free. One, they use the deposit scheme's free material: the chatbot, the digest of real decisions, and the guide on presenting a case. Two, they ask their most experienced property manager, who has seen a hundred check-outs and has a gut feel. Three, they rely on the inventory software they already pay for, which sells its templates as dispute-ready. Four, they use a free chatbot on their phone. An agent can drop three photos into a free tool and get a tidy paragraph about a stained carpet in ten seconds. That last point is the one to sit with: the writing-up part of this is already free and good enough. And the biggest substitute of all is doing nothing and splitting the difference with the tenant, which is what happens in most cases before anyone reaches adjudication.

Where it is defensible

Four things here are genuinely hard for someone else to copy. First, the 6,000 cases where you hold both the condition of the property and the ruling that followed. Photos on their own are not rare, everybody has photos. Photos with the answer attached are rare outside the deposit schemes. Second, your distribution: 300 paying clients and 90 inspectors. A software firm with a clever feature still has to win those clients one phone call at a time, and you already have them on contract. Third, you own your own scheduling and reporting software, so you can add something new without asking another company for permission or paying them a cut. Fourth, twelve years of one consistent method means the same flat can be compared across three different tenancies. That answers a question nobody else can answer: was that mark already there when the last tenant moved in. Worth noting, the most defensible version of this needs no machine learning at all. It is a second opinion service: a person with access to your archive telling an agent whether a claim is worth making.

Where it is exposed

Six real weaknesses. One, the deposit schemes are one feature away from owning this. The Tenancy Deposit Scheme already has a claim-strength chatbot on its site. If it deepens that using its full decision history, nothing you hold can match it, and it has no reason to share the data with you. Two, your dataset is thinner than 6,000 sounds. Split it by issue type, then property type, then region, then year, and each pocket becomes small. Worse, an adjudicator's decision is usually one money figure covering a whole case, not a verdict on each item or each photo. Connecting a specific photo to a specific outcome is slow manual work that somebody has to pay for before anything clever can happen. Three, permission. Those photos were taken for your clients, inside people's homes, showing tenants' belongings. Your client contracts may not allow you to re-use the material for a new product, and the images contain information about identifiable people. That is a contract and privacy question for a solicitor, and it needs answering before you spend anything. It is not a licensing problem, so it does not breach your stated red line, but it is a genuine blocker if the answer is no. Four, you have no machine learning capability, so you would buy it in. Bought capability is available to your competitors on the same terms. The clever model is never the moat, the archive is. Five, the rare-event problem. It is hard to charge a monthly fee for something an agent uses twice a month, so this probably has to be priced per case, which caps the revenue. Six, the ground is moving. The Renters' Rights Act 2025 brings a new landlord ombudsman and a national database of rented properties, with a government roadmap published in November 2025 pointing to late 2026 and beyond. Past decisions may become a weaker guide to future ones. And if you tell an agent a claim will win and it loses, expect to be blamed, so any such service must be worded as an opinion, never a promise.

A considered next step

Sell the service by hand, for money, before you build anything. This week, pick the 20 agency clients who send you the most check-outs. Offer them one thing: before they submit a deposit claim, they send you the check-in and check-out reports, and within two working days a senior member of your team compares them against similar decided cases in your archive and sends back a single page. That page says how strong the claim looks, which items an adjudicator is likely to cut, what extra evidence to add before submitting, and a realistic figure to claim. Charge a fixed fee of £45 to £60 per claim and invoice it. No software, no artificial intelligence, no development spend, just one experienced person and your own files. Take payment up front, discounted for the first ten if you must, but never free, because a free trial tells you nothing about willingness to pay. Set your decision rule now, in writing: if fewer than 6 of the 20 clients buy, and fewer than 25 claims come in over six weeks, the claim-prediction idea is not the one to build and you should stop. If they do buy, you will have written 25 of these pages by hand, and you will know precisely what a tool would need to do, which is the only sound basis for spending money on one.

Check it yourself

Who it's for

UK residential letting agents and build-to-rent operators managing roughly 200 to 5,000 tenancies each — the same firms already on your books — with the specific buyer being the person who signs off deposit deductions at the end of a tenancy.

Why they'd pay

They already pay you per inspection, so there is a live budget and no new purchase process; a losing deposit dispute costs them the deduction plus staff hours plus an angry landlord, so a tool that tells them, before they claim, what an adjudicator is likely to allow has clear money attached. Realistically they would pay a modest amount per tenancy end or a monthly fee per branch, not a large licence fee, because they are cost-conscious and buy in small steps.

Further ideas worth pursuing (best first)

1) Claim letter writer that quotes the evidence we hold worth testing now

It writes from the actual file for that tenancy, including photo references, instead of producing a blank template the agent must fill in by hand

What it's built on: Our own inspection reports and photo files for that exact property, check-in and check-out, using an AI that writes the finished claim submission, item by item, pointing to the specific photos and report lines that back each item.

What we found: I checked what agents actually complain about and what today's inventory apps already do. Published research on agents' views found the paperwork and the burden of proving the claim is their main gripe, and current apps do produce side-by-side before-and-after reports but still leave the agent to write the claim itself. So the gap is real, but this is a busy corner of the market and the only thing protecting it is that you hold the file.

How it scored: (each out of 5) hard to copy 3, people will pay 4, easy to build 4, reusable elsewhere 3, built to last 3 (18 out of 25 overall).

Risk to you: 4/5 (low — a mistake would be minor, and a refund would put it right)

A cheap first test: Write ten real claim letters this way for existing clients this month and ask if they would pay extra for it going forward, since building it needs little more than assembling data you already hold.

Check it yourself

2) Second opinion on a claim the tenant is fighting worth testing now

It is advising from recorded outcomes of similar pushbacks, not general reasoning

What it's built on: Decisions where tenants pushed back and what happened next, using an AI that reads the tenant's objection and the file, then advises settle, reduce or hold firm, with a rough likely outcome.

What we found: I checked whether there is a real settle-or-hold moment and who already helps with it. Roughly a quarter of deposit disputes are now settled by agreement before any formal decision, and the schemes offer free talks between the parties but no view on what you would probably be awarded if you pushed on. That missing number is exactly what your past pushback outcomes can supply.

How it scored: (each out of 5) hard to copy 5, people will pay 4, easy to build 3, reusable elsewhere 2, built to last 4 (18 out of 25 overall).

Risk to you: 3/5 (moderate — show your sources; insurance for handling personal data is recommended)

A cheap first test: Offer a few agents a free second opinion the next time a tenant pushes back on a claim, telling them settle, reduce, or hold firm with a rough likely outcome, and ask if they would pay for this going forward.

Check it yourself

3) Will this deduction stick? A before-you-claim check worth testing now

A free chatbot has never seen a single real adjudicator decision with the matching photos, so it can only guess at the rules of thumb we can actually measure

What it's built on: Our 6,000 finished deposit disputes, where we have the photos from check-in and check-out plus the adjudicator's written decision and how much they actually allowed, using an AI that reads the photos and the draft claim, finds the closest past cases, and gives a plain answer: likely allowed in full, likely reduced, or likely refused, with the past decisions it is copying from.

What we found: I checked how many deposit claims actually reach a formal decision and whether those decisions are public. About one in a hundred protected deposits goes to a formal decision, roughly 47,000 cases a year, and the written decisions stay private between the two parties, with only a short anonymised summary of one case published each month. The process is free but binding with no appeal on the facts, so the thing a buyer is weighing is the money at stake, not a fee.

How it scored: (each out of 5) hard to copy 5, people will pay 4, easy to build 2, reusable elsewhere 2, built to last 4 (17 out of 25 overall).

Risk to you: 3/5 (moderate — show your sources; insurance for handling personal data is recommended)

A cheap first test: Put up a simple one-page offer to letting agents: send us your draft deduction and photos, we will tell you if it is likely to be allowed in full, reduced, or refused, for a small fee, and see if anyone pays before building anything.

Check it yourself

4) Missing-evidence warning before check-out day worth testing now

It knows which gaps actually lost money in our own cases, and it can read our specific file for this property

What it's built on: The pattern of which weak spots caused past claims to fail, for example no check-in photo of the oven interior or no meter reading, using an AI that reads the check-in file and lists what the inspector must photograph at check-out to keep a claim winnable.

What we found: I checked whether advice on what evidence you need is already free, and whether the failure patterns are known. General checklists are published by the schemes, and it is widely written up that claims collapse without a signed check-in record and photos, so the broad lesson is not private. What nothing public can do is read this particular property's check-in file and name what is missing before the tenant hands the keys back, which is where your value sits.

How it scored: (each out of 5) hard to copy 3, people will pay 3, easy to build 3, reusable elsewhere 3, built to last 3 (15 out of 25 overall).

Risk to you: 4/5 (low — a mistake would be minor, and a refund would put it right)

A cheap first test: Offer a small group of agents a free one-month trial of an emailed pre-check-out warning list and ask afterwards whether they would pay to keep receiving it.

Check it yourself

5) Cleaning versus redecorating: which claim to make worth testing now

The distinction is a quirk of how adjudicators decide, learned from our decision texts, not from anything published

What it's built on: Past decisions where a claim failed only because it was described as the wrong thing, using an AI that reads the photos and advises which category to claim under and what the wording should say.

What we found: I checked how big this issue is and whether the distinction is really unpublished. Cleaning has been the single most common reason for deposit claims for five years running and redecoration is a large slice too, so the volume is there. But the rule that cleaning is judged separately from normal ageing is already written up publicly, so the only private part is the wording that actually held up in your own decided cases.

How it scored: (each out of 5) hard to copy 3, people will pay 4, easy to build 3, reusable elsewhere 2, built to last 3 (15 out of 25 overall).

Risk to you: 3/5 (moderate — show your sources; insurance for handling personal data is recommended)

A cheap first test: Offer agents a free check on their next ten cleaning or redecoration claims, telling them which category to use and why, and see if they ask to keep paying for it.

Check it yourself

6) Fair wear and tear line-drawer for carpets, walls and gardens worth testing now

The judgement is a picture-matching job against thousands of real examples, not a rule you can look up, and a free model has no examples to match against

What it's built on: Twelve years of check-in and check-out photos of the same items, so we know what four years of normal use looks like versus real damage, using an AI that looks at a photo of a carpet, wall or garden, is told the tenancy length and who lived there, and says where it sits between normal ageing and chargeable damage.

What we found: I checked whether this judgement is already available free and whether rivals do photo work. The deposit schemes and the landlord bodies publish expected item lifespans and the general principles for nothing, and at least one inventory app already sells photo analysis, so those parts are not yours alone. What I could not find anywhere public is a graded judgement built from matched before-and-after photos of the same item, which is the part that rests on your archive.

How it scored: (each out of 5) hard to copy 4, people will pay 3, easy to build 2, reusable elsewhere 2, built to last 3 (14 out of 25 overall).

Risk to you: 3/5 (moderate — show your sources; insurance for handling personal data is recommended)

A cheap first test: Offer a handful of agents a free trial where you manually grade five photos each against your archive and ask if they would pay monthly for this as a standing service.

Check it yourself

Didn't make the cut

We came up with these too, but set them aside — either because they'd be too easy for someone else to copy, or because not enough people would pay for them.

Inspector report grader that catches sloppy work same day

It scores against our own record of which report habits later lost disputes, and it plugs into the reports as they arrive

We set it aside because not enough people would pay for it (people-will-pay 2/5).

Search the archive in plain English

The archive is private; no outside model can search it

We set it aside because not enough people would pay for it (people-will-pay 2/5).

Check-in quality gate before the tenant gets the keys

It judges against our own evidence of what later fails, and it acts inside our workflow while there is still time to fix it

We set it aside because not enough people would pay for it (people-will-pay 2/5).

If you'd like us to help turn any of these ideas into a working tool — or you have any questions about this report — just email us at idea-uk@leopardess.uk.

idea.uk finds and tests AI product ideas for your business, so you can spend your time building the ones most likely to pay off.

Your report would be about your idea, researched the same way, with its own sources you can check. It costs £29, a person reads it before it is sent, and if it turns up nothing worth acting on you get your money back.

Site

Company

© 2026 idea.uk