AI Blockers
The reasons a CRM contact is not safe for an AI agent to act on
An AI Blocker is a specific, checkable reason a CRM contact is not safe for an AI agent or automated outreach to act on. Not a vague quality score. One concrete problem, mapped to the failure it causes. This is the full taxonomy, and what Datuma checks for each one.
100 contacts · 5 minutes
Your AI does exactly what you tell it, on data you never checked
An AI SDR, a voice agent, or an automated sequence acts on whatever your CRM says. It does not stop to ask whether the person still works there, whether the email still lands, or whether you already have them twice. It just sends.
Every AI Blocker below is a place where acting on the record does damage. The tool works as designed. The data was never safe to act on.
The AI Blocker taxonomy
9 blockers in four families. Each one is a problem, a downstream failure, and the check Datuma runs to catch it.
Currency blockers
Is what your CRM records still true today?Stale role
The problem. The contact has changed role, so the job title you hold is out of date.
What breaks. Personalised outreach references a job they no longer do. It reads wrong, gets ignored, or is marked as spam.
How Datuma catches it. Every contact is checked against the live LinkedIn profile on the day you upload, as standard. A likely role change is flagged with the previous title attached, and a person confirms it before it lands.
Stale organisation
The problem. The contact has left the organisation your CRM has them at.
What breaks. An AI sequence pitches them at an employer they left. At worst it reaches them after they moved to one of your customers or a competitor.
How Datuma catches it. Flagged as a likely organisation move, with the previous organisation attached. You review every flag before you act, and nothing changes without you.
Reachability blockers
Can the channel actually deliver a message?Unverified or undeliverable email
The problem. The address was never checked at the mail server, or it bounces.
What breaks. Hard bounces push up your bounce rate, damage your sender reputation, and can throttle the whole sending domain.
How Datuma catches it. Every work email is checked at the mail server, as standard, and returned as Valid, Risky, Invalid, Unknown, or Unverified.
Catch-all only email
The problem. The domain accepts every address, so a valid result does not prove a real mailbox exists.
What breaks. Automation treats catch-all as deliverable and sends anyway. Some bounce, some disappear silently, and you cannot tell which.
How Datuma catches it. A catch-all result is returned as Risky or Unknown and is not counted as Verified, so it never reads as a green light.
No verified phone
The problem. There is no number, or the number was never checked against the live network.
What breaks. An AI voice agent dials a dead or reassigned number, burning call minutes and reaching whoever holds that number now.
How Datuma catches it. Every phone number is checked against the live network, as standard, before it counts toward readiness.
Identity blockers
Is this one real, matchable person?Duplicate record
The problem. The same person appears more than once, often spread across different lists.
What breaks. An AI sequence emails the same person from two records, two reps work the same lead, and you end up with territory and commission conflict.
How Datuma catches it. Duplicates are caught across every list you have ever uploaded, at £0.00, before enrichment spends a penny. Your team confirms each close call.
Role-account email
The problem. The address is a shared mailbox like info@ or sales@, not a person.
What breaks. Personalised automation addressed to a shared inbox reads as spam, and there is no individual to build a relationship with.
How Datuma catches it. Work-email discovery looks for the person's direct address so the record points at a human, not a shared inbox.
Missing identity key
The problem. The record has no email and no LinkedIn URL, so nothing can match or verify it.
What breaks. The AI acts on a name and a guess, with no way to confirm the person is real or still current.
How Datuma catches it. The contact is graded Not Verified, with Missing Email as the AI Blocker to clear before anything can act on it.
Safety blockers
Should you be contacting this person at all?Protected contact treated as a prospect
The problem. The record is one of your own customers, partners, or investors, a competitor, or someone on an active deal.
What breaks. Your AI SDR cold-pitches an existing customer or a competitor, which is a commercial and reputational own-goal.
How Datuma catches it. Protected accounts are pulled into their own file and kept out of your prospecting list, so nobody contacts them by accident. The flagging and separation is free, and those contacts still come back enriched and scored in their own file.
Verified Readiness
Clear the blockers and a contact earns a readiness tier
AI Blockers are the problems. Verified Readiness is the grade of what your AI tools can safely do with a contact once those problems are cleared. Datuma sorts every contact into four tiers, from Fully Verified to Not Verified, and lists the blockers still standing against each one.
A contact reaches Fully Verified when a deliverable work email, a first name, an organisation, a verified LinkedIn URL, and a phone number are all present. That is when email sequences, LinkedIn outreach, and AI voice calling can all act on it. Lower tiers unlock fewer channels, and the AI Blockers field tells your team exactly what to fix to move a contact up.
A person confirms every close call
Datuma checks every contact and proposes what it finds. It never changes your records on its own. Likely role and organisation changes, suggested LinkedIn matches, and duplicate pairs are all presented for your team to confirm or reject before they land in your results.
The 0 to 100 Trust Score is the summary of those checks. Nothing changes without you.
AI Blockers, common questions
What is an AI Blocker?
What are the AI Blockers in CRM data?
Why do AI Blockers matter for AI outreach?
How do I know if my CRM is ready for AI?
Does Datuma fix AI Blockers automatically?
See your own AI Blockers on real contacts
Upload a spreadsheet and get back every contact scored, with a Verified Readiness tier and an AI Blockers field showing exactly what to fix. Duplicates and protected-account separation are free.
100 contacts · 5 minutes