At a glance
- Industry
- Financial Services / Payments
- Platforms
- Anthropic Claude, Claude + CRM Integration, Salesforce Sales Cloud
- Service
- Integration
- Results
- Aug 2026 delivered · 4 connected workstreams
The challenge
Sync Payments wanted to use Claude with its Salesforce data to support practical leadership tasks: recognizing team contributions, reviewing pipeline health, and preparing for business decisions. Leadership also wanted a repeatable version of its existing approach to using Claude for deal evaluation. The engagement needed to distinguish between reading business information and changing it. Recognition would remain subject to human review and posting, while any pilot that created Salesforce records would require approval before each action.
What we did
- Team recognition: built “Daily Wins,” a recurring Claude workflow that reviews Salesforce activity, identifies contributions worth recognizing, checks attribution and prior recognition, and drafts messages for leadership. A person reviews and posts each message; the workflow does not automatically publish recognition.
- Pipeline and forecasting analysis: scheduled, read-only analysis of aging opportunities, shifting close dates, pipeline coverage, and forecast quality, producing prioritized recommendations for sales leadership and checked against Salesforce reports.
- Deal-evaluation support: turned leadership’s existing approach into a reusable Claude workflow for organizing terms, comparing tradeoffs, and preparing for vendor and partner discussions. The workflow supports decision-making rather than replacing legal or financial review.
- Human-approved follow-up tasks: designed and piloted a limited workflow for creating Salesforce follow-up tasks from pipeline analysis, using restricted access, approval before each write, and an audit trail.
The results
The engagement was delivered in August 2026, with Daily Wins entering production that month and the remaining analysis workflows and controlled write pilot delivered by month-end. Sync Payments gained recurring recognition drafts grounded in Salesforce activity, scheduled pipeline and forecasting reviews, and a documented approach to deal evaluation. The work also established and tested a limited path from analysis to action through a completed pilot for human-approved follow-up task creation. Leadership retained control over recognition posts, business decisions, and every Salesforce write in the pilot.
Does this sound like your team?
- Leaders paste Salesforce exports into a chat window to ask questions, then rebuild the same prompt from scratch every week.
- Good work by reps goes unmentioned because nobody has time to comb through activity history and confirm who actually did it.
- You want AI to suggest next steps in the CRM, but nobody has agreed which records it may change or who signs off.
Where we would start
Start by separating questions that only read Salesforce from actions that would change it. Read-only work, such as reviewing stale opportunities or drafting recognition for a person to post, can run on a schedule with little risk. Any step that writes records should begin as a narrow pilot with limited permissions, sign-off on every write and a log of what changed. Checking each analysis against existing reports builds trust before scope grows.