Topic 04 of 10
Multi-agent & integration
Several agents and the systems around them: protocols, handoffs to people, memory, browser automation and the standards an enterprise sends you.
- 11 lessons
- 7.3 h reading
- 12 h hands-on
Start lesson 01
Progress0 / 11
Lessons
- 01Supervisor, router or swarm?Pick the multi-agent shape that fits the task, and know what each one costs. Pro
- 02Crews that mirror the customer's teamRoles, delegation and coordination mapped onto real approval structures. Pro
- 03Fan-out and sandboxed parallel workProcess thousands of records fast, isolated, and inside the budget. Pro
- 04Agents talking to other vendors' agentsA2A interoperability, and what to agree before agents trust each other. Pro
- 05Handing off to a person, wellSpecialisationEscalations with context, so whoever takes over doesn't start again. Pro
- 06SaaS connectors and auth flows LabOAuth, tokens and pagination: the integration work that is most of the job. Pro
- 07A gateway for bursty agent trafficQuotas, rate limits and fallbacks that protect the customer's upstream systems. Pro
- 08Agents started by events and schedulesBegin work from business events, not from someone opening a chat. Pro
- 09Automating systems with no APISpecialisationBrowser agents for the legacy screen everyone depends on, made reliable. Pro
- 10Live web search as a toolBring in market and regulatory facts, with sources and freshness checked. Pro
- 11FHIR and electronic prior authorisationSpecialisationRead and write healthcare data in the standard the hospital side actually sends. Pro