Literature assistant for a research group
A private AI tool that searches the group's papers, preprints and notes, answers questions with citations and drafts structured summaries, while access stays restricted to approved members of the team.
Oxford organisations hold vast, specialist knowledge. We build the custom applications, secure data platforms and AI tools that help experts find, analyse and use it.
Oxford's economy runs on knowledge.
Its university and research institutes anchor a wide life sciences and health research community; academic and educational publishing has been part of the city for centuries; and Oxfordshire's science and technology campuses host work in space, energy, quantum and advanced engineering. Much of this activity depends on information that is complex, specialised and often sensitive — clinical data, unpublished research, manuscripts, rights records.
Generic software tends to fit badly. Researchers need tools that follow their methods; publishers need workflows that respect editorial judgement and licensing rules; life sciences teams need traceability that satisfies partners and regulators. Oxford organisations also carry a strong sense of stewardship over their data and reputation. What they look for is a technical partner willing to learn the domain properly before proposing anything, and to build systems that specialists will trust.
A private AI tool that searches the group's papers, preprints and notes, answers questions with citations and drafts structured summaries, while access stays restricted to approved members of the team.
A single record of licences, territories and permissions for every title and asset, linked to the production workflow so that nothing is published or syndicated without the right clearance in place.
A system that tracks samples from receipt through storage, processing and analysis, with barcode scanning, chain-of-custody history and links to results, replacing paper logs and shared spreadsheets. Auditors can follow any sample's full history in one place.
Our Oxford projects reward patience at the start. Discovery includes structured interviews with the specialists who will use the system, a review of the data and its governance, and a written summary of what we understood, which you correct before anything is scoped. A fixed-scope proposal in pounds follows. We then deliver in short cycles with a demo at each review, document as we build, and plan for handover to in-house research software engineers if you have them.
We start by understanding the goal, the constraints, and what success actually looks like — not just a feature list. You leave this phase with a clear scope and a plan.
We design the system and the experience before writing production code. The right architecture early is what lets the product scale later without a rewrite.
We ship in increments and demo often, so you see real progress every week. Tests, code review, and CI are built in — quality is not a phase at the end.
Before go-live we harden performance, security, and reliability, and wire in observability — so launch day, and the busiest day of the year, hold up.
We don't disappear at handoff. We operate what we build, monitor it, and keep improving it — because we treat your software like a product, not a project.
Yes, with the right design. We keep data in environments you control, choose model providers and settings that do not retain or train on your content, enforce access by role and log every query. Where risk is high, we keep humans responsible for final outputs.
Yes, as well as for institutes, publishers and companies. Grant-funded projects often need clear deliverables and documentation, which suits our discovery-first, fixed-scope way of working. We can also support your research software engineers rather than replace them.
UK projects are contracted through SMARTECHOR LIMITED, a UK-registered company, under a fixed-scope proposal priced in GBP that follows discovery. The proposal sets out deliverables, phases and acceptance criteria, so finance and procurement teams have everything they need to approve it.
Tell us about your goals and your team. We'll discuss the architecture and the most sensible way to deliver it.