Data Steward
Find and fix duplicate and incomplete records in any Salesforce object, with changes approved by your team.
The fit.
- Cloud
- Platform (works with Sales, Service, Nonprofit, Education and custom objects)
- Who it’s for
- Salesforce admins, data stewards and RevOps teams whose AI, reporting or automation suffers from messy data
- Typical go-live
- 2–4 weeks
- Runs as
- Agentforce agent in your org
Built to order in your org. Not a template.
Fixed price, USD
- Simple
- $5,000
- Standard
- $10,000
- Complex
- Fixed quote after the Readiness Assessment.
What needs to change.
Duplicates and half-filled records quietly break everything that runs on your data: reports, routing, marketing, and now AI agents. Rules at entry do not replace a review workflow for existing data, and reviewing near-matches by hand takes time.
What it does.
- Scans any object you choose, standard or custom, for likely duplicates, on a schedule and as records are created
- Explains every match in plain language (for example, "same email after clean-up, names 96% similar, phone missing on one side")
- Merges duplicates after you approve, keeping the values you choose and moving related records across, including on custom objects Salesforce can't merge natively
- Scores every record's completeness and flags missing, malformed or stale data
- Suggests fixes from related records ("this Contact's city is blank, its Account says Toronto") for one-click approval
- Recovery steps are recorded in the log. Native merge recovery depends on the losing record still being in Salesforce's Recycle Bin and may require restoring relationships.
Price.
Prices in USD. Simple and Standard are fixed scope. Complex is scoped after the assessment.
| Tier | Your org looks like | Arkeforce price (USD) |
|---|---|---|
| Simple | 1–2 standard objects, native merge only | $5,000 |
| Standard | 3–5 objects, one custom object, Slack | $10,000 |
| Complex | 6+ objects, several custom objects, Data Cloud, high volumes | $12,000 to $21,500Typical range. Fixed quote after the Readiness Assessment. |
- What you pay Salesforce
- Agentforce entitlements and Flex Credits, billed by Salesforce
- What you pay anyone else
- Nothing for this agent.
- What you pay Arkeforce
- The tier price above, fixed before we start.
Outcomes we target.
| Outcome | How we measure it | Typical 90-day target, tracked with you |
|---|---|---|
| Fewer duplicates | Open duplicate groups ÷ records | −50% from the first scan |
| More complete records | Average quality score | +15 points |
| Safe changes | Merges undone ÷ merges | Under 2% |
Targets are targets, not measured results or guarantees. Tracked with you during the 30-day support window and under a retainer after that.
Sample exchange.
Illustrative examples, not results from a client engagement.
You:Show me the top 10 Donor duplicates.
Agent:Here are 10 groups, highest confidence first. The first: Jon Smith and John Smith, same email, same postal code, phone missing on one. Want to preview a merge?
You:Merge them, keep the newer phone number.
Agent:Here's the preview: John Smith survives, 4 gifts and 2 notes move over, phone comes from the newer record. Confirm and I'll merge.
You:Delete all the duplicates.
Agent:I can merge records after you review each group, or open the Data Steward Console for bulk review. A native merge deletes the losing record; the recovery steps are in the log.
Built for your org.
- We configure Data Steward to your objects, fields and matching rules, and it speaks your language, "Donors" and "Households", not "records". It reuses your existing Salesforce duplicate rules where they work and adds smarter matching where they don't. It only sees and changes what each user is already allowed to.
Installed as a package in your org, source visible to your admin.
Scheduled scans and the Console use no Flex Credits. Chat actions can consume Flex Credits. Agentforce testing, including sandbox tests, can consume requests and credits. We confirm metering against your licences.
Compared with what’s already out there.
| Option | What it covers | When it may be enough |
|---|---|---|
| Native Agentforce | Duplicate and matching rules identify duplicate accounts, contacts and leads.Source · 2026-10-28 | Your standard-object matching needs fit the built-in rules. |
| Salesforce in Claude | Sales skills can inspect a seller's account data. Scheduled duplicate governance and approved merges need a separately scoped workflow.Source · 2026-10-28 | Check the job, access and scope against your org. |
| AgentExchange | clearMDM Data Steward Agent offers conversational duplicate review and stewardship inside Salesforce. It is a paid marketplace listing.Source · 2026-10-28 | Its matching, review and stewardship scope fits your data model. |
| Arkeforce | Find and fix duplicate and incomplete records in any Salesforce object, with changes approved by your team. | |
Verified on 2026-10-28. Availability and licensing can change; we confirm them for your org.
What’s included.
- Discovery of your org and a written fit report
- Configuration to your objects, fields and matching rules
- Sandbox testing with your data, including adversarial tests
- A report-only first week in production (scans and scoring live, no merges)
- Admin and steward training
- 30 days of post-launch support
Prerequisites.
- Agentforce enabled with Employee agents, and entitlements for chat actions (the Console and scheduled scans use no Flex Credits)
- Enterprise, Performance or Unlimited Edition with Salesforce Foundations or an Agentforce 1 Edition; required add-ons depend on the actions used.
- Optional: Data Cloud (adds identity-resolution matching), Slack (weekly digest and approvals)
Retainer (optional, monthly).
Tuning, threshold changes and re-checks after Salesforce releases come out of your retainer hours.
- Essentials $1,000/mo with 3 hours
- Growth $3,000/mo with 10 hours
- Partner $5,500/mo with 20 hours plus one integration per quarter
Retainer prices in USD.
Trust and safety.
- Merges and fills need explicit approval from your team. The agent writes its own scores, flags and log records.
- Never deletes a record outside a merge you approved. Where our own merge runs, the losing record is archived, not deleted.
- Respects each user's existing Salesforce permissions
Start with the job.
Tell us what needs to happen and where it gets stuck. We’ll check the fit and agree the scope before anything is built.
Not sure yet? Explore the Readiness Assessment →