Responsible use of prospect data and outbound workflows

Responsible use starts with knowing what was found, where it came from, what has been verified and what remains unknown. Keystone's workflow is intended to retain useful provenance, distinguish inferred classifications from facts and surface exceptions for review.

• ✓Source-aware company and contact research
• ✓Qualification against an agreed brief
• ✓Verification state where applicable
• ✓Unknowns remain unknown
• ✓Human review at acceptance and outreach boundaries

Subscriber treatment can differ between corporate subscribers and individual subscribers, including sole traders and certain partnerships. If identifiable professional contact data is processed, UK data-protection obligations can still apply in a B2B setting.

• ✓Distinguish corporate and individual subscribers
• ✓Assess the campaign purpose and necessity
• ✓Consider reasonable expectations and proportionality
• ✓Confirm the permitted channel before outreach

Keystone may research public business sources and retain source evidence where useful to the record. That does not turn every published detail into permission for every use. Fairness, transparency, purpose, minimisation and the customer's campaign context still matter.

• ✓Public company and website information
• ✓Source and collection context
• ✓Data minimisation for the agreed purpose
• ✓Privacy information and transparency handled appropriately

The workflow should distinguish verified facts, inferred classifications and unknowns. Verification, duplicate handling, qualification and suppression are controls with their own states; an AI-generated classification is not an email verification result.

• ✓Retained provenance where appropriate
• ✓Verification status and date where available
• ✓Duplicate and canonical identity handling
• ✓Issue flags and exception review
• ✓No invented accuracy or freshness promises

Suppression lists and objection handling should be maintained so a contact is not simply deleted and accidentally re-acquired later. Opt-out and objection routes should be clear, easy to action and reflected in the campaign controls.

• ✓Import existing customer suppressions
• ✓Check people, email, domain and company exclusions where agreed
• ✓Apply a final pre-send suppression check
• ✓Stop sequences on objection or opt-out
• ✓Keep suppression logic deterministic

AI may interpret source evidence and draft recommendations. Deterministic systems should handle verification, deduplication and suppression where appropriate. Outbound begins with human approval, and automation increases only after observed stable no-edit operation and explicit client approval.

• ✓AI drafts or recommends
• ✓Human reviews, edits and approves
• ✓Corrections feed evaluation and adaptation
• ✓Low-confidence cases stay review-gated
• ✓No full autonomy on day one

The broad flow can be described as discovery → website research → contact discovery and verification → qualification and review → suppression → CSV, Sheet, CRM or agreed delivery. The final privacy notice must reflect the actual vendors, processor roles, regions, retention periods and agreements in the deployed stack.

• ✓Separate live routes from configured but inactive fallbacks
• ✓Confirm controller and processor roles
• ✓Record hosting and data regions where relevant
• ✓Define retention and data-subject handling
• ✓Link detailed information to the privacy notice

Keystone can configure agreed process controls and surface exceptions. The customer remains responsible for confirming the lawful basis, permitted audience and channel, message claims, sector restrictions and any legal review needed for its campaign.