Draft resource — work in progress
This is an exploratory, draft resource compiled as part of an internal research exercise. It does not represent the views, plans, endorsements, or official positions of the University of Sheffield. Content is subject to change and has not been formally reviewed or approved by the University.
This knowledge base is a comprehensive, point-in-time snapshot of the AI products and services landscape relevant to university research, giving the University the evidence base needed to make informed commissioning, support, and strategic planning decisions.
The landscape map serves as the foundational input to a proposed AI Catalogue — a curated, dynamic list of approved AI tools for researchers spanning University-licensed tools, reviewed public tools, and development environments for research and innovation. Where the catalogue will be a living, governed resource, this landscape map is the structured market-research exercise that makes the catalogue possible: understanding what exists, how tools relate to one another, and where the clearest opportunities lie for the University to act.
This is part of an attempt to balance the rapid “bottom-up” adoption by researchers (shadow IT) with the “top-down” requirements for governance, security, and strategic investment. The approach is to provide supportive resources, not mandates on AI usage.
Quick navigation
Capability Areas · Tools Index · Watch List · Reports
Scope
In scope: AI-enabled products and services that researchers and scholars across all faculties (Arts & Humanities, Social Sciences, Science, Engineering, Health) might realistically use in their day-to-day research work and that R&I IT would need to commission, license, or support. This includes:
- Standalone AI tools and platforms (e.g. literature review assistants, coding copilots, transcription and qualitative analysis tools, data visualisation and exploration tools, writing and editing aids).
- AI features or plugins embedded within tools the University already supports (e.g. AI capabilities in NVivo, SPSS, Adobe Creative Cloud, Microsoft 365 Copilot).
- Specialist research AI services with cross-disciplinary or faculty-specific relevance (e.g. image analysis, computational notebooks, lab automation assistants).
Out of scope: Underlying infrastructure (HPC, cloud GPU provisioning), foundational model APIs and SDKs aimed at developers or ML engineers, bespoke model training pipelines, and tools that sit exclusively within a single research group’s technical stack.
This map excludes commodity AI — i.e. generic LLMs like Google Gemini, which is already provided by IT Services.
This project is not the AI Catalogue itself, but the research that informs any such catalogue’s first edition. Neither is it an approval or procurement process — all tools identified as promising will still need to go through the University’s standard review, information security, and procurement workflows. It is not a technology strategy, but its findings should be a key input to one.
Objectives
-
Map the current landscape — Identify and catalogue the AI products and services available today that are relevant to university researchers, organised by capability area (e.g. literature discovery, qualitative coding, data analysis, writing support, image/media generation, research data management).
-
Reveal the functional network — Produce functional-cluster diagrams that show how tools relate to one another: where capability areas overlap, where tools complement each other, and where there are gaps or redundancies.
-
Identify low-hanging fruit — Highlight tools and capability areas that represent high-impact, broadly-useful, mature products — the clearest candidates for early adoption, commissioning, or expanded support.
-
Surface emerging opportunities — Identify capability areas or products not yet mature enough for broad adoption but showing strong potential, with an indicative timeline for viability.
Generative AI Principles
This landscape map assists the University in the proactive adoption of AI tools in research while adhering to our Principles for Using GenAI in Research and Innovation and Principles for the use of Generative AI for productivity purposes.
The map ensures adherence to data protection and accountability by including “Trust & Governance” as a core evaluation criterion, specifically assessing tools for explainability, information security, and compliance features.
The map addresses appropriateness and accuracy by categorising tools based on their maturity and specific research lifecycle utility — distinguishing, for example, between specialised chat tools that minimise hallucinations by grounding responses in academic databases, and more experimental autonomous laboratory tools.
Evaluation Criteria
Each tool is assessed against a consistent framework to enable like-for-like comparison and prioritisation. Every criterion maps to one or more structured fields in the tool note schema, making assessments machine-readable and comparable across the catalogue.
| Category | What we assess | Schema fields |
|---|---|---|
| Fit | Which research activities the tool supports; whether it serves a single discipline or many; which faculties benefit | capability-areas, generality, faculty-relevance |
| Maturity & timing | Production-readiness today; when broad adoption becomes realistic for emerging tools | maturity, viable-from |
| Impact & value | Potential effect on research outcomes; effort required to deploy and learn; quality of documentation and support resources | research-impact, ease-of-implementation, learning-curve, documentation-quality |
| Sustainability & cost | Licensing terms; indicative per-user cost; whether the University already holds a licence; ongoing operational and maintenance burden | licensing-model, institutional-pricing, cost-per-user-gbp, existing-university-license, ongoing-maintenance |
| Trust & governance | Data-training practices; GDPR compliance; data residency; SSO/SAML support; whether the tool explains or cites its outputs; published accessibility statement | data-training-policy, gdpr-compliant, data-residency, sso-supported, output-explainability, accessibility-statement |
| Strategic alignment | Confidence in vendor roadmap; Jisc/CHEST procurement availability; size and health of the user/developer community | vendor-roadmap-confidence, jisc-chest-available, community-ecosystem |
| Human oversight & fairness | All AI-generated outputs require verification by a staff member; known sources of bias are documented per tool | human-oversight-required, bias-risk-notes |
Capability Areas
Explore research capabilities and the tools available in each:
- Ideation and Grant Writing
- Literature Review
- Hypothesis Generation
- Data Collection
- Automatic Transcription
- Autonomous Laboratories
- Data Analysis
- Scholarly Writing
- Qualitative Coding
- Research Data Management
Intelligence
Reports and deeper investigations to support commissioning decisions: