At a glance
- Industry
- Biotechnology
- Platforms
- Salesforce Service Cloud, Salesforce Agentforce, MuleSoft
- Service
- Integration
- Results
- 30% support time saved · 19 knowledge articles powering ai
The challenge
A three-person support team — one person handling 90% of cases — was overwhelmed by growth, while an administrator manually moved payment data between Salesforce, NetSuite, and Stripe.
What we did
- Service Cloud with six case record types and 19 knowledge articles to power AI responses
- An Agentforce agent that drafts email replies, summarizes cases, and recommends knowledge articles in the brand’s tone
- MuleSoft integrations connecting Salesforce, NetSuite, and Stripe with automated payment sync and error handling
- A phased rollout starting internal-only on the most common case type, expanding as the knowledge base grew
The results
The AI agent produced send-ready draft responses for 30% of question-type cases, manual payment entry was eliminated, and the team gained a framework to scale support with case volume rather than headcount.
It’s pretty freaking brilliant.
Client, Biotechnology
Does this sound like your team?
- One support person answers most cases, and every new customer adds to a queue that no one else can clear.
- Replies to common questions are written from scratch each time because answers live in people’s heads rather than articles.
- An admin copies payment records between Salesforce, the ERP and the payment processor by hand, and mismatches surface later.
Where we would start
Start by sorting cases into record types and finding the question type you see most often. Write knowledge articles for that type first, because an AI agent can only draft good replies from content you have approved. Keep the agent internal at first so staff review each draft, then widen scope as the knowledge base grows. In parallel, replace manual payment entry with a MuleSoft integration that reports sync errors.