Eq.ip.ai — Equitable IP Intelligence for Collaboration

Elevator pitch

EQ.IP.ai is an AI co-pilot that teaches, automates, and audits IP across collaborations—turning messy ownership into fair, transparent value.

Why it was developed (Royal Free example)

High-stakes partnerships kept stalling on unclear rights, licences, and attribution. The Royal Free–DeepMind episode showed how promising tech can be derailed by gaps in transparency and governance. Eqip.ai exists to make fairness and compliance the default: teach the basics, assemble the right agreements, and tie contributions to a tamper-evident trail so teams can move fast without breaking trust.

Inspiration

Researchers, clinicians, SMEs, and public bodies want to collaborate—but templates, licences, and credit paths are confusing.
Valuable “little lights” (small insights, code, data prep) often go unrecognised.
We imagined a co-pilot that makes the right thing (clarity, consent, attribution) the easy thing—so collaboration accelerates and trust holds.

What it does

Educates: Plain-language, conversational explainers on IP/licensing tailored to role and context.
Automates: Guided agreement builder using sector templates (e.g., NHS/HDR), with version control and tracked sign-off.
Audits: Provenance ledger linking data → work → outputs → benefits; attribution reports you can share with partners and funders.
Integrates: Optional hooks for SharePoint, Data Safe Havens, GitHub, ORCID, and SDE audit logs.
Assures: Role-based access, DPIA-friendly logging, and exportable audit packs.

How we built it

Chat interface + RAG engine: Domain prompts + retrieval over curated IP guidance, policy, and template libraries.
IP ontology & policy pack: Normalised terms (ownership, background/foreground IP, licensing, benefit-sharing) mapped to decision trees.
Agreement composer: Structured Q&A that instantiates approved templates; captures assumptions and exceptions.
Provenance layer: Hashes key artefacts and decisions to an append-only ledger; optional IPFS/immutable store for audit proofs.
Interoperability: Lightweight APIs and webhooks for document stores, code repos, and identity (ORCID).
Governance & security: RBAC, least-privilege design, event logs, and privacy-by-design patterns (data minimisation, purpose binding).

Challenges we ran into

Ambiguity by default: Different partners interpret the same clause differently; we built explainers plus “compare view” to surface gaps.
Template sprawl: Reconciling multiple template families and licence variants without losing nuance.
Adoption vs. friction: Automate enough to help, while keeping a human-in-the-loop for legal sign-off.
Provenance granularity: Capturing just enough detail to be audit-ready without creating admin overload.
Procurement reality: Making the tool light-touch and integrable so it fits within existing Trust/university processes.

Accomplishments that we’re proud of

Click-through agreement builder that outputs clean, review-ready drafts from approved templates.
Attribution reports that trace contribution → output → impact in minutes.
Plain-language IP glossary co-designed with users (useful in training and onboarding).
Pilot design for an HDR UK SDE node with clear metrics (time-to-agreement, literacy uplift, audit readiness).
Configurable governance pack aligning roles, responsibilities, and benefit-sharing options.

What we learned

Education is the unlock: Short, contextual explainers reduce escalation and speed decisions.
People need visibility, not perfection: A good provenance trail beats a perfect one no one keeps.
Configurability matters: Partners want to tweak clauses and approval paths without starting from scratch.
Incentives drive behaviour: When attribution is visible, contributions increase—and meetings get shorter.
Trust is socio-technical: UX, policy, and process must land together, not as afterthoughts.

What’s next for Eqip.ai

HDR UK SDE pilot (6 months): Ship explainer, agreement builder, and audit dashboard; measure time-to-deal, literacy, and audit outcomes.
Engaging the developer community with ongoing management.
Template library expansion: Add common university–industry and charity-funded variants; internationalise licensing packs.
Deeper SDE integrations: Pull write-only activity logs for provenance; enrich reports with execution context.
Onboarding toolkit: Playbooks, micro-courses, and sandbox scenarios for rapid team uptake.
Fairness & impact analytics: Optional modules to visualise benefit flows and surface inequities in attribution.
Scale paths: From single node → multi-node networks; from health → broader research and data-trust consortia.