Case study · Enterprise medical company
AI workflows that turn service calls into tasks on the right account
Service reps at a medical company with about two billion dollars in annual revenue were logging call notes and tasks by hand. We put five Agentforce agents into production on Service Cloud: three that turn what a rep says about a call into tasks on the right record, and two that score sentiment on every case.
- Healthcare
- Six months, as a subcontractor to a Salesforce partner
5
Agentforce agents in production in six months
The problem
Reps logged every call by hand: the notes, the follow-up tasks and the record they belonged to. It took a long time, and it was time away from the next customer.
Finding the right record was the hard part. One contact could be related to many accounts, so a simple name lookup often picked the wrong one, and the follow-up landed where nobody would see it.
The team also wanted a sentiment read on every case from the call transcripts and emails already on it in Service Cloud, without anyone reading them by hand.
What we delivered
- Three voice to task agents, each built for a different use case: a rep gives the agent what was said on a call, and it creates the tasks on the right contact and account
- Fuzzy matching that finds the right contact and account even when a contact is related to many accounts, refined over three versions
- A sentiment agent that scores each call transcript and email on a case
- A parent sentiment agent that rolls those scores up into one sentiment for the case when it closes
- Built on Service Cloud, with email to case and voice transcript records related to each case
- External systems connected to the agents through MuleSoft
Have a project like this?
Thirty minutes with the architect who would do the work. You leave with a recommended next step.
Book a free 30-min scoping call