Relevance AI pricing and review

Reachkit makes a competing cold email tool. Every price here was checked on October 15, 2026 against Relevance AI’s own pricing page and docs.

Relevance AI pricing starts at $29 a month for Pro, or $19 a month billed yearly, according to its pricing docs. The public pricing page shows no prices: on October 15, 2026 it listed Enterprise alone, behind “Talk to sales”. The docs still describe upgrading in the app, but it isn’t clear whether new accounts can buy Pro or Team without sales. Plans are priced per organization, not per seat.

Two meters. An Action is one run of a Tool, and a failed run still counts. Vendor Credits pay for the AI model “with no markup”; your own model API keys skip them. Plan Actions reset at each renewal. One worked example in the plans doc shows Pro with 3,000 Vendor Credits rather than 10,000.

Extras. Top-ups cost $80 per 1,000 Actions and $20 per 10,000 Vendor Credits. Team adds calling and meeting agents; Enterprise adds SSO, audit logs, agent evaluations and Bosh, the BDR agent.

What a typical setup costs (my math): Pro works out to about $11.60 per 1,000 Actions, so top-ups cost roughly seven times as much. Pro plus 4,000 extra Actions is $349, the price of Team with its 7,000. An outbound agent running three tools per prospect (my assumption) uses about 3,000 Actions per 1,000 prospects, past Pro’s allowance.

Relevance AI repriced in September 2025, so some third-party pages still show a $199 Team plan or the retired Free plan.

Relevance AI is a platform where teams build AI agents for sales, support and operations work. You build from Tools (multi-step automations), Agents (a model with instructions, knowledge, tools and triggers) and Workforces (several agents wired together, with approvals). Invent, relaunched as Invent 2.0 in September 2026, builds these from a plain-language request.

For sales, the homepage lists ready-made agents such as Research & Enricher and Outbound Prospector. The personalized outbound guide reads each prospect’s CRM and LinkedIn context, drafts a message and queues it for a rep to approve. Contact data comes through integrations such as Apollo and ZoomInfo, and the guide tells you to use sending caps and a warm-up schedule on new domains.

Agents run on models from OpenAI, Anthropic, Gemini and others, connect to over 1,000 apps, and are reachable through an API and a hosted MCP server (free; runs bill as usual), the standard way assistants like Claude and ChatGPT use other apps. Bosh, the BDR agent that handles two-way email, is Enterprise only and set up per company.

Strengths

• Quick to build. A G2 reviewer says setting up agents “takes just a few minutes” (via search results).
• Model costs at cost. No markup on Vendor Credits, and your own API keys work on every plan (plans doc).
• Open to developers. An API, a JavaScript SDK, MCP guides for Claude, ChatGPT and Cursor, and a Claude Code plugin.
• Security paperwork. SOC 2 Type II and GDPR on every plan.

Drawbacks

• Bills are hard to predict. Two meters, failed runs that still cost an Action, and pricey top-ups. Cost is the theme critical reviews repeat most.
• Key features sit higher up. Calling agents need Team; evaluations, SSO and Bosh need Enterprise.
• Not a cold email sender. No contact database of its own, and the outbound guide tells you to add sending caps and warmup on new domains.
• Strict terms. Credit cards only per its plans doc, fees “non-refundable for change of mind”, and unused Actions expire on cancellation (terms).

Relevance AI’s homepage says “4.5 stars on G2”, while search results for its G2 listing showed about 4.3 from around 21 reviews in October 2026. Trustpilot has just two reviews, both negative, with no reply from Relevance AI.

Fans mention how fast agents come together. Some reviewers raise cost (a G2 reviewer, a CEO, calls it very expensive with most features kept for higher tiers), onboarding that “can be a little confusing”, and support: the one-star Trustpilot review’s title starts “There is 0 customer support service”.

Relevance AI fits a sales or ops team with at least one person who enjoys building and wants custom agents for research, meeting prep, CRM updates and outbound drafts in one place.

Look elsewhere if you need a free plan, a fixed monthly bill, or a tool that sends cold email on its own. In outbound, I’d point an agent builder at account research and list work, where AI saves real hours, and leave inboxes and warmup to a sender.

• Lindy: an AI assistant in Slack and email, per active user from $29.99 or $49.99 a month (both shown in October 2026). See the Lindy profile.
• Clay: enrichment tables with AI research per row, a Free plan, and Launch at $185 a month, which includes its email Sequencer. See the Clay profile.
• Artisan: Ava, an AI BDR that sends outreach and answers replies itself, from $280 a month. See Reachkit vs Artisan.
• Apollo: a data source Relevance AI connects to, with its own sequences, a free plan and seats at $65 a month ($49 billed yearly). See the Apollo profile.
• Reachkit (our product): the sending side Relevance AI leaves to you. Verified leads from 500M+ profiles, your own inboxes connected free or new Google or Microsoft inboxes set up and warmed for you ($5 each a month, plus $15 a year per domain), and every reply in the Unibox. First lines an agent writes import from a CSV as merge tags. Growth is $39 a month per workspace, no seat fees. It doesn’t build general agents or answer replies for you.