Automotive / Mobility
AI workflow automation
A major European automotive group runs several brands, each processing industry developments, partner activity, customer signals and internal knowledge. Research was manual, signals took too long to reach action, and commercial teams kept researching the same companies from scratch. They wanted one shared system rather than more isolated experiments, without giving up human control of what gets published.
What was built
The pieces, and how they fit
All of it is configuration on a single instance.
Specialised agents per stage of research, analysis, content and commercial work, with a coordinator granted agent-to-agent calls to a defined set of them
Shared knowledge in collections and in tables the agents read through named queries
Stores carrying context between stages, so a workflow spanning monitoring, analysis and drafting keeps what it learned
Scheduled functions polling the chosen sources, then assessing relevance per brand and preparing content
Connectors into CRM, collaboration tools, content systems and communication channels
Approval-gated functions on anything that publishes or commits, so the workflow stops and waits for a person
A namespace per brand on shared infrastructure, keeping knowledge, tone of voice, users and assets separate
The result
One reusable foundation instead of disconnected assistants. Market intelligence moves faster from discovery to action, commercial teams get structured briefings without starting over, and a new brand or workflow no longer means a new AI backend.
AI agents
Agent-to-agent delegation
Scheduled functions
State stores
Collections
Named queries
CRM connectors
Approval-gated functions
Role-based access control
Namespaces per brand
Execution tracking
Configuration as code
Get started