Responsible AI use in Implest
A short guide for customer administrators who configure, supervise or authorise the use of Implest.
1. What Implest AI does
Implest can understand Missions, analyse business context, prioritise opportunities, generate and classify messages, interpret replies, qualify opportunities and help execute actions. AI can be wrong, miss context or present an inference as if it were a fact.
2. Your organisation remains responsible
Your company decides the commercial purpose, audience, data supplied, channel and applicable lawful basis. Authorising autonomous mode does not remove human accountability for configuration, supervision and use.
3. Data and minimisation
Use only the professional data necessary for the Mission. Do not provide passwords, API keys, tokens, payment credentials or unnecessary sensitive data. Public availability does not automatically make every use or communication lawful.
4. Copilot, supervised and autonomous
Copilot recommends or prepares. Supervised allows automation with review points. Autonomous may execute permitted actions without step-by-step human approval, but remains subject to compliance rules, limits, suppression and account controls.
5. Opt-out and suppression
If a person asks to stop, objects, unsubscribes or expresses an equivalent preference, contact must stop and suppression must be respected. Do not try to bypass a suppression by changing channel or account.
6. Transparency and identity
Where required, Implest may disclose that a person is interacting with AI. Do not use fictitious human identities as if they were real people or connect accounts without the account holder's authorisation.
7. Prohibited or separately reviewed uses
Do not use Implest by default for employment, credit, insurance, health, education, essential services, social scoring, biometrics or sensitive-data targeting. These cases require separate review and may be prohibited.
8. Human review and escalation
Stop and escalate when the channel's legal route is unclear, sensitive information appears, the system may materially mislead someone, a security incident may have occurred or the use falls outside intended B2B GTM.