AgentSpot, with personal memory, shared company facts, and workflows he can see

François builds agents around the work he already does. AgentSpot connects to the company’s tools, adds his personal context and shared company facts, and turns a conversation into a workflow he can inspect. The results arrive in his calendar, his inbox, or a customer visit plan.

His personal memory learns from conversations and connected tools. Company memory gives everyone the same starting context, including product names and terminology an outsider would not know.

Those company facts are managed by admins. Agents can be shared across the company, but when someone runs one, it uses that person’s credentials and permissions.

François starts by describing the outcome. AgentSpot asks clarifying questions and builds a visible workflow. Each step can use a specialized agent, while the workflow gives the process a defined structure, including conditions and parallel steps.

He keeps refining it in conversation: add source links, make the email format consistent, scan only fresh Gong calls, and simplify slow Slack searches. Building the first version is the beginning; feedback makes it useful.

The meeting briefing appears in a calendar event beside the customer meeting. François borrowed that placement from his assistant Monique’s process: the prep is already there when he opens the meeting, instead of buried in a message thread.

The briefing combines account details, previous conversations, internal chatter, company news, adoption metrics, and relevant customer stories approved to share. He estimates that doing this preparation by hand would take at least an hour per meeting.

At the time of the conversation, people could already share agents and company memory. His next focus was bringing more people and agents into the same conversation, so they could work together with the right context and permissions.