Illustrative example · fictional business
This is what you would receive
A complete AI Opportunity Audit, walked end to end for three made-up firms. None are real clients. Pick the one closest to yours. Every number is built from a business's own figures and shows its working.
The full audit, walked through for a made-up firm: Kestrel Trade Supplies, an independent building & plumbing merchant we invented for this page (trade counter, phone and email orders). It's not a real client. Every number is built from a business's own figures and shows its working.
Executive summary
The prize, in four numbers
Kestrel's team is quick and knowledgeable, but half the day goes on re-typing orders, pricing quotes by hand, and chasing deliveries. That manual load is a large, measurable time cost, and it's squarely in AI's wheelhouse.
Two kinds of number, kept apart. The HARD floor is time saved, from Kestrel's own order, quote and pricing volumes (46 working weeks, conservative coverage). You can bank it, and it pays for the build in aboutfour months on its own. The growth upside (more quotes won on speed) is shown per opportunity and proven in the pilot, never bundled into the headline. Provenance on every figure: D your data · M measured from your systems · B benchmark · A assumption (flagged, pilot-proven).
What we found
Four themes, in their own words
From a short operations review and a look at the order log, quote history and supplier price files, each grounded in a real number and a real quote.
Orders arrive as messages, not data
~300 orders a week come in by phone, email and WhatsApp, each re-keyed into the trade system by hand. It is the single biggest drain on the office.
“Half the day is just typing orders in from emails and voicemails.” (Sales office)
Slow quotes lose the job
~120 quotes a week, ~20 minutes each to price up from the book. Trade buyers ring three merchants and buy from whoever quotes first.
“By the time we've priced it up, they've ordered elsewhere.” (Branch manager)
Weekly repricing is a manual grind
Supplier price files land most weeks; thousands of lines are updated by hand, and pricing errors slip through to quotes.
“Every Monday is copying supplier prices into the system.” (Owner)
“Where's my delivery?” owns the phones
Variable lead times and pallet networks generate a constant stream of chasing calls, all landing on the trade-counter line.
“The counter phone never stops with delivery chasing.” (Office)
Your business, as we understand it
Where your information lives today
The work is excellent; the constraint is that every order, quote and price update runs through people, because the systems don't talk to each other.
Four disconnected islands: people are the integration. It is the single biggest limiter on cheap automation, and it shapes the sequencing below.
The good news: the two highest-value wins don't need these joined up first. They run from an order export and the price book. That's why we lead with them.
The opportunity landscape
Every opportunity, ranked
Seven opportunities scored on impact (what it's worth) against feasibility (how quickly and cheaply we can ship it). Bubble size shows annual value, and the leaders are the ones that hand your team its day back.
| # | Opportunity | What the AI does | Value / yr | When |
|---|---|---|---|---|
| 1 | Order & quote desk assistant | Reads emailed/messaged orders & quote requests, drafts them into the trade system for a person to approve | £22 to 32k hard · win-rate upside | Wave 1 |
| 2 | Supplier price-list ingestion | Turns weekly supplier price files into a validated, ready-to-import update | ~£12 to 18k hard | Wave 1 |
| 3 | Delivery-status & chasing deflection | Answers “where's my order” from courier + system data, on phone and email | ~£8 to 12k hard | Wave 1 |
| 4 | Trade-counter product & substitute finder | Helps staff find the right part or an in-stock substitute in seconds | ~£6 to 10k hard · margin upside | Wave 2 |
| 5 | Statement & overdue-account chasing | Drafts statements and runs polite, staged dunning on trade accounts | ~£7 to 11k hard | Wave 2 |
| 6 | Back-in-stock & reorder prompts | Tells customers when a line returns; nudges predictable reorders | ~£8 to 16k upside | Wave 2 |
| 7 | Demand & stock forecasting | Predicts fast movers and reorder points from sales history | strategic | Wave 3 |
Opportunity deep-dive · 1 of 2 shown
★ Order & quote desk assistant
Turn every emailed order and phoned-in quote request into a drafted, priced entry your team just checks and sends, in minutes, not half a day.
How it works
~300 orders + 120 quotes a week re-keyed and priced by hand; ~60 office hours a week on typing, and quotes that land a day late lose to faster merchants.
Orders and quotes arrive pre-drafted and priced; the team checks and sends; most quotes go out same-hour.
HARD (proven): your logs show ~300 orders + 120 quotes a week entered by hand, about 60 office hours (M). Drafting 45 to 65% of that automatically frees ~27 to 39 hours a week; at £18/hr over 46 weeks that is ~£22,000 to £32,000/yr you can bank. That is the floor, and it pays back the build in ~4 months on its own.
Hard: ~27 to 39 hrs/wk × £18 × 46 ≈ £22 to 32k (M) · Upside: faster quotes → higher win-rate (A, pilot-proven)
UPSIDE (to prove): same-hour quoting wins jobs that slow quotes lose today. On ~120 quotes/week even a few points of win-rate is real revenue, but win-rate is an assumption (A) until measured, so the pilot proves it on your own quotes before you count it.
Opportunity deep-dive · 2 of 2 shown
Supplier price-list ingestion
Every supplier price file turned into a checked, ready-to-import update, so Monday stops being repricing day, and quotes stop going out on stale prices.
How it works
Thousands of lines updated by hand each week; hours gone, and pricing errors that quietly cost margin on quotes.
The file is mapped, validated and checked in minutes; margin-critical moves are flagged; prices are current.
This one is pure HARD: time saved plus fewer pricing errors, all from your own files (M). No growth assumption; it simply gives the week back. The delivered report details every Wave-1 opportunity to this depth; this sample shows two.
The numbers
Shown as ranges, not single figures
The bars show the proven hard floor: the time saved you can bank. Growth upside is shown per opportunity and proven in the pilot, not rolled up here. The tick is our central, conservative estimate; the band is the plausible range.
| Wave | Net / yr | Build | Payback |
|---|---|---|---|
| W1 quick wins | ~£42 to 62k floor | ~£18k | ~4 mo |
| W2 scale | ~£15 to 22k floor | ~£12k | ~7 mo |
| W3 strategic | TBD | TBD | TBD |
| Total identified | ~£57 to 84k floor + upside to prove | n/a | n/a |
These bars show the PROVEN hard floor: time saved, fromKestrel's own volumes (46 working weeks, conservative coverage per opportunity). Growth upside (quotes won on speed, reorder revenue) is shown per opportunity above and proven in the pilot, not rolled up here. Every figure is source-tagged (D/M/B/A). Running AI costs are billed to Kestrel directly.
Technical & data readiness
What to tidy first, so the plan holds up
The trade system has an API that is switched off, and price data is inconsistent. That shapes the sequencing: the Wave-1 wins are chosen because they need only a data export to begin.
Strengths: an experienced counter team and a real order history. Gaps: a trade system whose API is switched off, and inconsistent price data.
Fix first
- Get a simple <strong>order & quote export</strong> (CSV). It unlocks the lead win with no integration.
- Turn on the <strong>trade-system API</strong> (read-only) so drafts can write back in Wave 2.
- Gather a handful of recent <strong>supplier price files</strong>, the grounding for the ingestion tool.
The roadmap
Sequenced, not dumped
Wave 1 starts immediately and needs only an export; the system integration runs underneath to unlock the later waves.
Order & quote desk · Price-list ingestion · Delivery-status, export-only to begin
Statement chasing · Product & substitute finder · Back-in-stock prompts
Demand & stock forecasting, once the sales history is joined up
Trade-system API & clean price data, the enabling work that unlocks Waves 2 to 3.
Security & governance
What each tool is allowed to touch
In plain English, not a compliance certificate. Nothing gets a staff login; each automation gets a narrow key you create and can switch off in seconds.
- The lead win needs no access to your trade system to start: it drafts from an order export and your price book, so there is nothing to break.
- Your data is not used to train AI. Models run on no-training terms; only the fields an automation needs ever leave your systems.
- Running costs sit on your own accounts. The AI and API usage are billed to you directly, no lock-in, no markup.
- A person approves every order and quote before it is sent. The AI drafts and prices; someone signs off. Every action is logged.
A live demo of the biggest quick win
Watch the order & quote desk at work
Not a slideshow: this is the assistantturning a real order or quote request into a drafted, line-by-line entry for your team to approve. Pick a scenario and watch it play out, including when it's unsure and hands off to a person.
Morning, can you price up 20× 110mm soil pipe (3m), 4× single branch and 10 brackets for collection Thursday? Cheers, Jonno
A staged walk-through for this page: the real build reads from yourfull price book and account rules. Seeing it handle your own orders is what turns 'sounds good' into a decision.
The full audit, walked through for a made-up firm: Harbourline Property, a residential lettings & management agency we invented for this page (~600 managed units). It's not a real client. Every number is built from a business's own figures and shows its working.
Executive summary
The prize, in four numbers
Harbourline's growth is capped by how much maintenance, arrears and compliance admin a small team can carry, all of it manual, repetitive and constantly interrupting. That load is a large, measurable time cost, and it's squarely in AI's wheelhouse.
Two kinds of number, kept apart. The HARD floor is time saved, from Harbourline's own maintenance, arrears and compliance volumes (46 working weeks, conservative coverage). You can bank it, and it pays for the build in aboutfour months on its own. The growth upside (fewer void days, better retention) is shown per opportunity and proven in the pilot, never bundled into the headline. Provenance on every figure: D your data · M measured from your systems · B benchmark · A assumption (flagged, pilot-proven).
What we found
Four themes, in their own words
From a short operations review and a look at the maintenance log, arrears report and compliance tracker, each grounded in a real number and a real quote.
Maintenance triage never stops
~120 maintenance issues a week arrive by call, email and portal, each needs triaging, details gathering, a contractor matching and the tenant keeping updated. It swallows the managers’ days.
“We spend more time coordinating repairs than managing properties.” (Property manager)
Arrears chasing is inconsistent
Chasing rent arrears is manual and stop-start; some cases slip, and the ones that slip get expensive.
“When we're busy, the arrears chasing is the first thing to drop.” (Accounts)
Compliance is tracked in a spreadsheet
Gas safety, EICR, EPC and PAT expiries live in a spreadsheet; a missed certificate is a legal and safety risk, not just admin.
“One missed gas safety and it’s not just a fine, it’s serious.” (Compliance)
The same tenant questions, all day
“How do I report a repair?”, “when’s my inspection?”, “can I get a reference?”, a steady stream of repeat questions on the inbox.
“Half the inbox is the same ten questions.” (Office)
Your business, as we understand it
Where your information lives today
The team cares and the tenants are looked after; the constraint is that maintenance, arrears and compliance all run through people re-keying between systems that don't talk.
Four disconnected islands: people are the integration. It is the single biggest limiter on cheap automation, and it shapes the sequencing below.
The good news: the highest-value wins don't need these joined up first. They run from exports and the inbox. That's why we lead with them.
The opportunity landscape
Every opportunity, ranked
Seven opportunities scored on impact (what it's worth) against feasibility (how quickly and cheaply we can ship it). Bubble size shows annual value, and the leaders are the ones that hand your managers their week back.
| # | Opportunity | What the AI does | Value / yr | When |
|---|---|---|---|---|
| 1 | Maintenance triage & contractor dispatch | Triages a tenant’s issue, gathers details/photos, matches an approved contractor and keeps the tenant updated | £24 to 34k hard · retention upside | Wave 1 |
| 2 | Compliance-certificate tracker | Tracks gas/EICR/EPC/PAT expiries, drafts renewal instructions, flags anything overdue | £8 to 12k hard + risk avoided | Wave 1 |
| 3 | Arrears chasing assistant | Runs polite, staged arrears comms and escalates strictly to your policy | ~£8 to 12k hard | Wave 1 |
| 4 | Tenant & applicant enquiry handling | Answers the repeat questions on email and portal; books viewings and inspections | ~£10 to 15k hard | Wave 2 |
| 5 | Renewals & rent-review prep | Flags upcoming tenancy ends and drafts the renewal / rent-review pack | ~£6 to 10k hard | Wave 2 |
| 6 | Void-to-relet accelerator | Checklist, marketing copy and scheduling to cut the days a property sits empty | ~£12 to 24k upside | Wave 2 |
| 7 | Portfolio & arrears forecasting | Predicts arrears risk and void exposure across the portfolio | strategic | Wave 3 |
Opportunity deep-dive · 1 of 2 shown
★ Maintenance triage & contractor dispatch
Take a tenant’s repair report, triage it, gather what’s needed, match an approved contractor and keep everyone updated, so your managers stop being a switchboard.
How it works
~120 issues a week, each hand-triaged and coordinated; managers interrupted all day; tenants left waiting for updates.
Routine issues triaged, matched and scheduled in minutes; managers handle only judgement calls; tenants kept informed automatically.
HARD (proven): your log shows ~120 maintenance issues a week (M), each taking ~20 to 30 minutes of coordination. Handling 45 to 60% end-to-end frees ~20 to 30 hours a week; at £20/hr over 46 weeks that is ~£24,000 to £34,000/yr you can bank. That is the floor, and it pays back the build in ~4 months on its own.
Hard: ~20 to 30 hrs/wk × £20 × 46 ≈ £24 to 34k (M) · Upside: retention → fewer lost managements (A, pilot-proven)
UPSIDE (to prove): faster, better-communicated repairs improve tenant retention and landlord satisfaction, fewer lost managements. That is an assumption (A) until measured, so the pilot tracks it on your own portfolio before you count it.
Opportunity deep-dive · 2 of 2 shown
Compliance-certificate tracker
Every gas, electrical, EPC and PAT certificate tracked, chased and renewed on time, turning a spreadsheet risk into a system that never forgets.
How it works
Expiries tracked by hand in a spreadsheet; renewals chased when someone remembers; a missed certificate is a real legal and safety exposure.
Every expiry watched; renewals drafted ahead of time; nothing lapses silently, and the risk is provably managed.
The hard number here is admin time saved (M); the bigger value is avoided risk: a lapsed safety certificate is measured in fines and liability, not hours. The delivered report details every Wave-1 opportunity to this depth; this sample shows two.
The numbers
Shown as ranges, not single figures
The bars show the proven hard floor: the time saved you can bank. Growth upside is shown per opportunity and proven in the pilot, not rolled up here. The tick is our central, conservative estimate; the band is the plausible range.
| Wave | Net / yr | Build | Payback |
|---|---|---|---|
| W1 quick wins | ~£40 to 58k floor | ~£17k | ~4 mo |
| W2 scale | ~£16 to 25k floor | ~£13k | ~7 mo |
| W3 strategic | TBD | TBD | TBD |
| Total identified | ~£56 to 83k floor + upside to prove | n/a | n/a |
These bars show the PROVEN hard floor: time saved, fromHarbourline's own volumes (46 working weeks, conservative coverage per opportunity). Growth upside (fewer void days, retention) is shown per opportunity above and proven in the pilot, not rolled up here. Every figure is source-tagged (D/M/B/A). Running AI costs are billed to Harbourline directly.
Technical & data readiness
What to tidy first, so the plan holds up
The property manager holds the data but exports awkwardly, and compliance lives in a spreadsheet. That shapes the sequencing: the Wave-1 wins need only exports and inbox access to begin.
Strengths: a caring team and good property records. Gaps: awkward exports from the property manager and compliance living in a spreadsheet.
Fix first
- Export the <strong>maintenance log & contractor list</strong>. It unlocks the lead win.
- Get the <strong>compliance spreadsheet</strong> into the tracker so nothing lapses in the gap.
- Agree an <strong>escalation policy</strong> for safety issues and arrears, the rules the assistants follow.
The roadmap
Sequenced, not dumped
Wave 1 starts immediately from exports and the inbox; the deeper system write-back runs underneath to unlock the later waves.
Maintenance triage · Compliance tracker · Arrears chasing, export & inbox to begin
Enquiry handling · Renewals prep · Void-to-relet accelerator
Portfolio & arrears forecasting, once the data is joined up
Property-manager write-back & a single tenant record, the work that unlocks Waves 2 to 3.
Security & governance
What each tool is allowed to touch
In plain English, not a compliance certificate. Nothing gets a staff login; each automation gets a narrow key you create and can switch off in seconds.
- The lead win needs no write access to your systems to start: it triages from an export and the inbox, and a manager approves every dispatch.
- Your data is not used to train AI. Models run on no-training terms; only the fields an automation needs ever leave your systems.
- Running costs sit on your own accounts. The AI and API usage are billed to you directly, no lock-in, no markup.
- A person approves anything that spends money or touches a tenancy. Safety issues escalate to a human immediately. Every action is logged.
A live demo of the biggest quick win
Watch the maintenance triage at work
Not a slideshow: this is the assistanttaking a real tenant maintenance report, triaging urgency, gathering what's needed and matching an approved contractor. Pick a scenario and watch it play out, including when it's a safety issue and escalates to a person.
Water's coming through the kitchen ceiling, and there’s a light fitting right where it's dripping.
Water near a light fitting is a safety issue: the console asked the tenant to switch that circuit off, flagged it urgent, and escalated to a manager to dispatch an emergency plumber now.
A staged walk-through for this page: the real build reads from yourcontractor list, property records and access rules. Seeing it handle your own reports is what turns 'sounds good' into a decision.
The full audit, walked through for a made-up firm: Ledgerwell & Co, a bookkeeping & accountancy practice we invented for this page (~180 SME clients). It's not a real client. Every number is built from a business's own figures and shows its working.
Executive summary
The prize, in four numbers
Ledgerwell's people are its product, but the biggest slice of their time goes on keying documents, chasing clients for records and tracking deadlines. That manual load is a large, measurable cost, and it's squarely in AI's wheelhouse.
Two kinds of number, kept apart. The HARD floor is time saved, from Ledgerwell's own document and client volumes (46 working weeks, conservative coverage). You can bank it, and it pays for the build in aboutthree months on its own. The growth upside (capacity for more clients, advisory work) is shown per opportunity and proven in the pilot, never bundled into the headline. Provenance on every figure: D your data · M measured from your systems · B benchmark · A assumption (flagged, pilot-proven).
What we found
Four themes, in their own words
From a short operations review and a look at the job log, document volumes and deadline tracker, each grounded in a real number and a real quote.
Document keying is the biggest cost
~4,000 documents a month (invoices, receipts, bank lines) read and keyed into the ledgers by hand. It is the single largest slice of chargeable-but-low-value time.
“Most of a junior’s week is typing in receipts and invoices.” (Practice manager)
Chasing clients for records eats the month
Every deadline means chasing clients for missing paperwork, repetitive, awkward and never-ending.
“We spend the last week of every month chasing missing records.” (Bookkeeper)
Deadlines are tracked by memory and lists
VAT, CT and Self-Assessment deadlines are tracked across spreadsheets; a slip means a client penalty and an awkward conversation.
“A missed filing is a penalty and a lost client’s trust.” (Partner)
The same client questions, every day
“What do you need from me?”, “when’s my VAT due?”, “how much should I set aside?”, a steady stream of repeat questions.
“I answer the same handful of questions ten times a day.” (Client manager)
Your business, as we understand it
Where your information lives today
The work is excellent; the constraint is that documents, client chasing and deadlines all run through people, because the tools don't talk to each other.
Four disconnected islands: people are the integration. It is the single biggest limiter on cheap automation, and it shapes the sequencing below.
The good news: the highest-value win doesn't need these joined up first. It runs from the document stream and your chart of accounts. That's why we lead with it.
The opportunity landscape
Every opportunity, ranked
Seven opportunities scored on impact (what it's worth) against feasibility (how quickly and cheaply we can ship it). Bubble size shows annual value, and the leader is the one that gives the practice its capacity back.
| # | Opportunity | What the AI does | Value / yr | When |
|---|---|---|---|---|
| 1 | Document-to-ledger processing | Reads invoices, receipts & bank lines and drafts categorised entries for review | £34 to 48k hard · capacity upside | Wave 1 |
| 2 | Client query & records chaser | Answers routine client questions and chases missing records automatically | £10 to 15k hard | Wave 1 |
| 3 | Deadline & filing tracker | Tracks VAT/CT/SA deadlines per client, drafts reminders and prep checklists | £6 to 10k hard + penalties avoided | Wave 1 |
| 4 | VAT-return prep assistant | Assembles the quarter and flags anomalies for the accountant to review | ~£10 to 16k hard | Wave 2 |
| 5 | Client onboarding & records collection | Guides new clients through providing exactly the right records | ~£5 to 9k hard | Wave 2 |
| 6 | Advisory-insight prompts | Flags clients who’d benefit from tax or cashflow advisory, new fee income | ~£10 to 20k upside | Wave 2 |
| 7 | Practice capacity & workflow forecasting | Predicts crunch periods and capacity across the practice | strategic | Wave 3 |
Opportunity deep-dive · 1 of 2 shown
★ Document-to-ledger processing
Turn the mountain of invoices, receipts and bank lines into categorised, review-ready entries, so your team checks and codes instead of typing.
How it works
~4,000 documents a month keyed by hand; most of a junior’s week on data entry; capacity capped by typing, not expertise.
Documents arrive pre-extracted and coded; the team reviews and posts; the same people handle far more clients.
HARD (proven): you process ~4,000 documents a month (M), at ~3 minutes each keyed and coded, about 200 hours a month. Automating 55 to 70% of the extraction and coding frees ~110 to 140 hours a month; at £22/hr over the year that is ~£34,000 to £48,000/yr you can bank. That is the floor, and it pays back the build in ~3 months on its own.
Hard: ~110 to 140 hrs/mo × £22 × 12 ≈ £34 to 48k (M) · Upside: more clients per head, no hiring (A, pilot-proven)
UPSIDE (to prove): the same team can then carry more clients without hiring, real growth capacity. But headcount leverage is an assumption (A) until measured, so the pilot proves the true time-per-document on your own mix before you count it.
Opportunity deep-dive · 2 of 2 shown
Client query & records chaser
Answer the repeat client questions and chase the missing paperwork automatically, so the last week of the month stops being a chase.
How it works
Routine questions answered one at a time; missing records chased by hand near every deadline; awkward, repetitive, never-ending.
Routine questions answered instantly; records chased automatically and politely; the team only handles the genuine queries.
Pure HARD: admin time saved, from your own query and chase volumes (M). No growth assumption; it simply gives the month-end back. The delivered report details every Wave-1 opportunity to this depth; this sample shows two.
The numbers
Shown as ranges, not single figures
The bars show the proven hard floor: the time saved you can bank. Growth upside is shown per opportunity and proven in the pilot, not rolled up here. The tick is our central, conservative estimate; the band is the plausible range.
| Wave | Net / yr | Build | Payback |
|---|---|---|---|
| W1 quick wins | ~£50 to 72k floor | ~£18k | ~3 mo |
| W2 scale | ~£15 to 25k floor | ~£13k | ~7 mo |
| W3 strategic | TBD | TBD | TBD |
| Total identified | ~£65 to 97k floor + upside to prove | n/a | n/a |
These bars show the PROVEN hard floor: time saved, fromLedgerwell's own volumes (46 working weeks / 12 months, conservative coverage per opportunity). Growth upside (more clients per head, advisory fees) is shown per opportunity above and proven in the pilot, not rolled up here. Every figure is source-tagged (D/M/B/A). Running AI costs are billed to Ledgerwell directly.
Technical & data readiness
What to tidy first, so the plan holds up
The accounts software holds the ledgers but documents arrive in every format under the sun. That shapes the sequencing: the Wave-1 wins need only the document stream and an export to begin.
Strengths: structured accounts software and a disciplined team. Gaps: documents arrive in every format, and deadlines live across spreadsheets.
Fix first
- Point the <strong>document stream</strong> (uploads, email, bank feeds) at one place. It unlocks the lead win.
- Confirm your <strong>chart of accounts & coding rules</strong> so drafts match how you post.
- Consolidate the <strong>deadline lists</strong> into the tracker so nothing slips in the gap.
The roadmap
Sequenced, not dumped
Wave 1 starts immediately from the document stream; the software write-back runs underneath to unlock the later waves.
Document-to-ledger · Client query & chaser · Deadline tracker, document stream to begin
VAT-return prep · Onboarding assistant · Advisory-insight prompts
Practice capacity forecasting, once the workflow data is joined up
Accounts-software write-back & one document store, the work that unlocks Waves 2 to 3.
Security & governance
What each tool is allowed to touch
In plain English, not a compliance certificate. Nothing gets a staff login; each automation gets a narrow key you create and can switch off in seconds.
- The lead win needs no write access to your accounts software to start: it drafts entries from the document stream for your team to post.
- Client data is not used to train AI. Models run on no-training terms; only the fields an automation needs ever leave your systems.
- Running costs sit on your own accounts. The AI and API usage are billed to you directly, no lock-in, no markup.
- A person reviews and posts every entry, and approves every filing. The AI drafts; someone qualified signs off. Every action is logged.
A live demo of the biggest quick win
Watch the document processing at work
Not a slideshow: this is the assistantreading a real invoice, receipt or bank line and turning it into a categorised entry for review. Pick a scenario and watch it play out, including when a transaction is ambiguous and it flags rather than guesses.
A staged walk-through for this page: the real build reads from yourown chart of accounts and client rules. Seeing it handle your own documents is what turns 'sounds good' into a decision.
This sample clears the 5× findings guarantee many times over, which is the point. We only take on work where the numbers are there.
See what your own numbers look like
Answer a few questions about how your business runs. We'll tell you honestly whether there's enough here to be worth doing.