Demo listing
Mira
Support Agent
Answers routine customer questions by email and chat, and hands complex cases to your team with a summary.
Illustrative
This is an illustrative example created to show how an AI employee could be used. Figures are not from a real deployment.
Illustrative example
E-commerce
Faster replies for an online store's support team
Business context
A 25-person online retailer selling home goods, with a 4-person support team.
Challenge
Most tickets were repeat questions about delivery and returns, so replies slowed down during sales.
Previous workflow
- Step 1Customer emails or chats
- Step 2Ticket waits in a shared inbox
- Step 3Agent searches the help centre
- Step 4Agent writes a reply
AI employee introduced
A support AI employee answers routine questions from the help centre and hands the rest to staff.
Implementation approach
Help-centre articles were reviewed first, then the AI employee ran alongside the team for two weeks before answering on its own.
Integrations and human oversight
Integrations
Human oversight
Refunds and complaints always go to a person. The team reviews a sample of answers every week.
Before and after
Before
- Step 1Customer emails or chats
- Step 2Ticket waits in a shared inbox
- Step 3Agent searches the help centre
- Step 4Agent writes a reply
After
- Step 1Customer emails or chats
- Step 2AI employee answers routine questions
- Step 3Complex tickets go to an agent with a summary
- Step 4Agent resolves and approves refunds
The outcomes
Median first reply time
Before
Illustrative6 hours
After
Illustrative1 hour
Before6 hours
After1 hour
Tickets handled by staff per week
Before
Illustrative400 tickets
After
Illustrative160 tickets
Before400 tickets
After160 tickets
Limitations and lessons learned
- Answers are only as good as the help-centre articles
- Peak days still need extra staff