The whole base, healed in one run
Nobody hand-fixes a database — so it rots. The Data Agent works your entire base in one approved run: formats fixed, duplicates merged, records refreshed from live LinkedIn, missing contacts found and verified. While your team recruits.
One run, live
Formats fixed, duplicates merged, missing mobiles found — in place, on the record.
Six ways it works your base
Tick the actions you want and the agent shows you how many records match before anything runs. Everything below is what it actually does, not a wish list.
Parse
Extract what's already in the CVs you hold.
- Employment history
- Education
- Skills
- Years of experience
- Location
- Languages
Cleanup
Standardise and fix what's rotted.
- Phone numbers to international format (+44…)
- Job titles de-emojied
- Skills deduped (JS → JavaScript)
- Casing fixed (McDonald, O'Brien)
- N/A · TBD junk removed
- Duplicates merged
Classify
Match to your own master lists, not someone else's categories.
- Industry
- Sector
- Tags (seniority, function, work type)
Generate
AI writes the profile content you present to clients.
- AI overview (2–3 paragraphs)
- AI summary (elevator pitch)
Refresh the record from the live profile.
- Enrich and merge latest profile
- Generate a CV from LinkedIn
Contact Finder
Find what's missing, verified, charged only on find.
- Find email
- Find phone
- Waterfall: Recruitly → Apollo → Lusha → FullEnrich
You approve every run
An agent that touches 12,640 records answers to someone. This one scopes the run, waits for an admin, and reports exactly what it fixed, merged and found.
- Point it at a saved search, or the exact rows you selected in the grid.
- An itemised cost estimate appears before anything runs. Unspent budget is refunded.
- Nothing executes until an admin presses Approve & run.
- A summary email lands when it finishes, and every change is stored field by field — before and after.
Candidates. Contacts. Companies.
Rot doesn’t stop at candidates — your client book and account lists age the same way. The agent runs on all three record families.
Candidates
The full catalogue — parse, cleanup, classify, generate, LinkedIn, contact finder. CVs become structured records; stale profiles become today's.
Contacts
Clients move jobs too. Titles cleaned, employers refreshed from LinkedIn, missing emails and dials found before your BD call.
Companies
Names cased properly, industry and sector classified, size and location enriched — so account lists stop lying to you.
Pay per fix, not per seat
Enrichment is charged only when something is actually found — a miss costs nothing. Set a monthly budget and the agent stays inside it.
$0.15
Work email
found and verified
$0.25
Personal email
found and verified
$0.25
Phone number
found and verified
$0.10
LinkedIn profile
discovered and linked
$0.10
Full profile refresh
role, history, skills, location
$0.10
Company enrichment
size, industry, location
Volume discounts at 1,000 (10%), 5,000 (20%) and 10,000 (30%) enrichments · monthly budget control from $25 · 100% bonus credits on your first purchase, up to $500.
And you can watch it work
Data Health scores every record family before and after the run — missing mobiles, stale employers, unclassified sectors. The number goes up, and the base is worth selling from again.
See the whole Living Database story →Data Health
Missing mobiles, stale employers, unclassified sectors — surfaced per record family, fixed in bulk.
Questions recruiters ask
What exactly can the agent change?
Six families of actions, and you pick which ones run: Parse (pull missing fields out of CVs you already hold), Cleanup (phone formats, title junk, skill duplicates, placeholder values, duplicate records), Classify (industry, sector and tags from your own master lists), Generate (AI overviews and summaries), LinkedIn (refresh the record from the live profile) and Contact Finder (find missing emails and phones).
Can it run without someone approving it?
No. The agent scopes the run first — how many records match, which actions will run, and an itemised cost estimate. Nothing executes until an admin presses Approve & run.
How do I choose which records it works on?
Point it at a saved search, or select rows in the grid and launch it on exactly those. Candidates, client contacts and companies each have their own action set.
What does a run cost?
Cleanup and classification run on your plan. Enrichment uses credits, itemised before you approve: work emails from $0.15, personal emails and phone numbers $0.25, LinkedIn and company enrichment $0.10 — charged only when something is found. If the run spends less than estimated, the difference is refunded. Volume discounts kick in at 1,000, 5,000 and 10,000 enrichments.
How do I know what it changed?
Every change is stored field by field, before and after, per record. A summary email lands when the run finishes with the full tally — records healed, duplicates merged, contacts found — and Data Health scores show the base improving.
Your database rots a little more every day you think about it.