Complete AI Training
Sign inGet my AI kit

Your job's AI kit

Get your AI kit

Tell us who you are and what you do. We show you your kit right away and email you the link: skills, prompts, AI agents, MCP servers and courses for your job.

500+ jobs ready, and we make a kit for any other job. No payment needed to look.

Share

AI agent for chief digital officers

Master Data Duplicate Resolution Agent

Reduce duplicate records safely without breaking any downstream report

Data cleanupExecutives and strategy
See how it worksGet access
Master Data Duplicate Resolution Agent: what goes in, what the agent does and what you get

What it does

Sales sees 52,000 customers and finance sees 46,000, and the gap is mostly duplicates. This agent scans master tables for likely duplicates and scores each pair on name, address, ids and behavior such as shared orders. It groups strong matches and proposes merges with a surviving record. Before anything changes, it applies the merge to a test copy and reruns the key reports to check that totals, such as revenue and customer counts, still tie out. If a total breaks, it rejects that merge, narrows the rule and tries again. The data steward approves merges in the live system. Edge case: two clinics with the same name at different addresses score high but fail the address check, so they stay separate.

How it works

Follow the arrows from top to bottom. The orange dashed arrow is the loop: when a check fails, the agent goes back and tries again.

Start and resultWhat it doesA check on its own workWaits for your OKGoes back and retries
Yes, continueApprovedNo 1 STARTS WHEN Weekly scan or bulk load completes 2 USES A TOOL Scan master tables for likely duplicates 3 DOES Score each pair on name, address, id and behavior 4 DOES Group strong matches and pick the surviving record 5 USES A TOOL Apply the merges to a test copy 6 USES A TOOL Rerun key reports on the test copy 7 CHECKS THE RESULT Do report totals still tie to the original? If not: Reject the failing merge, tighten the rule andrescore the group. Back to step 3. 8 YOU APPROVE Data steward approves merges in the live system 9 USES A TOOL Record the merge log with before and after counts 10 RESULT Merged records and an audit log
Read the steps as a list
  1. Weekly scan or bulk load completes
  2. Scan master tables for likely duplicates
  3. Score each pair on name, address, id and behavior
  4. Group strong matches and pick the surviving record
  5. Apply the merges to a test copy
  6. Rerun key reports on the test copy
  7. Do report totals still tie to the original?If not: Reject the failing merge, tighten the rule and rescore the group. Back to step 3.
  8. Data steward approves merges in the live systemThe agent waits here for your OK.
  9. Record the merge log with before and after counts
  10. Merged records and an audit log

How it decides

It merges only pairs scoring above the match threshold and passing the report tie-out. Pairs that fail go to manual review.

  • Merge only pairs scoring above 0.9 on the match scale
  • Send pairs scoring 0.7 to 0.9 to manual review
  • Reject any merge that changes revenue totals
  • Keep the record with the most complete history as the survivor

Make it yours

Every agent is a starting point. You choose these settings for your own situation.

  • Match threshold (default 0.9)
  • Manual review band (default 0.7 to 0.9)
  • Reports used in tie-out (default revenue and customer count)
  • Scan frequency (default weekly)

What keeps you in control

It always asks you first

  • Data steward approves merges before they run in the live system

Hard limits

  • Never merge records in the live system without approval
  • Never delete original records, only mark them merged
  • Test every merge before approval

It stops when

  • Done: Approved merges applied and logged
  • Stop: Tie-out fails repeatedly, so hand the group to the data steward

Set it up

We guide you through the set-up, step by step

Members get the full set-up guide for this agent. No technical skills needed: you copy, paste and upload.

10 minto set it up in your AI
5 AIsChatGPT, Claude, Copilot, Gemini, Grok
  • One set of instructions to paste into your AI, with the clicks for ChatGPT, Claude, Microsoft 365 Copilot, Gemini and Grok
  • The agent then walks you through connecting your own data, one source at a time
  • A downloadable copy with the flow chart, the rules and the full guide
Get access to this agent

An example run

What happensA scan finds 1,340 candidate pairs. 1,100 score above 0.9. On the test copy, revenue changes by $48,000 because 12 pairs belong to two sister accounts with separate billing. The agent rejects those and rescores the rest. The tie-out passes at 1,088 merges. The steward approves and customer count drops from 52,000 to 50,912.

More agents for chief digital officers

Decisions · IT and development

Blockchain Use Case Feasibility Screen Agent

A fair, documented screen of a blockchain proposal with a clear recommendation

10 steps2 checks6 jobs
Compliance · Executives and strategy

Data Retention and Deletion Schedule Agent

Data is deleted when the policy says, with proof, and nothing under hold or in use is touched.

10 steps2 checks6 jobs
Compliance · Risk management

Data Vendor and Third-Party Sharing Review Agent

Every vendor that receives data has a current agreement and review.

11 steps2 checks9 jobs
Compliance · Executives and strategy

Customer Data Consent and Preference Sync Agent

Consent and preferences match in every system, with unresolved records listed.

11 steps2 checks5 jobs