Robin AI

UK-origin AI contract-review platform — once a leading challenger in mid-market legal AI, fragmented in late 2025 with managed services sold to Scissero and engineering team absorbed by Microsoft.

1. Core Product / Service

Robin AI built a contract-review and contract-intelligence platform: upload a contract → Anthropic-Claude-based engine returns redlines, risk flags, clause extraction, and obligation tracking [1][3]. Three tiers:

• Self-serve SaaS (Free / Pro / Enterprise) — in-app review, playbook config, SSO/security at the top tier.
• Managed services — Robin-employed lawyers running contract review at "Professional" (next-business-day) and "Executive" (4-hour) SLAs, on top of the AI.
• API + integrations — feeds into CLM systems for automated redlining at scale.

The product was an early Anthropic-Claude poster child; Robin's founder Richard Robinson publicly evangelized why Claude's long-context + safety profile fit contracts better than GPT-class models in 2023-2024 [4].

2. Target Users & Pain Points

• Buyer: In-house legal at mid-market companies + UK / EU law firms (Robin's home turf), volume-heavy contracting functions (procurement, sales).
• Pain points: Reviewing high-volume, low-complexity contracts (NDAs, MSAs, vendor agreements) chews up junior counsel time. Robin's wedge was AI-only review for the easy 80%, with managed services for the harder 20%.
• Geographic skew: UK + EU + APAC stronger than US, where Harvey dominates BigLaw.

3. Competitive Landscape

4. Unique Observations

• Pricing model: Hybrid — per-seat SaaS for the platform, plus per-volume / per-contract pricing for the managed-services layer. No public price list (the platform tier collapsed to "contact sales" by 2024).
• Implied $/1M tokens consumed: A contract-heavy in-house user processes ~50-200 contracts/month, each ~10-50k tokens. ~1-10M tokens/seat/month. With historical platform pricing in the ~$300-$600/seat/mo range, the implied retail rate was roughly $60-$300 per 1M tokens — significantly lower than Harvey's BigLaw price point. The managed-services layer carried a much higher implied per-token rate because it included billable lawyer time. Underlying Claude API cost (2026): ~$3-$15/1M. Retail-to-API markup on the platform tier: ~10-50×. The managed tier's economics looked like services/BPO, not SaaS.
• Moat type: Workflow / data on the contract-review surface, but no distribution moat in BigLaw and no model moat. Founder Robinson admitted in late 2025 the unit economics never worked: enterprise legal SaaS requires expensive sales cycles and deep implementation, and the human-in-the-loop services layer compressed margins below SaaS norms [1].
• Customer profile: Mid-market in-house legal + UK/EU law firms — explicitly not AmLaw 100. This left Robin sandwiched between Harvey above and Spellbook below.
• Markup multiple: ~10-50× on platform; near 1× (cost-plus) on managed services. The mix dragged blended margins toward services-business levels.
• Why it broke (2025): Robin reportedly had ~$10M ARR + a $16M pipeline but failed to close a $50M round in mid-2025; the human-in-the-loop services drag plus a weak BigLaw distribution motion priced them out of late-stage AI multiples [1]. Outcome: managed-services arm sold to Scissero (Dec 2025); engineering team absorbed by Microsoft (Jan 2026); standalone SaaS continuity uncertain.

5. Financials / Funding

• Total raised: ~$45M cumulative (Series A-B), pre-2025 wind-down [2][5].
• Notable rounds: Series B Jan 2024, $26M led by Temasek; Series A 2022, ~$10M led by Plural.
• ARR (mid-2025): ~$10M; pipeline ~$16M.
• 2025-26 outcome: Failed to close $50M growth round; managed services divested to Scissero (Dec 2025); core engineering acqui-hired by Microsoft (Jan 2026); standalone product status uncertain.
• Status flag: pivoted — historically active SaaS, currently fragmented across acquirers.

6. People & Relationships

• Founders:
  • Richard Robinson (CEO) — ex-Clifford Chance lawyer, vocal Anthropic-Claude proponent.
  • James Clough (CTO) — ex-physicist, ML lead.
• Investors: Temasek (lead Series B), Plural, QuantumLight, Episode 1.
• 2025-26 acquirers: Scissero (managed services arm); Microsoft (engineering team).
• Related wiki: harvey (BigLaw competitor), spellbook (SMB-tier competitor).
• Lesson noted in harvey / spellbook entries: in legal AI, the squeezed mid-market positioning between BigLaw-grade SaaS and SMB-grade Word add-ons proved structurally hard.