█
LastWrite
  • > Curriculum
  • > Pricing
  • > For Educators
  • > About
  • > Contact
Log InGet Started

Questions, concerns, bug reports, or suggestions? We read every message, write to us at [email protected].

More ways to reach us →
LastWrite

Structured computer science lessons for aspiring developers and security professionals.

[email protected]

(201) 785-7951

Mon–Fri, 9 AM–5 PM EST

Learn

  • Curriculum
  • Pricing

Company

  • About
  • For Educators & Schools
  • Contact Us

Legal

  • Terms of Service
  • Privacy Policy
© 2026 LastWrite. All rights reserved.
Curriculum/AI Agents/Low-Code Agents/Relevance AI: Agent Teams Without Infrastructure
50 minIntermediate

Relevance AI: Agent Teams Without Infrastructure

After this lesson, you will be able to: Build a team of specialised AI agents in Relevance AI that collaborate on a complex task with role assignment and shared context.

Relevance AI's pitch: build a 'team' of agents, researcher, writer, reviewer, and orchestrate them like a tiny company. This lesson builds a content team that researches, drafts, and reviews, and coordinates handoffs between agents.

Prerequisites:Dify: Prompt IDE + Workflow Orchestration

The agent-team mental model

One agent ≈ one role. A researcher gathers facts; a writer drafts; a reviewer critiques and refines. The orchestrator passes context between them. It's how a real human team works, and how Relevance AI is structured.

Build a single agent first

  1. 1

    1. Sign up at relevanceai.com.

  2. 2

    2. Click 'Build an Agent'. Pick 'Custom'.

  3. 3

    3. Role: 'You are a research assistant. Search the web for facts and return 5 verified bullet points.'

  4. 4

    4. Add Tool: 'Web Search' (built in).

  5. 5

    5. Test: 'Research the top 3 trends in remote work for 2026.' Watch tool calls in the trace panel.

Build the team

  1. 1

    1. Create a second agent: 'Writer'. Role: 'You take research bullets and draft a 400-word blog post.'

  2. 2

    2. Create a third agent: 'Reviewer'. Role: 'You critique drafts for clarity, accuracy, and tone. Return revised draft.'

  3. 3

    3. Build a workflow: Researcher → Writer → Reviewer. Output of each agent feeds the next.

  4. 4

    4. Run end-to-end with one topic input. Inspect each agent's output in the trace.

💡 When agent teams are worth it

Single-agent works for ~80% of tasks. Multi-agent shines when: subtasks need different expertise, you want quality gates between steps, or you want each agent's output to be auditable separately. Multi-agent costs more tokens, don't reach for it by default.

Deployment options

Embed widget, drop on any site. API. POST to trigger the team. Slack/Teams integrations, agents that live in your team's chat. Scheduled, run the whole team on a cron.

Sign in and purchase access to unlock this lesson.

Sign in to purchase
←Dify: Prompt IDE + Workflow Orchestration
Back to Low-Code Agents
Memory, Tools, and RAG in Low-Code Agents→