FeeRecoup

1. One-liner

Audits a UAE SME’s bank charges against the published fee schedule and drafts the Sanadak complaint.

2. Trend signal — why now?

On 17 February 2026 the Central Bank of the UAE issued Circular No. 2/2026 — the SME Customer Protection Regulation — replacing the 2021 SME Market Conduct Regulation. It comes into force 13 September 2026, nine days from today.

That regulation converts a pile of soft “treat SMEs fairly” language into dated, checkable duties a bank either met or did not:

• Written acknowledgement of a complaint within 2 business days; final written response within 30 business days.
• Minimum 60 calendar days’ written notice before any change to T&Cs including fees takes effect.
• 30 calendar days’ notice before automatic annual renewal of a contract, explaining how to cancel.
• No closing or penalty fee where an account has been open more than 6 months.
• No fee for original paper statements.
• Fees must “remain consistent throughout a customer’s lifecycle” and may not exceed CBUAE fee caps.
• Free account, product and data portability — and the SME need not explain why it is switching.
• A Key Facts Statement before execution, not after.
• Three-business-day account opening for low-risk SMEs, with an obligation to document and explain delays or rejections.

Every one of those is a binary the SME can test against its own statements. And the redress rail already exists and is free: Sanadak, the CBUAE-established ombudsman, accepts complaints from “a natural person, sole proprietor or small to medium sized enterprise”, at no cost, online — after the SME has raised it with the bank and waited 30 days.

Meanwhile the demand side has been screaming for a decade with no tooling. Gulf News reader complaints: a Dh2,700 dormant-account charge, charges surfacing five years later, customers “never told about charges, never contacted by email or phone” until they tried to close the account. Gulf News’ own banking coverage calls fee schedules “deliberately hidden in most cases,” with banks answering that the charges are in the agreement the customer signed.

Today the only people who actually recover this money are law firms running forensic audits to prove how much a bank overcharged over the life of a loan, recovering millions for clients. That is a lawyer’s engagement fee for a service the average SME’s dispute is worth a fraction of. Priced out, not unserved.

Provenance:

• Signal 1 (demand): UAE SMEs report undisclosed and retroactive bank charges — Dh2,700 dormant-account fee, charges surfacing after five years, no notification — and law firms sell forensic bank-overcharge audits recovering millions, i.e. the work is real but lawyer-priced — https://gulfnews.com/uae/reader-complaints/uae-reader-complains-against-bank-charges-of-dh2700-for-a-dormant-account-1.1654870066510 and https://alsaadiadvocates.com/hidden-bank-charges-mohammad-alsaadi-advocate/ — observed 2026-09-04
• Signal 2 (feasibility): Sanadak accepts free online complaints from SMEs against any CBUAE-licensed institution once the bank has had 30 days; and banks must publish up-to-date fee schedules publicly online and file them with CBUAE — so both the claim rail and the reference data are open — https://sanadak.gov.ae/en/make-a-complaint/complaint-eligibility/ and https://rulebook.centralbank.ae/en/rulebook/article-5-business-conduct — observed 2026-09-04
• Signal 3 (economic): CBUAE Circular No. 2/2026 (SME Customer Protection Regulation), issued 17 Feb 2026, in force 13 Sep 2026, imposes dated fee, notice, complaint-timeline and portability duties on every licensed bank — while the only fee-audit software that exists (Trovata, Redbridge HawkeyeBSB, Nomentia, Kyriba at $100K+/yr) is built for corporate treasury departments — https://cms.law/en/are/legal-updates/new-sme-customer-protection-regulation-marks-a-shift-in-uae-central-bank-approach and https://www.redbridgedta.com/us/bank-fee-analysis-software/ — observed 2026-09-04 Category: Regulatory arbitrage

3. The opportunity

The bank fee-audit category exists and works. It is simply aimed at the wrong wallet. Trovata, Redbridge’s HawkeyeBSB, Nomentia’s Bank Fee Analyzer and Kyriba all do line-item billing validation against contracted rates — for corporate treasury teams, with weeks of onboarding and enterprise pricing that reportedly runs $100K+ a year at the top end. A Dubai trading company with four accounts and Dh180,000 of annual bank charges is invisible to all of them.

Below that line sit two non-products. The first is the law firm’s forensic audit — accurate, effective, and priced so that only a six-figure dispute justifies it. The second is a free blog post: a dozen UAE consultancies hand-maintain “best business bank account 2026” comparison tables listing RAKBANK at AED 50–99/month with AED 10–20K minimum balance, Emirates NBD at AED 200–250 with AED 50K, ADCB at AED 150 with AED 25K, FAB at AED 250+ with AED 50K. Those tables are lead magnets for company-formation agents. They are static, they compare headline prices only, and they never look at what your bank actually charged you last month.

The gap is the diff: what the published schedule says × what the regulation now requires × what actually hit your account. Nobody sells that to an SME, because the incumbents’ economics start an order of magnitude above this customer and the law firms’ economics start two orders above.

What makes it a business rather than a spreadsheet is that from 13 September the diff is actionable. Before the regulation, “the bank charged me a fee I didn’t expect” was an argument. After it, “the bank raised this fee without the 60 days’ written notice Article requires” and “the bank charged a closing fee on an account open 14 months” are assertions with a defined remedy path and a free ombudsman at the end of it. The product turns a fuzzy grievance into a filed claim with a citation.

The second, quieter revenue line: the same diff engine that finds an overcharge also tells the SME what it should be paying elsewhere. Portability is now free and the SME is not required to justify switching. That makes the switch recommendation credible for the first time.

4. Target market

Primary customer: The owner or finance manager of a UAE SME — mainland LLC or free-zone company, 5–150 staff, AED 5M–150M turnover, typically 2–6 bank accounts across one or two of RAKBANK / Emirates NBD / Mashreq / ADCB / FAB / Wio, in trading, logistics, contracting, professional services or F&B. There is usually one person who is “the finance person” and they are not a treasurer; they are an accountant with a bookkeeping package and a bank portal login.

Why they buy: Because they have the specific, recurring experience of finding a charge they cannot explain and having no leverage to argue it. In their words, from the public record: charges “were never told to us, they never contacted me by email or phone”; charges surfacing “after 5 years”; a Dh2,700 dormant-account bill. The Gulf News banking desk describes fee schedules as “deliberately hidden in most cases,” with banks replying that everything was in the signed agreement. The SME’s current workaround is to call the relationship manager, get told it is standard, and absorb it.

Rough TAM reasoning: SMEs are the overwhelming majority of UAE companies and the regulation is aimed squarely at them. I do not need a big share. At AED 400/month, 1,000 paying SMEs is roughly AED 4.8M (~$1.3M) ARR. There are far more than 100,000 banked SMEs in the UAE; I need a low-single-digit-thousand count at the ceiling of this plan. The constraint is distribution, not market size.

Why now for them: Two dates. 13 September 2026, when the duties bind. And the following 12 months, during which every bank in the country will be repricing and re-papering SME accounts to comply — which is exactly when notice periods get missed and fee tables get changed. The highest-yield year for this product is the first one.

5. Product sketch (MVP)

• Connect or upload statements for each business account (PDF/CSV export from the bank portal covers every UAE bank on day one; no open-banking dependency required to launch).
• Every debit that is a bank charge is extracted, classified and named in plain English — minimum balance non-maintenance, dormancy, remittance/SWIFT, FX markup, cheque return, chargeback, statement, closure.
• Each charge is checked against that bank’s currently published schedule of fees and against the SME CPR duty set: was it in the schedule, was it at the scheduled rate, was 60 days’ notice given before it changed, is it a closure fee on an account over 6 months old, is it a paper-statement fee.
• A running recoverable balance — the total of charges that fail a check — with the reason and the statement line for each.
• One-click complaint pack: a dated letter to the bank citing the specific charge, the schedule version and the regulation duty; then, if 30 days pass without a satisfactory answer, the Sanadak submission built from the same file, in English and Arabic.
• Notice watch: monitors each bank’s published fee schedule for changes and tells the SME the date a new fee becomes lawfully chargeable — so the SME knows before the charge lands rather than after.
• Switch view: what the same 12 months of activity would have cost at the other major UAE business accounts, with the minimum-balance opportunity cost included.
• Timeline tracker: for each complaint filed, the 2-business-day acknowledgement and 30-business-day response clocks, with the Sanadak escalation unlocking automatically when the bank runs out of time.

6. AI angle — what’s load-bearing

Two places, both genuinely load-bearing.

First, normalisation. UAE bank statements are inconsistent PDFs with truncated, abbreviated, sometimes bilingual narration strings — “CHRG MIN BAL NON MAINT”, “SWIFT CHG OUR”, “SERV CHG Q3”. Mapping thousands of those variants onto a canonical fee taxonomy, per bank, is exactly the fuzzy-matching-at-scale job that was a data-labelling project two years ago and is now a cheap model call. Published fee schedules are the same problem in reverse: multi-page PDFs, updated without changelogs, that must be parsed into a comparable rate table and diffed version over version. Remove the model and you are back to a human building a per-bank parser and re-checking it every quarter — which is precisely why nobody has bothered for a market this size.

Second, drafting. The output is a complaint that cites the charge, the schedule clause and the regulatory duty, in register, in two languages. That is the step that converts a spreadsheet finding into something the customer will actually send — and it is the step the law firms currently charge for.

If you strip the AI out, this is a manual bookkeeping service. The AI is what makes the unit economics work at AED 400/month instead of a lawyer’s retainer.

7. Localization angle

This is a UAE-first product and the localization is the moat, not decoration.

• Arabic + English output. Bank correspondence and Sanadak filings work in both; the complaint pack must be credible in Arabic.
• The rail is local. Sanadak has no analogue elsewhere in the GCC in this exact form, and its SME eligibility plus zero filing fee is what makes small claims economically worth pursuing.
• The fee vocabulary is local. Minimum-balance non-maintenance is the single most complained-about UAE bank charge and barely exists as a category in US/EU treasury tools.
• Local price point. AED 350–600/month is normal for a UAE SME software line item and is roughly a tenth of one avoidable monthly charge stack.
• Adjacent expansion is regional, not global. Saudi (SAMA) and Bahrain run comparable conduct regimes with their own ombudsman equivalents; the same engine re-points with a new rulebook and fee-schedule corpus. Going to the US instead means competing with Trovata, which is a bad trade.

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

• Pricing: AED 399/month ($109) for up to 4 accounts and 2 banks — Standard. AED 899/month ($245) for up to 12 accounts, multi-entity, Arabic filings and notice-watch alerts across all banks — Pro. Plus an optional 20% success fee on recovered amounts for SMEs that prefer to pay nothing up front; this converts the sceptics and is the acquisition wedge.
• ACV: AED 6,000 ($1,630) blended, before success fees.
• Rough math to $1M ARR: ~620 SMEs at the blended AED 6,000 ACV ≈ AED 3.7M ≈ $1.0M. Achievable at roughly 15 net new customers a month for three and a half years, or faster with one good accounting-firm channel partner.
• Rough math to $5M ARR: ~3,100 SMEs, or ~1,800 SMEs plus a meaningful success-fee book plus one GCC market (Saudi) opened. Realistically this needs the accountant channel to be the primary engine rather than direct sales, and a second regulator’s rulebook loaded.
• Expansion path: accounts → entities → banks → jurisdictions. A group with five free-zone entities is five times the Standard plan. The genuinely attractive expansion is the accounting-firm reseller seat: a 30-client Dubai accounting practice paying per client under one login, which turns one sale into thirty.

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

• The free audit as the entire top of funnel. Upload 12 months of statements, get the recoverable balance and the top three failed checks free; pay to get the complaint pack and the ongoing watch. This is the only pitch that works, because the SME does not believe there is money there until it sees the number from its own statements. Target: 500 free audits, ~15% conversion.
• Company-formation and accounting agents are the list, and they are already publishing. A dozen UAE consultancies hand-maintain “best business bank account” comparison tables (Meydan FZ, RadiantBiz, JB Consultants, ProServices, Business Setup Experts and others) as lead magnets. They have the SME relationships, they have the banking conversation already, and their tables are static. Offer white-labelled audits as their content upgrade and a revenue share on conversions. Twenty partners × 5 clients each = the first 100.
• Free-zone SME communities. DMCC, IFZA, Meydan Free Zone and SHAMS run member events, newsletters and WhatsApp groups with thousands of member companies. A “what the 13 September regulation entitles you to” session is a legitimately valuable, non-salesy talk with a hard call to action. Three sessions × 60 attendees × 10% = ~18 customers.
• The regulation itself is the content engine. A public, maintained tracker of every UAE bank’s published fee schedule and every dated change — with the date each new fee becomes lawfully chargeable under the 60-day rule — is a thing that does not exist, that finance people will bookmark, and that ranks. It also doubles as the product’s own reference data.
• The complaint cohort. Public reader-complaint columns and bank review threads name specific SMEs and specific charges. That is a small, warm, directly addressable list for the first twenty conversations.

10. Build complexity — justification

Low. Statement ingestion is PDF/CSV parsing plus model-based classification — no open-banking licence, no bank integration, no regulatory permission, because the customer supplies its own data and the product never touches money. The fee-schedule corpus is public documents, scraped and diffed. The complaint pack is templated generation over a rules table. Nothing here is novel; the work is in the per-bank fee taxonomy and keeping the schedule corpus current, which is grind, not risk. A technical founder plus a part-time UAE banking-domain advisor ships a credible v1 in 8–10 weeks, covering the four or five banks that hold most SME accounts.

The one non-trivial piece is the rules table encoding the SME CPR duties precisely enough that a generated complaint does not embarrass the customer. That is a legal-review line item, not an engineering one — budget one advisory engagement.

11. Gating checklist

Gate	Pass?	Note
Legal in target market	✅	The SME is the customer and files in its own name; the product is analysis and drafting, not regulated financial advice or debt collection. No CBUAE licence required to help a company read its own statements. Success-fee model should be reviewed against local rules on claims-handling before launch.
Ethical — no harm / dark patterns	✅	Helps small businesses enforce rights a regulator just granted them. The failure mode to avoid is spamming banks with weak claims — the product must suppress low-confidence findings, not inflate the recoverable number.
Market exists (evidence above)	✅	Documented complaints, law firms selling the manual version, enterprise software selling the corporate version.
1–5 person team can build this	✅	Solo technical founder + domain advisor.
Launchable with <$50K / ₹40L	✅	Realistically $15–25K including legal review of the rules table and the Arabic drafting quality pass.

12. Feasibility score

Axis	Weight	Score	Notes
Problem intensity	20	16/20	Real money, felt monthly, and now with a remedy attached. Docked because for many SMEs the annual number is an irritant (low thousands of AED) rather than hair-on-fire — the pain is sharp but the amount is often modest, and some owners will shrug rather than fight their bank.
Demand evidence	15	12/15	Strong and independent: documented complaint history, lawyers monetising the manual version, an enterprise software category proving the job-to-be-done. Docked because there is no direct evidence UAE SMEs will pay a subscription for it — the willingness-to-pay is inferred from adjacent proof, not observed.
Build feasibility	15	13/15	Off-the-shelf throughout, no licence, no bank integration. Docked for the per-bank fee taxonomy grind and the Arabic-quality bar.
Distribution clarity	15	11/15	The agent/consultancy channel is named, real and already publishing on this exact topic, and the free-audit wedge is self-demonstrating. Docked because channel-partner motions are slow to start and the direct-to-SME conversion rate is a guess.
Revenue mechanics	15	11/15	Pricing sits comfortably in UAE SME software norms and the value is denominated in recovered dirhams. Docked because the success-fee line is unproven and churn risk is structural (see risks — the SME may cancel once the backlog is cleaned out).
Time to first revenue	10	8/10	The free audit converts inside one meeting when the number is good. First paying customer plausible 4–6 weeks post-launch.
Defensibility	10	5/10	Execution moat plus an accumulating fee-schedule history corpus that gets more valuable each quarter — a competitor arriving in month 12 cannot retroactively prove what a bank’s schedule said in March. But the core is copyable and a UAE fintech or an accounting-software vendor could bolt it on.
Total	100	76/100

13. Qualitative modifiers

Founder-fit tags

technical-heavy · domain-expertise-required

The build is straightforward; the credibility is not. This needs someone who can get a UAE banking or regulatory-compliance advisor onto the cap table or a retainer, and who can operate in Arabic-language business contexts. A pure offshore build will produce complaint letters that get ignored.

Key assumptions to validate (3–5)

1. Assumption: A typical UAE SME’s 12 months of statements contain at least AED 3,000–5,000 of charges that fail a defined check, often enough to justify a subscription. How to test: Get 20 SMEs to hand over 12 months of statements under NDA. Audit by hand. If the median recoverable is under AED 2,000, the pricing collapses and this becomes a one-off service, not SaaS.
2. Assumption: Banks actually reverse charges when presented with a citation-backed complaint, at a meaningful rate, without escalation. How to test: File 10 real complaints across 4 banks in month one and measure the reversal rate and time-to-resolution. Under ~30% reversal and the product’s promise is Sanadak escalation, which is slower and less saleable.
3. Assumption: SMEs will pay a recurring fee rather than treating this as an annual clean-up. How to test: Offer both a one-off audit and a subscription at launch and watch which sells; measure month-4 retention on the first cohort.
4. Assumption: Company-formation and accounting agents will white-label rather than build or ignore. How to test: Pitch 15 of the consultancies currently publishing bank-comparison tables. Three signed pilots inside six weeks is the pass mark.
5. Assumption: Published fee schedules are complete and current enough to be a reliable reference. How to test: Pull every major UAE business-banking schedule now, re-pull in 60 days, and check whether observed charges reconcile to published rates. If schedules are systematically incomplete, half the checks become unprovable.

Risk flags

1. Churn by design: The first audit clears years of backlog and delivers most of the lifetime value in month one. If the ongoing notice-watch and switch-monitoring is not genuinely useful, this is a one-off service wearing a subscription costume. This is the single biggest threat to the model and should shape v1 — the recurring surface must be built, not bolted on later.
2. Bank relationship risk: SMEs depend on their bank for credit. Some owners will not pick a fight over AED 4,000 with the institution holding their facility. Expect a real segment to opt out for reasons that have nothing to do with the product’s accuracy, and expect banks to notice a tool that files complaints at volume.
3. Regulatory-interpretation risk: The duties are new and untested. Early complaints will surface CBUAE and bank interpretations that differ from the rules table, and being confidently wrong in a customer’s complaint letter is a brand-killing failure mode. Ship conservative — suppress marginal findings.
4. Platform dependency (soft): Statement formats and fee-schedule publication practices are controlled by the banks. A format change breaks ingestion; a bank that stops publishing a granular schedule blinds a whole check class.
5. Fast-follow risk: A UAE accounting-software vendor or one of the neobanks (Wio, Mashreq NeoBiz) could ship a version of this as a switching lure. A neobank has every incentive to prove what your incumbent bank overcharged you.

14. Structured verdict

Score:                  76/100
Verdict:                GO
Confidence:             Medium
Best-fit builder:       Technical founder based in or well-connected to the UAE,
                        paired with a banking-compliance advisor; Arabic capability
                        in the team, not outsourced
Time to revenue:        6–10 weeks from launch
Capital to launch:      $15–25K (AED 55K–92K)
Top 3 assumptions to validate first:
  1. Median recoverable balance ≥ AED 3,000 per SME per year — hand-audit 20 sets
     of real statements under NDA before writing code
  2. Banks reverse citation-backed complaints at ≥30% without escalation — file 10
     real complaints across 4 banks and measure
  3. Consultancies publishing bank-comparison content will white-label — pitch 15,
     require 3 signed pilots in 6 weeks
Kill criteria:
  - Abandon if the median recoverable balance across 20 hand-audited SMEs is under
    AED 2,000/year — the subscription cannot be justified
  - Abandon if fewer than 3 of 10 filed complaints produce a reversal or credit
    within 60 days
  - Abandon if month-4 retention on the first paying cohort is under 60% — that
    confirms this is a one-off service, not SaaS
  - Abandon if a UAE neobank or accounting-software vendor ships an equivalent
    free audit before v1 reaches 50 paying customers

15. Next step — 1-week validation sprint

• Day 1–2: Pull and parse the current published schedule of fees for the six banks holding most UAE SME accounts (RAKBANK, Emirates NBD, Mashreq, ADCB, FAB, Wio). Build the rules table for the SME CPR duties that are objectively checkable from a statement: closure fee after 6 months, paper-statement fee, fee changed without 60 days’ notice, fee charged above published rate. This is the reference asset and it is needed regardless of the outcome.
• Day 3–4: Recruit 20 UAE SMEs — via two free-zone WhatsApp communities and three accounting-firm contacts — to share 12 months of statements under NDA in exchange for a free hand audit. Do the audits manually. Record, per SME: total bank charges, total that fail a check, and the single largest failed charge.
• Day 5: Take the five SMEs with the largest recoverable balances and ask for the order: AED 399/month, we file the complaints and watch the schedules. Also file two real complaints to start the reversal-rate clock.

Falsifiable outcome: median recoverable balance across the 20 audits, and paid conversions out of 5 asks. Go if median recoverable ≥ AED 3,000 and ≥2 of 5 convert to a paid plan. Anything less and this is a lawyer’s service line, not a bootstrapped SaaS — and I should walk.