The project story
The brief
Storage inquiries often arrived when staff were helping members, working the docks, or off the clock. By the time someone returned the message, basic details were still missing: boat length, storage type, preferred start date, and contact preference.
The solution had to feel useful rather than robotic. It needed to answer routine questions, recognize when a person should step in, and avoid promising availability the system could not verify.
How we approached it
Strategy before production.
Map the real qualification sequence
The conversation follows the questions an experienced marina coordinator already asks, in an order that feels natural to a boat owner.
Set clear automation boundaries
The bot can explain general options and collect requirements, but availability, pricing exceptions, and unusual vessel needs route to a person.
Pass context, not transcripts
CRM records receive a concise lead summary with the details staff need for the next conversation.
Production graph
Where the work went.
This is a scope mix, not a client-performance claim.
What shipped
A complete handoff, not loose files.
- Conversation and escalation map
- Hermes Bot configuration
- Structured storage-intake workflow
- CRM field mapping and lead summaries
- Scenario testing and staff handoff guide
Delivery timeline
The engagement arc.
- 01 Staff interview and flow map
- 02 Bot voice and knowledge setup
- 03 CRM routing integration
- 04 Scenario QA and launch
The outcome
What changed after delivery.
The resulting workflow gives prospects an immediate, structured response without pretending the bot can replace marina judgment. Staff start follow-up with vessel and timing context already captured.
For the agency, the project became a repeatable example of practical AI: narrow scope, visible boundaries, and a handoff that improves the human conversation.
The best intake bot does not try to close the deal. It makes the next human conversation worth having.