Claude-Ready Salesforce. Governed before it's switched on.
Claude makes a clean org fast. It makes a messy org dangerous, faster.
Your reps are about to talk to Salesforce in plain English, and Claude will read, update and report on your records with whatever access you hand it. We make your org and your Claude Enterprise tenant safe, governed and worth the money before that happens.
Readiness Check: 12 questions, about 3 minutes, scored in your browser.
What changed
What changed this fall
Anthropic's October release review made the list long. Four items matter most if you run Salesforce.
- Beta
Salesforce inside Claude
Reps talk to Salesforce from Claude: they update records and pull reports. Your Salesforce admin has to apply for access.
- Rolling outOff by default
Unified Claude
Chat and Claude's agentic workspace in one cloud surface on desktop, web and mobile. It needs Chat and Code execution on, then its own toggle, and it can be scoped by role.
- Smart reports: public betaBudgets: available
Smart reports and pooled budgets
See where Claude spend went, by work stream. Set department caps and individual caps; whichever is hit first applies.
- BAA via account teamHooks: GA
HIPAA coverage expands, inference hooks GA
BAA coverage now reaches more of Claude's local tools, through your account team. Inference hooks check policy before every prompt and every tool call.
The risk
Why a messy org becomes a risky org
An AI layer doesn't fix your org. It inherits it. Every flaw you've lived with becomes something Claude can act on at conversation speed.
Field-level security exposure
A rep asks Claude for "everything on this account." Claude returns the margin field and the contract notes, because that user could always see them. Nobody ever checked.
Bad writes, no validation
A rep says "mark it closed won, $50k." No validation rule on close date or stage, so a deal with no close date and the wrong product lands in your forecast.
Duplicates, confident wrong answers
Three records for the same company. Claude picks one and tells your rep the renewal date with total confidence. It's the wrong record.
Automation cascades
A rep says "close out this renewal." Claude updates the opp. Four flows, a trigger and an approval process fire. Did anyone map that?
What we cover
Two sides, one owner
We do both sides, because the risk lives where they meet.
Your Salesforce org
- Profiles, permission sets and field-level security for every user who gets Claude
- Validation, required fields and restricted picklists on the objects Claude can write
- Duplicate rules and a data-quality clean-up list
- A full automation inventory: what fires when a record changes
- Integration and connected-app users reviewed
Your Claude tenant
- Unified Claude rollout by role
- Efficient mode and model access by role
- Artifact access for Docs, Slides, Design, Dashboards and Motion
- Pooled department budgets and individual caps
- Smart reports, skills and plugin approvals, and inference hooks
- SIEM and DLP rules checked after the file-ID change
Pricing
Packages
Fixed prices. You buy an outcome, not a meter.
Package A
Claude-Ready Org Audit
$7,500fixed
About 2 weeks
Know exactly what Claude will see, write and set off in your org.
- Profile, permission set and FLS map for every user getting the Salesforce app in Claude
- "AI write risk" score for each object: validation coverage, required fields, duplicate rules, picklist hygiene
- Full automation inventory per object: flows, triggers, process builders, workflow and assignment rules
- The "what fires when Claude updates a record" map
- Data-quality hot list: duplicates, orphaned lookups, stale owners
- Prioritized remediation plan with hour estimates
- Written report plus a 60-minute walkthrough
Package B
Claude Enterprise Admin Setup
$4,500fixed
1 to 3 days
Every new admin setting chosen on purpose, not left at its default.
- Unified Claude rollout plan by role
- Efficient mode by role: protect the budget without slowing power users
- Artifact access by role, plus pooled department budgets and individual caps
- Smart reports on, department-head access scoped, attribution decision documented
- Skills and plugins: scanning confirmed, approval flow, who can publish, share-to-group
- SIEM and DLP check for the file-ID activity-feed change, plus an inference-hooks policy baseline
- Salesforce app beta application and connector governance
Package C
Claude-Ready Retainer
From $3,500/ month
Monthly, 3-month minimum
Every monthly release reviewed. Fixes shipped ticket by ticket.
- Org fixes from the audit, worked down ticket by ticket
- Monthly release review: what changed for you, what we turned on, what we left off and why
- Monthly smart-reports summary: where spend went, the waste we found, the skills worth building, and we build them
- Salesforce release checks three times a year against your Claude setup
- A written record of every setting we change
- Includes upkeep for modules we rebuild for you
Claude-Ready Launch: both A and B for $10,500. Save $1,500.
Process
How we work
Readiness Check
Twelve questions, three minutes, scored in your browser. You see your gaps before we talk.
Audit
Read-only access to your production org. We map access, write risk and automation, and you get the report and the fix list.
Fix and configure
Every org change is built in a sandbox and passes an independent review gate. Sandbox and production stay separated, and the production step is owned by a person, not a script. We run an automation inventory before every data change.
Monthly release review
Anthropic ships monthly. Every month we tell you what changed, what we turned on, and what we left off.
Fit
Who this is for
- Heads of RevOps who want reps using Claude without wrecking the forecast
- Salesforce admins and owners who will get the call when Claude does something strange
- CIOs and IT leads who just bought Claude Enterprise and inherited a page of new settings
- Compliance leads in regulated work: healthcare, financial services, education, and nonprofits holding donor PII
Background
Credentials
- JD
- AIGP
- CIPP/US (IAPP)
- Salesforce Consulting Partner
- Anthropic Claude Partner Network
Questions
FAQ
Do we need Claude Enterprise?
For most of Package B, yes. The admin controls we configure, like role-based access, pooled budgets, smart reports and plugin approvals, are tenant-level settings. If you're on another plan, the org audit still applies in full.
Is the Salesforce app in Claude generally available?
No. It's in beta, and your Salesforce admin has to apply for access. We handle the application as part of Package B and set the connector up with governance from day one.
Will Claude bypass our Salesforce permissions?
Claude works within what the connected user can access, so the user's access is the control. That's exactly why profiles, permission sets and field-level security matter more now, not less.
We already use Agentforce. Do we need this?
Agentforce is Salesforce's own agents, inside Salesforce. This is Claude reaching into Salesforce from the outside. Both inherit the same org, so the org work overlaps, and our Agentforce Governance & Privacy Assessment covers the Agentforce side.
Can you work in our sandbox only?
For changes, yes: every change we make is built and tested in a sandbox first. The audit itself needs read-only access to production so it reflects reality. Nothing reaches production without a person on your side approving it.
Do you handle HIPAA?
We configure and document the controls so your Claude rollout is audit-ready. Business Associate Agreement coverage comes through your Anthropic account team. We are not your legal counsel, so loop yours in on the contract side.
How fast?
The audit takes about two weeks. Admin setup takes one to three days. Bundled, both fit inside three weeks.
What does the monthly retainer actually include?
Audit fixes shipped ticket by ticket, the monthly release review, the smart-reports spend summary, and Salesforce release checks three times a year. Scope is agreed up front, and larger orgs are quoted separately.