Honest answers to
common questions.

We begin with a scoping conversation — typically with a Head of Digital Policing or equivalent — to understand current workflows, volumes, and integration requirements. From there, we configure a force-specific deployment and run a supervised pilot. We don't do generic demos followed by a handover; onboarding is collaborative from the start.

Contact us via the contact page to start the conversation.

Evidentful provides an API-led integration layer for case management systems. Core export functionality is available now; force-specific integrations are configured and validated during onboarding.

If you have a specific CMS requirement, raise it early in the evaluation process so we can confirm compatibility and scoping.

Evidentful is designed for accessibility — it works in any modern browser on any device, including basic smartphones. No app download is required. Where a witness genuinely cannot access the platform, the traditional officer-led statement process remains available. Evidentful augments the system; it doesn't replace the fallback.

Each completed statement is automatically assessed across several dimensions: completeness (are all PEACE phases covered?), evidential coherence, specificity of detail, and consistency. The result is a quality score visible on the officer's dashboard alongside flagged gaps — specific areas where the interview may benefit from follow-up.

Quality scoring is an officer tool, not a gatekeeping mechanism. Officers review every statement regardless of its score. A high score does not bypass officer review; a low score prompts it.

Evidentful is designed to produce MG11-format witness statements, including the required declaration and witness review/signature process. As with any witness statement, admissibility and evidential use remain subject to the circumstances of the case, force procedure, and legal review.

The witness reviews, edits, and digitally signs the statement — the signed document is the legal instrument, not the AI draft. Evidentful is transparent about AI involvement. Every statement is clearly marked as AI-assisted, and the immutable audit trail (StatementStream) records every action from invite to signature.

Evidentful's primary deployment model is Azure UK South / UK West. For forces with stricter hosting requirements, the architecture has been designed to support containerised deployment, subject to technical scoping.

If on-premise or force-managed hosting is a requirement, raise it during initial conversations so we can assess feasibility and timeline.

Support arrangements, SLAs, and escalation paths are confirmed as part of the deployment agreement. We work closely with each force during the initial deployment period to monitor usage, address issues, and iterate on configuration. Ongoing support is included in the service agreement.

Specific SLA terms — including response times and availability commitments — are available on request.

The police officer handling your incident has sent you a secure link to give your statement through Evidentful. This approach lets you share what you know from wherever you are, at a time that suits you, without needing to travel to a police station or wait for an officer to be available.

Your statement is just as important as one given in person. The process is designed to help you tell your account fully and clearly.

Giving a witness statement is voluntary in most circumstances. The officer who contacted you can explain whether anything different applies in your situation. Choosing not to respond to the invite is a decision you can make — but your account may be valuable to the investigation.

Yes. Your statement is encrypted in transit and at rest, stored only in UK data centres, and accessible only to the officer and authorised supervisors associated with your incident. Evidentful is ISO 27001 certified and processes data in accordance with UK GDPR and the Data Protection Act 2018.

Your information will never be used to train AI models. It is shared only within the police force handling your case.

Yes. Evi conducts interviews in over 60 languages. When you begin the process, you'll be able to choose the language you're most comfortable in. Your statement will be clearly marked with the language used.

Before you sign your statement, you have the opportunity to review and edit it. Once you have signed it digitally, the signed statement is final — in the same way a traditionally signed paper statement would be. If you need to add information after signing, contact the officer in charge of your incident directly.

No. Evi conducts the witness interview and generates a draft statement. All decisions — about investigation, prosecution, or any other policing action — remain entirely with the officer. The officer reviews every statement, and nothing happens automatically as a result of what a witness says to Evi.

Evidentful analyses statement completeness and structure to help officers identify gaps. It does not assess credibility, determine evidential weight, classify offences for charging, or recommend investigative outcomes.

Yes — and this is precisely why officer review is a structural requirement, not an optional step. Evi generates a draft based on the witness's account. Every statement is marked as AI-assisted and reviewed by a qualified officer before it is treated as evidentially significant.

Witnesses also review and sign their own statement before it is finalised. If the draft does not accurately reflect what they said, they can correct it.

No. All AI provider agreements explicitly prohibit use of Evidentful data for model training, fine-tuning, or evaluation. Witness data is not retained on third-party AI infrastructure after inference.

Evidentful uses a LiteLLM abstraction layer to route AI calls — this means no single AI provider has persistent access to witness data, and the configuration can be audited by force IT security teams.

Evidentful uses leading commercial large language models, accessed via the LiteLLM proxy layer. Specific model selection is disclosed to forces during technical evaluation. The LiteLLM abstraction means Evidentful is not locked to a single provider — models can be swapped without architectural changes as the landscape evolves.

Model provenance is documented in alignment with the NCSP AI/LLM Cyber Standard v1.1.

Evi follows the PEACE framework as specified in College of Policing guidance: Preparation and planning, Engage and explain, Account, Closure, and Evaluation. This is the same interview structure used by trained police officers for witness and suspect interviews.

The interview is conversational and adaptive — Evi probes for detail, follows up on gaps, and closes systematically. It does not ask leading questions or apply pressure. The full PEACE methodology and how Evi implements it is documented on the AI & officers page.