TroncPilot — AI tronc-master-in-a-box for UK indie pubs and restaurants

1. One-liner

AI tronc copilot for UK indie pubs and restaurants — drafts the policy, allocates tips, survives tribunal.

2. Trend signal — why now?

Three things converged in eighteen months and most indie operators have not caught up.

• The Employment (Allocation of Tips) Act 2023 went live on 1 October 2024. Every UK hospitality employer must now pass 100% of qualifying tips to staff, publish a written tipping policy, allocate fairly within one month of receipt, and keep 3-year records. Tribunals can fine up to £5,000 per affected worker. Conor Sheridan, Nory CEO, on record: “There is a growing concern that the Act may lead to higher menu prices… margins are already razor-thin, and with this new law in place, businesses can no longer rely on tips to plug the gaps.”
• The Employment Rights Act 2025 strengthening pack lands October 2026. Government consultation closed 1 April 2026. From October 2026 employers must consult workers (or union/worker reps) when developing or revising their written tipping policy and review it at least every three years, plus publish an anonymised written summary of staff views. That is brand-new compliance work nobody is set up for.
• Indies are already drowning. Excel-and-Casio is still the default at the small end. JustTip’s June 2025 explainer literally calls it out: “systems consisting of an excel spreadsheet and a casio calculator.” AccountingWEB threads from owner-operators ask things like “do I need a troncmaster” and what to do when the husband-and-wife director-shareholders cannot legally be the troncmaster. Peter Davies, MD of WMT Troncmaster Services, on the government’s guidance: “For those expecting further detail and clarity on the practical issues arising from trying to implement the new rules there will only be further frustration.”

Money has started moving. Grateful raised £1.5M seed from Calculus Capital in October 2025 specifically to scale tronc software for hospitality. TiPJAR charges £30/day-equivalent for single-venue plus a £99 setup, plus 4% on every transaction, plus £100/mo for the SuperTronc add-on that actually does the allocation. Buzzacott and WMT sell troncmaster-as-a-service to multi-site groups. Every one of those products is built for the operator who has a head of finance. Indies don’t.

Provenance:

• Signal 1: UK Tipping Act live Oct 2024 + Oct 2026 strengthening (worker consultation, anonymised summaries, 3-year review cycle); £5k/worker tribunal fines — gov.uk consultation, Make Work Pay strengthening pack — 2026-02-05
• Signal 2: Indie operators still on Excel + Casio; AccountingWEB threads of confused owner-directors asking who can be troncmaster; manual process burden — JustTip blog 2025-06-04, AccountingWEB Any Answers threads — 2025
• Signal 3: Grateful raised £1.5M seed (Calculus Capital) Oct 2025 for tronc; TiPJAR established at £60-130/mo + 4% txn for single sites; WMT/Buzzacott sell manual troncmaster service to chains — UKTN, BeBeez, TiPJAR pricing page — 2025-10-15 Category: Regulatory arbitrage

3. The opportunity

The market has been bifurcated. At one end, TiPJAR / Grateful / Toast tipping sit on top of EPoS hardware integrations and chase multi-site groups that can absorb £200/mo+ and 4% transaction fees. At the other end, Buzzacott / WMT / Moore Kingston Smith sell troncmaster-as-a-service — humans on retainer, £200–500/mo for a small operator, designed for groups that already have a finance director.

The 70,000-ish single-site indie pubs, restaurants, and cafés in between have nobody. They cannot afford either tier. They are running tronc on Excel or not at all, and the Oct 2024 Act + Oct 2026 strengthening just made “or not at all” a £5k-per-worker tribunal risk. The wedge is:

“We replace your tronc Excel and your accountant’s £200/mo tronc service. £49 a month, no transaction fee, written policy drafted in plain English, tribunal-ready records, and the new Oct-2026 worker-consultation runs itself. Connect your rota and your card-machine CSV — done in 10 minutes.”

AI is load-bearing here twice — drafting the fairness policy in language that survives a tribunal, and running the worker-consultation cycle (collect responses, anonymise, summarise) without the owner doing it.

4. Target market

• Primary customer: Owner-operator of an independent UK pub, restaurant, or café with 5–25 staff and 1–2 sites. Annual revenue typically £400k–£2M. They run payroll themselves or with a £30/mo small-business payroll tool, and rely on a high-street accountant for year-end.
• Why they buy: “I need a written tipping policy, I’m scared of the tribunal cap, I haven’t started the worker consultation that kicks in this October, my accountant just quoted me £250/mo to do tronc and I can’t afford that.” The Oct 2026 deadline is the trigger.
• Rough TAM reasoning: ONS reports 176,685 UK hospitality businesses (March 2025), 99.6% SME. 99,296 licensed venues. Filter to indies under 25 staff and you get a serviceable addressable market north of 70,000 venues. At £49/mo that is a ~£40M/year category — too small for a venture-scale incumbent, perfect for a bootstrapper who can take 5–8% of it.
• Why now for them: Oct 2026 Employment Rights Act tipping pack. The first time a worker formally requests their tip records under section 27I and the operator can’t produce them, the tribunal letter follows. That fear is loud right now.

5. Product sketch (MVP)

• Policy generator. Owner answers 15 questions about their venue. AI drafts a written tipping policy in plain English, mapped to the statutory Code of Practice fairness factors. Output is a PDF + posted-in-the-staffroom version + an employee-handbook insert.
• Allocation engine. Connect Square / Toast / EposNow / SumUp via CSV upload or basic API. Connect a rota tool (Deputy, RotaCloud, Planday) or upload a rota CSV. The engine runs the chosen weighting (hours × role multiplier × discretionary points) and produces a monthly allocation file ready for payroll.
• Tribunal-ready records. Every allocation, payslip line, and policy version is stamped and stored for the statutory three years. One-click export when an employee requests their statement.
• Worker-consultation runner. Anonymous staff survey link, AI summary of responses, dated record that the consultation happened. Templated for both the initial Oct 2026 rollout and the rolling 3-year review.
• Plain-English Q&A bot. Trained on the Act + Code of Practice + HMRC E24 guidance. Owner asks “can my head chef be the troncmaster” and gets the answer with a citation, not a £200 phone call to the accountant.
• Fairness sanity check. Flags allocations that look risky — e.g. front-of-house consistently getting <60% of the pool when policy says equal-by-hours. “This may be challenged at tribunal — here’s why.”

6. AI angle — what’s load-bearing

Two places, real work, no decoration.

• Policy drafting. The Code of Practice has fourteen-ish fairness factors. Translating an owner’s 15 answers into a venue-specific written policy that uses the right legal hooks is exactly the job an LLM does well. Without it the product is a £49/mo spreadsheet template — boring, copy-paste-able, dies on price.
• Worker-consultation summarisation. Oct 2026 onward employers must publish an anonymised written summary of staff views from the consultation. Doing that manually means reading 18 free-text responses, removing PII, and writing a one-pager. AI does it in 30 seconds and the output looks defensible.

Strip the AI out and you’re left with a tronc allocator. Useful, but indistinguishable from a spreadsheet template. The AI is the wedge.

7. Localization angle (if any)

UK-only by design. The Tipping Act, the Code of Practice, the Oct 2026 strengthening pack, the £5k tribunal cap, and the National Insurance treatment of tronc payments are all UK-specific. Selling “global tronc software” is a category error — there is no tronc outside the UK. That focus is a feature, not a bug; it lets the product be opinionated and the marketing copy be specific. Expansion later is into Ireland (similar service-charge culture, no act yet) and Australia (different framework — distant). Not in the first 18 months.

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

• Pricing:
  • Solo £29/mo — single site, up to 10 staff, monthly allocation, policy generator, records.
  • Standard £49/mo — single site, up to 25 staff, plus consultation runner, plus Q&A bot. This is the wedge tier.
  • Multi £79/mo per site — 2–5 sites, plus accountant export.
  • One-time setup and onboarding: free. (Indies hate setup fees; eat the cost in CAC.)
  • Optional: independent Troncmaster signoff add-on at £39/mo — partner with a chartered accountant, take a margin.
• ACV: ~£600/year average across mix. That’s £49/mo × 12 with some upgrade headroom.
• Rough math to $1M ARR (~£800k): 1,400 venues × £49/mo × 12 = £823k. Of 70,000 addressable venues, that’s 2%.
• Rough math to $5M ARR (~£4M): 6,500 venues × £49/mo or 4,500 venues × £75/mo blended (with multi-site upgrades + Troncmaster add-on). 9% market share of indie addressable. Tight but real.
• Expansion path: Multi-site upgrade as operators add a second venue. Troncmaster signoff add-on. Eventually a payroll-line export to Sage/Xero/QuickBooks Online (charged as Plus tier). Long-term: integrate with PAYE submission so we replace the customer’s current £30/mo payroll tool too.

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

Not “SEO and content.”

• AccountingWEB and the Tronc Troubleshooter PDF audience. Owner-directors who have asked “do I need a troncmaster” on AccountingWEB Any Answers in the last 18 months. ~200 named threads. Reply with a one-line answer + a Loom of TroncPilot drafting their policy from the thread details. Expect 8–10% conversion to demo, 30% demo-to-paid. ~5 customers from this alone.
• Cold email the 1,200 “small group” indie operators in the Restaurant Collective + NCASS member directories. Personalised — pull the venue name, the staff count, send them their draft tronc policy as a free deliverable, ask £49/mo to keep it live. 2% reply, 25% close = ~6 customers. Repeat monthly.
• The “Oct 2026 panic” content engine. Three SEO assets: “tronc policy template UK 2026”, “do I need a troncmaster small restaurant”, “Oct 2026 tipping consultation explained.” All ranked queries with thin existing answers. Paid distribution via UKHospitality and Restaurant Collective newsletters (£500–800 per blast). Drives 30–60 leads/blast.
• Accountant referral programme. 200 small-firm accountants currently turn down tronc work because it is fiddly and they can’t bill enough for it. White-label TroncPilot as their product, give them 25% rev share. They sell to their existing book. 30 firms × 5 clients each = 150 venues.
• One viral moment. First small operator to win a tribunal because they had TroncPilot’s audit trail = a press story we can put in front of every operator on the fence.

If two of these five fire at expected conversion, we hit 100 paying customers in 90 days.

10. Build complexity — justification

Medium. Off-the-shelf for almost everything: Next.js + Postgres + Stripe + Clerk for auth, GPT-4-class for policy + summarisation, plain CSV ingest for rota and EPoS. The non-trivial bits are (a) the allocation engine — pure deterministic logic, weighting + edge cases for variable hours, sick days, terminations — and (b) the audit-grade record store with versioned policy snapshots. A two-person team ships v1 in 12–14 weeks. PoS deep integrations (Toast, Square, EposNow APIs) come in v1.5.

11. Gating checklist

Gate	Pass?	Note
Legal in target market	✅	UK statutory framework explicitly contemplates third-party tronc software.
Ethical — no harm / dark patterns	✅	Tool literally helps workers get fair tips — pro-worker.
Market exists (evidence above)	✅	Tipping Act + Oct 2026 strengthening + £1.5M Grateful raise + paying TiPJAR base.
1–5 person team can build this	✅	Two-person team in ~14 weeks.
Launchable with <$50K / ₹40L	✅	Hosting + Stripe + a few API credits. <£15k to revenue.

12. Feasibility score

Axis	Weight	Score	Notes
Problem intensity	20	16/20	Real fines, real records demand, real worker requests. Not life-or-death but acute and dated. Operators ask their accountant about it monthly.
Demand evidence	15	12/15	TiPJAR paying base, Grateful £1.5M raise, AccountingWEB threads, multiple law firm explainer pages. Not yet a screaming consumer-style market — operators are confused, not demanding.
Build feasibility	15	12/15	Off-the-shelf stack. The allocation engine + audit log is the only real engineering. ~14 weeks for two people.
Distribution clarity	15	11/15	Named directories (NCASS, Restaurant Collective), named threads (AccountingWEB), named partners (small accountants). No mass channel. Will require sustained outbound.
Revenue mechanics	15	11/15	£49/mo is below the pain threshold for a 5–25-staff venue. Path to 1,400 venues for £1M ARR is plausible. £5M needs accountant channel to fire.
Time to first revenue	10	7/10	6–10 weeks. Indies decide fast on a £49/mo line item, especially with Oct 2026 in the air.
Defensibility	10	7/10	Soft moats: the policy library compounds, the accountant referral channel locks in once seeded, audit-trail data has switching cost. Grateful and TiPJAR can copy features but won’t drop their per-transaction model — that is the moat.
Total	100	76/100

13. Qualitative modifiers

Founder-fit tags

domain-expertise-required · technical-heavy

The product needs someone who can read the statutory Code of Practice, the HMRC E24 guidance, and the Employment Rights Act 2025 amendments and translate them into product. A hospitality-savvy ex-accountant + a technical co-founder is the ideal pairing. Solo technical founder works only with a paid tronc-advisor on retainer.

Key assumptions to validate (3–5)

1. Assumption: Indie operators (5–25 staff) will pay £49/mo for tronc compliance software when their current alternative is Excel or a £200/mo accountant retainer. How to test: 30 cold calls with a 1-page mock-up + Stripe checkout link. Target 10% LOI / pre-paid first month.
2. Assumption: The Oct 2026 worker-consultation requirement is a strong enough trigger to drive purchases without further sales effort. How to test: Run the same outreach split — half referencing Oct 2024 act only, half referencing Oct 2026 strengthening. Compare reply rates.
3. Assumption: Small accountants (1–5 partner firms) will white-label and resell rather than just refer. How to test: 20 accountant calls. Target 3 signed letters of intent.
4. Assumption: The allocation engine can handle real-world rotas (zero-hours, split shifts, last-minute drops) without manual fixup more than 5% of the time. How to test: Build engine on synthetic data, then run 5 real venues’ last 3 months of rota+tip data through it manually. Track failure rate.
5. Assumption: AI-drafted policies pass solicitor review at first attempt. How to test: Pay a hospitality-employment solicitor £500 to review 10 generated policies. Target ≥7 with no material edits.

Risk flags

1. Regulatory drift: The Code of Practice and consultation rules can shift between now and Oct 2026 implementation. The product must follow them in near-real-time. Mitigation: a chartered-accountant advisor on retainer (£200–400/mo) who reviews each Code of Practice update.
2. Incumbent moves down-market: Grateful’s £1.5M raise explicitly mentions “expanding into new markets” — they could launch a low-tier SKU. Mitigation: the per-transaction fee is core to their unit economics; it would cannibalise their existing book to drop it. Time advantage of 12+ months is realistic.
3. Channel concentration: Accountant referral channel could end up being the only one that scales. Mitigation: keep direct outbound active even at lower ROI to retain a non-partner-dependent acquisition path.

14. Structured verdict

Score:                  76/100
Verdict:                GO
Confidence:             Medium
Best-fit builder:       Hospitality-savvy ex-accountant + technical co-founder, UK-based
Time to revenue:        8–10 weeks from build start
Capital to launch:      £12-18k ($15-22k) — hosting, advisor retainer, first cohort onboarding
Top 3 assumptions to validate first:
  1. £49/mo willingness-to-pay (30 cold calls, target 10% pre-paid LOI)
  2. AI-drafted policies pass solicitor review (£500 spot-check, target ≥7/10 clean)
  3. Accountant white-label channel (20 calls, target 3 LOIs)
Kill criteria:
  - Abandon if <5% of 50 cold-outreach indies show meaningful interest after 4 weeks
  - Abandon if solicitor review rejects >3/10 generated policies as materially non-compliant
  - Abandon if Grateful or TiPJAR launches a sub-£50/mo, no-transaction-fee SKU before our v1 ships

15. Next step — 1-week validation sprint

Falsifiable.

• Day 1–2: Pull 100 indie hospitality operators from NCASS + Restaurant Collective + Companies House (under-25-staff filter). Build a one-page landing — “TroncPilot. Your tronc policy, allocation, and Oct 2026 consultation, done. £49/mo. Beta open.” Wire to a Cal.com 15-min slot.
• Day 3–4: Cold email 100, with a personalised “here’s a draft tronc policy for your venue” Loom. Pay a hospitality-employment solicitor (£500) to review one such draft policy and confirm it would survive a tribunal challenge.
• Day 5: Decision rule: ≥10 demos booked AND ≥4 paid £49 deposits via Stripe AND solicitor signoff with no material edits = GO. Anything less = re-shape or kill.