Relevance AI
Build your AI workforce, no codeNo-code platform for building AI agents and multi-agent 'workforce' teams, with a visual tool builder and 2,000+ app integrations.
- Maker
- Relevance AI
- Founded
- 2020
- Pricing
- Free tier · from $19/mo
- Last checked
- October 2026
Verdict
No-code platform for building AI agents and multi-agent 'workforce' teams, with a visual tool builder and 2,000+ app integrations.
Score breakdown
vs n8n, the category leaderOps and revenue teams that want to assemble multi-agent workflows across their existing SaaS stack without hiring engineers
Key features
- Workforce — visual canvas for building teams of agents that delegate to one another
- Agent builder with custom instructions, knowledge bases and tool assignment
- Invent — generates a working agent from a plain-text prompt
- Tool builder with branching conditionals, native integrations, custom APIs and webhooks
- Bosh — prebuilt AI BDR agent for prospect research, outreach, objection handling and CRM updates
- 2,000+ app integrations
- Model-agnostic execution across OpenAI, Anthropic and Google models
- Enterprise controls: SSO, RBAC, audit logs, agent evaluations and A/B testing
Strengths
- +Genuine multi-agent orchestration — the Workforce canvas lets several specialized agents hand work between each other, not just run single-task bots
- +2,000+ native integrations covering Slack, Salesforce, HubSpot, Google Sheets and Zapier, plus custom APIs and webhooks
- +Model-agnostic across OpenAI, Anthropic and Google, so you are not locked to one provider's pricing or capabilities
- +Real free tier (200 Actions/month) and a $19/mo Pro plan make it genuinely testable before committing
- +Vendor Credits are billed at wholesale LLM cost with no stated markup, which is more transparent than competitors that bundle opaque 'credits'
Weaknesses
- −Two-meter billing (Actions + Vendor Credits) is hard to forecast, and overages are steep at $80 per 1,000 extra Actions — cost surprises are the most common user complaint
- −'No-code' oversells it: single agents are easy, but custom API calls, branching conditionals and multi-step workflows need real technical effort
- −Reviewers report a busy UI that sometimes fails to sync recent edits, which bites hardest on the complex workflows the platform is sold on
- −Thin review base (roughly 20-50 G2 reviews) and sparse community troubleshooting material compared to larger automation platforms
Overview
Overview
Relevance AI is a no-code platform for building AI agents, founded in 2020 by Daniel Vassilev, Jacky Koh and Daniel Palmer. The company is headquartered in Sydney, Australia with a second office in San Francisco, and raised a US$24M Series B in May 2025 led by Bessemer Venture Partners, bringing total disclosed funding to roughly US$37M.
The product’s distinguishing idea is the “AI workforce”: rather than configuring one chatbot, you assemble a team of specialized agents on a visual canvas and let them delegate work between each other. Agents are built from plain-language instructions plus tools — reusable actions that call native integrations, custom APIs or webhooks. The platform is model-agnostic, routing to OpenAI, Anthropic or Google models, and ships a prebuilt sales agent called Bosh that handles prospect research, outreach and CRM updates.
Pricing is where buyers should pay attention. Relevance AI bills on two meters: Actions (counted each time an agent runs a tool) and Vendor Credits (the underlying LLM and tool compute, passed through at wholesale). The transparency is welcome, but the forecasting burden is real, and overage rates of $80 per 1,000 Actions have generated the platform’s most persistent complaints. Reported customers include Qualified, Activision and SafetyCulture; the company said 40,000 agents were created on the platform in January 2025 alone.
Key Benefits
- Multi-agent by design: The Workforce canvas treats agent-to-agent delegation as a first-class primitive, not an afterthought bolted onto a single-agent product.
- Broad connectivity: 2,000+ integrations mean most workflows can be wired to existing systems of record without custom engineering.
- No model lock-in: Switching between OpenAI, Anthropic and Google models is a configuration choice, which protects against single-vendor pricing changes.
- Low-cost entry: A functional free tier and a $19/mo paid plan let teams validate a use case before committing budget.
Use Cases
- Outbound sales development — Bosh and custom BDR agents research prospects, personalize outreach, handle replies and objections, and write results back to HubSpot or Salesforce.
- Lead enrichment and routing — Agents pull data from multiple sources, score inbound leads and route them to the right rep automatically.
- Internal operations automation — Multi-step workflows spanning Slack, Google Sheets and internal APIs, handling approvals, reporting and data hygiene.
- Market and competitor research — Research agents gather and synthesize information on a schedule, delivering structured summaries into existing team channels.
Pricing
Checked October 2026 · USD- 200 Actions per month
- 1 build user
- One-time bonus Vendor Credits
- Access to the agent and tool builders
- 2,500 Actions per month
- $20/month in Vendor Credits
- 2 build users
- Full integration library
- 7,000 Actions per month
- $70/month in Vendor Credits
- 5 build users plus 45 end users
- Overages at $80 per 1,000 Actions
- Custom Actions and Vendor Credit allowances
- Unlimited users
- Calling and meeting agents
- SSO, RBAC, audit logs, agent evaluations and A/B testing