The Opportunity

Racing content is stuck in 1998.

UK & Irish racing wagering: £13B annual handle. Consumer racing media & tipping (our TAM): ~£250m. Both segments served by products that read, look, and measure the same way they did before the smartphone existed.

What We've Built

An AI-native racing publisher
with radical measurement.

Every piece of content — from a daily tip to a long-form Papers article — is generated by a structured LLM pipeline, rendered through distinctive editorial brands, and measured against actual race results in a public-facing ledger.

The Moat

One pipeline. Eight content products.

One data pipeline. One LLM contract. One prompt library. Multiple branded content formats. Measurement loop closes the flywheel.

See Appendix A for BoD + Pick Winner engine deep-dives →

Content Products

Eight products. Four brands. One engine.

Y1 growth focus Pick Winner · Bet of the Day · Race Papers. The other formats stay in maintenance mode until revenue mechanics prove out.

The Consumer Wedge

One interface. Every signal.
Zero cost to serve.

The DEN is Saturday Racing's chat-native concierge — a hybrid deterministic + agentic layer that flattens the whole platform into a single question: what do you want to know? Free Cubs see the signals. Paid Cubs ask LLaMa to investigate why.

The strategic direction: The DEN is the compound-interest engine on top of the content moat. Every new signal Saturday ships becomes a free-tier hook, a paid interrogation target, and a daily habit trigger — automatically. One consumer surface. Every product. Zero marginal cost of surfacing.

The Engineering

Production-ready platform, early market validation.

A production-grade platform engineered to enterprise standards. The metrics below aren't projections — they're what an acquirer, technical partner, or investor can inspect today.

The Defensible Edge

Continuous model validation is the moat.

Every tipping site claims accuracy. None publish the maths. Saturday Racing's measurement layer is the first honest track record in racing content — and it compounds with every settled race.

Token Economics Prompt caching + structured JSON + deterministic fallback keep per-pick LLM cost under 1p at 100k users — the AI margin story scales without server-cost surprises.

Where We Are

Behavioural engagement, zero marketing spend.

Saturday Racing has been built in public, launched quietly, and grown organically. Registration counts alone under-sell the story — the behavioural signals below are the real proof that the concept resonates.

Preliminary snapshot · GA4 verified before Q1 pitching

Platform Output / week ~250 races analysed · ~300 predictions published · £0 marketing spend · +22% MoM organic traffic · ~38% newsletter open rate

Revenue Streams

Five paths to revenue. Two ready to switch on.

The 24-Month Plan · Plus Optionality

Two years to earn Year 3.

This raise funds Year 1 and Year 2 execution. Year 1 turns readers into competitors; Year 2 turns competitors into subscribers. Long-term, the audience becomes a peer-to-peer liquidity pool — contingent on Y1–Y2 traction and follow-on capital.

See top callout for regulatory positioning. The long-term P2P optionality would require UKGC Betting Exchange licensing + dedicated compliance capability + a separate capital round — nothing about this raise depends on it.

The 3-Year User Path

Base 50k · Upside 100k · Exceptional 150k.

Starting from ~1,000 organic users, three catalysts drive step-change acquisition through Y1–Y3. Base case reaches ~50k users by end-Y3; the upside path reaches 100k with strong paid-acquisition efficiency; exceptional execution unlocks 150k+.

Base case assumes seed round closes Q1, Pick 6 launches Q3, subscription tier ships Q5, and organic + paid acquisition compound through Y2–Y3. Upside case requires best-in-class paid-acquisition efficiency + strong affiliate-driven conversion. Exceptional case additionally requires viral / editorial breakout moments.

Year 1 · Go-to-Market

Four channels. One acquisition engine.

Year 1 · Engineering Resource

Seven roles. One product line. Ready for 100k users.

The Field

We compete on a dimension nobody else does.

Incumbents lead on brand authority, distribution, and data depth. We lead on AI-native content architecture and measurement transparency — dimensions the incumbents can't easily replicate.

Distribution is our known gap — the Year 1 GTM plan on Slide 14 is the answer. The AI-native architecture + measurement moat are genuinely differentiated: difficult to replicate without significant investment and architectural change.

The Founder

Built solo. Now scaling deliberately.

Founder & CEO

Vas Rabani

18 moSolo build

70kLines of code

739Commits

2,688Automated tests

8Live products

£0External capital

30+ years immersed in horse racing — a punter's understanding of the market baked into every layer of the product. 10 years as a senior Product Manager with a strong engineering background that let him build Saturday Racing solo: architecting, coding, designing and shipping 8 branded AI-native products end to end. The rare founder who is his own domain expert, product lead and engineering team.

Senior Fractional CFO · 0.5 FTE

James Power

Senior finance leader and qualified accountant with a track record guiding early-stage founder-led startups from seed through institutional funding. Hands-on across cap-table discipline, board reporting, and investor governance. Focus at Saturday Racing: closing this seed round, monthly board packs from day one, and preparing the business for Series A.

Investment Round

£1.2m seed to fund Year 1.

The technology is built. This raise funds the team on Slide 15, the Year 1 GTM mix on Slide 14, and 18 months of runway to seed-to-Series-A metrics — strike-rate proof, 10k+ users, and 2+ bookmaker affiliate deals live.

[Founder / Team names + brief bios] · [Contact: name@saturday-racing.com] · [Website: saturday-racing.com] · Private and confidential. Not an offer of securities. Financial projections are indicative and depend on execution + market conditions.

Bet of the Day — Pipeline Deep-Dive

From ~500 runners to one bet in eight steps.

Eight deterministic steps, one calibrated LLM narrative. The moat is the loop — every bet becomes training data for the next.

Pick Winner — Pipeline Deep-Dive

Data ▸ Deterministic ▸ AI ▸ Guardrails ▸ Published.

Deterministic scoring picks the field. Claude picks the winner. Guardrails guarantee delivery. The fallback is the guarantee — every race gets a pick, every pick gets measured.

Horses to Watch — Story-Score Engine + Learning Loop

From today's declared runners to five worth watching.

The algorithm doesn't pick winners — it finds the five horses today with the strongest evidence across multiple independent signals, and tells the story. Empirically calibrated. Point-in-time correct. Continuously learning.

LLaMa Letters — Deterministic Scoring Engine → AI Editorial Analysis

The pick is calculated. The prose is written.

LLaMa is no longer the decision-maker — she is an editorial layer explaining a deterministic model. Every letter is measurable, testable, and auditable end-to-end. Deterministic scoring engine · AI editorial analysis · evidence-backed racing letters.