How Spark Hire Scaled Customer Account Reviews with AI in Slack

A custom AI agent designed and built by Agreeable Data on Anthropic’s Claude Managed Agents lets Spark Hire review customer accounts from Slack and capture the results in Salesforce.

How Spark Hire Scaled Customer Account Reviews with AI in Slack

Built with: Claude Managed Agents, Slack, Salesforce, Zendesk, Gainsight, HubSpot, and Outreach.

At a glance

  • Full portfolio coverage: Spark Hire completed customer account reviews across its entire portfolio.
  • Reviews from Slack: The team requested AI account reviews in batches directly through Slack.
  • Repeatable reviews: Future account reviews can be scheduled through Slack to run automatically.

About Spark Hire

Spark Hire provides video interviewing and applicant-tracking software for hiring teams. Its customer success team wanted a clear, repeatable way to understand account health and spot renewal risks across its portfolio. They needed a way to pull together customer account data, support tickets, and conversations without manually assembling every review.

A useful review process the team couldn’t sustain

For each account review, Spark Hire’s customer success team had to gather information from Salesforce and recent customer calls and emails, then write a three-paragraph summary of where the account stood and the risks to renewal.

The reviews were valuable, but the manual research and writing made them difficult to keep up with across the portfolio. The team needed a way to extend that same review process to every account.

A custom account-review agent in Slack, built on Claude Managed Agents

Agreeable Data designed and built a custom AI agent for Spark Hire on Anthropic’s Claude Managed Agents. Anthropic provides the infrastructure that runs the agent; Agreeable Data developed the integrations, review skills, and testing that make it work for Spark Hire’s customer success process.

Spark Hire supplied the business definitions and review criteria. Agreeable Data translated those into agent instructions and tested how the agent interpreted Spark Hire’s customer information.

The account-review agent in Claude Managed Agents, showing its connected systems, review skills, memory, and vaulted credentials The account-review agent in Claude Managed Agents: connected systems, review skills, memory, and vaulted credentials. Illustrative, with simplified configuration.

The team requests a review for one account or a list of accounts directly in Slack. The agent gathers customer account data, support tickets, and conversations, follows Spark Hire’s review criteria, and produces a summary of account health and renewal risks. Reviews can then be saved to the account’s Salesforce record.

The agent also has memory, so the team can correct it directly in Slack and have those corrections inform future reviews. This lets Spark Hire refine how the agent interprets customer information as the team uses it.

Claude Managed Agents enables the agent to delegate account reviews to parallel workers while keeping each account’s research separate. Credentials for connected systems are held in managed vaults. For Spark Hire’s implementation, Agreeable Data configured Salesforce write access to be limited to the account-review notes field.

A Slack request for nine numbered accounts fanning out to one parallel review worker per account One Slack request fans out to a parallel worker per account, each with its own research. Illustrative workflow.

Account reviews across every customer

Spark Hire’s team used the agent to complete reviews for every customer, running the process themselves. A review process that had been difficult to sustain manually could now cover the full portfolio, giving the customer success team a consistent view of account health and renewal risks.

The team can also schedule future reviews through Slack to run automatically. This makes portfolio-wide reviews a repeatable process, reducing the manual work required to keep account assessments current.

“Our team found the account reviews valuable, but they couldn’t keep up with the manual work. With the agent, we’ve been able to review every customer and make that a repeatable process. That gives our customer success team a much more practical way to stay on top of account health and renewal risks.”

— Tom Cruger, Spark Hire

Looking beyond individual accounts

With reviews captured in Salesforce, Spark Hire now has a foundation for meta-analysis across its customer portfolio. The team can analyze reviews together to look for recurring renewal risks, shared customer challenges, and patterns that may be less apparent when reviewing accounts individually.

A Salesforce record labeled Account 1 with an agent-written Account Review field showing health, summary, and renewal risks Each finished review lands in the account’s Account Review field, the only field the agent can write to. Illustrative account review.

This opens the door to using account-level findings to inform broader customer success decisions, such as where to focus outreach or which issues warrant attention across multiple customers.

“Putting the ability to do broader customer analysis in our leadership’s hands will be huge for us. Once the reviews are captured in Salesforce, we can look across accounts for recurring risks and shared challenges. That gives us a way to step back from individual customers, see the bigger picture, and make more informed decisions about where to focus our team.”

— Tom Cruger, Spark Hire

Spark Hire owns the agent, which runs in its own Anthropic account. The team can continue to develop this capability by updating the agent’s skills as its review criteria and business needs evolve.

Now booking new engagements

Make AI work on your company data.

Connect your systems, clean up your data, and build agents your team can trust in the tools they already use.

Trusted by data teams at RudderStack, Enverus, Flatiron School, Marqo, CoLinx, and more.