Responsible intelligence
Data & AI Transparency
Emi scores are indicators, not determinations.
R&D probability scores and recommendations are generated from available evidence to support prospect prioritisation. They may be incomplete or incorrect and should be reviewed by a person before commercial, legal or tax decisions are made.
What Emi analyses
Depending on the workflow, Emi may use Companies House information, company status, SIC codes, incorporation details, officers, official website content and relevant external technical evidence.
What Emi produces
Outputs can include an R&D probability score, confidence level, evidence summary, likely R&D activities, likely projects, risk factors, current-activity signals and recommended next actions.
What a score means
A score is an evidence-based prioritisation signal. It is not confirmation that a company conducts qualifying R&D, has made an R&D tax claim, owns a patent or will qualify for any tax, grant or legal treatment.
Evidence quality
Public sources can be incomplete, stale or ambiguous. Website discovery can identify the wrong organisation, technical sources may relate to similarly named entities, and absence of evidence is not necessarily evidence of absence.
Human review
Users remain responsible for reviewing outputs, checking important source material and deciding whether a prospect should be researched, contacted, qualified or excluded.
People and profiling
RND Emi is designed primarily to assess companies. If identifiable people are included in prospecting activity, users should avoid unnecessary personal profiling and should not use Emi to make solely automated decisions with legal or similarly significant effects on an individual.
Data minimisation
Store only the information needed for the prospecting purpose. Do not add sensitive, private or special-category personal data simply because it is available elsewhere.
Corrections and objections
Where an individual challenges inaccurate personal data or objects to marketing, users should have a process to correct the record, stop inappropriate processing and retain only any minimal suppression information that is genuinely needed.
Evidence hierarchy used in practice
- 1. Official company data: Companies House and other authoritative company records.
- 2. First-party company evidence: official websites, technical pages, product information and company publications.
- 3. External technical evidence: relevant public patent, grant, research, publication or collaboration signals where available.
- 4. AI inference: Emi's interpretation of the available evidence, shown with a score and confidence rather than represented as a verified fact.
Last reviewed: 30 August 2026. This notice should be updated when RND Emi's data sources, scoring framework or AI/infrastructure providers materially change.