Relevance AI

Relevance AI operates a low/no-code platform for building, deploying and managing AI agents and multi-agent "workforces" that autonomously complete business tasks in a manner the company compares to human employees. Users can generate an agent from a natural-language description using Invent, clone pre-built agents, tools and workforces from a marketplace, or build from scratch; agents are equipped with no-code tools for actions such as sending email, updating a CRM, web search or calling arbitrary APIs, and with a retrieval-augmented "Knowledge" layer fed by uploaded files or syncs from Google Drive, SharePoint, Notion and websites. Multiple agents can be linked on a visual canvas so each handles a narrow step and the workforce manages handoffs, with escalations, approval workflows and triggers configured before deployment.

The company positions its offering around narrow "specialist agents" that each own a single task, contrasted with general frontier-model exploration tools. Marketed agent examples cover sales and go-to-market work — research and enrichment, pre-meeting briefs, post-call actions, meeting scheduling, outbound personalization, pipeline forecast roll-ups, deal review and proposal drafting — as well as customer success, marketing, HR, support, operations and research functions. The platform runs evaluations against live runs to chart pass rates and detect drift, and benchmarks multiple third-party models (including Anthropic, Google, OpenAI, GLM and Kimi families) per agent to select the lowest-cost model that clears a quality bar. Enterprise controls include SSO, RBAC, audit logs, A/B testing and analytics.

Founding story

Co-founders Daniel Vassilev, Jacky Koh and Daniel Palmer started the company on the premise that roughly 80% of business data is qualitative — text, images, audio, video — and that most businesses use only a fraction of it; making vector-based machine learning accessible to teams without large data science functions was expected to raise productivity across industries. The founders later described traditional automation tools as hitting a ceiling: effective for rule-based processes but unable to handle tasks requiring judgment and contextual decision-making, which motivated the move to AI agents.

Business model

Relevance AI sells access to its agent platform to businesses, with an Enterprise tier priced on a custom, sales-led basis ("Talk to sales") that bundles unlimited agents, tools, users, projects and workforces, 2,000+ integrations, calling and meeting agents, enterprise triggers, agent evaluations, A/B testing and analytics, SSO, RBAC and audit logs, custom actions, custom vendor credits and a dedicated account manager. A self-serve entry path exists via a free trial and an instantly usable chat product. Alongside software, the company offers an embedded deployment team that maps use cases, builds a first set of agents and then trains customer staff to build further agents themselves.

Subscription/usage-based enterprise software with custom pricing negotiated through sales, including custom vendor credits for model usage; a free trial and self-serve sign-up provide entry access.

Traction

The company reports 40,000 agents created on the platform in January 2025 alone and says it has helped thousands of teams deploy agents running millions of tasks a month; its earlier vector platform was cited as serving millions of end users with 100 million weekly API requests. Headcount grew from 4 at Galileo's initial investment to 18, then 19 in 2023 and over 80 by May 2025. Published customer outcomes include Qualified generating $7M pipeline in six months with 35+ agents and a 10x output increase, Send Payments saving 40 hours weekly across global operations, and Zembl reporting 30% higher customer conversion and 60% faster average call time.

Latest developments

In May 2025 the company announced a $24M Series B led by Bessemer Venture Partners with King River Capital, Insight Partners and Peak XV participating, bringing total funding to $37 million; no valuation was disclosed. Concurrently it launched Workforce (visual multi-agent builder) and Invent (text-to-agent generator), opened a San Francisco office, and said it would use the funding to enhance agent product capabilities and support customers in Australia and the U.S. Later blog activity includes posts on adaptive context management for production AI agents and on submitting to the Relevance AI marketplace.