
A GTM engineer builds your data workflows
You want 200+ data providers in 1 place, waterfalls across 150+ databases, and Functions your builders publish to reps' AI tools.
RevOps trains the agents once for everyone, and each rep runs them from the record or by asking Eva.

GTM engineers orchestrate. Reps use what they package.

New plays and sequencer pushes wait for an Ops-built Function
GTM data and workflow platform for GTM engineers, with a chat layer for reps
Ops builds and orchestrates. Every rep asks Eva.






By default, any rep can ask Eva for a new play
Workspace of specialized digital colleagues, built for teams

You want 200+ data providers in 1 place, waterfalls across 150+ databases, and Functions your builders publish to reps' AI tools.
RevOps trains ready-made agents once on your ICP, personas and value proposition, and reps launch them from the record or ask Eva in Slack, Teams or their AI tool.

Keep your builders' Clay workflows. Evergrowth works on the same CRM records: findings go to EG fields, emails and phones to fields you map.

Clay counts 17,000+ customers. Each card below comes from Clay's own pages, read 2 Oct 2026.
Waterfalls run across 150+ databases, and you pay only for the data that is found.
Describe a play and Sculptor or the CLI turns it into a workflow builders can open, test on 1 record and adjust. Workflows is in open beta.
Clay MCP works in 6 AI tools, including Claude, ChatGPT, Copilot and Glean, within credit limits Ops sets per rep.
Once Ops builds it, a play can start when a record joins an audience, on a schedule or from a webhook.
Clay describes its own team as a centralized build for decentralized use: builders package each play, reps call it. In Evergrowth, RevOps, GTM engineers or Marketing Ops set the agents up once, and every rep reaches them through Eva.

GTM engineers orchestrate. Reps use what they package.
GTM engineers or RevOps build each play as a Function, agent or workflow.



New plays and sequencer pushes wait for an Ops-built Function, and Ops controls CRM writes. Search and default lookups need no Function.
Ops builds and orchestrates. Every rep asks Eva.
Ops builds the training and orchestrates the agents.















By default, any rep can ask Eva for a new play.
Eva answers in Slack, in a direct message or when someone mentions her in a channel, in Microsoft Teams in a direct message, and in ChatGPT, Claude and Copilot through MCP. It is the same Eva as in the workspace, with each rep's own permissions, and her answers link back to the workspace.
The Evergrowth connector brings Eva into ChatGPT, Claude and Copilot. She answers from the same workspace research, and before she launches anything she shows the cost and asks.

What do we know about Acme Corp, and where should we focus?

The workspace research points to a RevOps-led consolidation project.


Help me prepare for a conversation with Dana Whitfield at Acme Corp.

The answer is being provided in the Evergrowth panel.

Run our saved “Find RevOps contacts” workflow on Acme, Northwind and Contoso.

Estimated execution: 6 credits, plus credits for each new verified email. Conversation usage is separate.
Shall I run it?
Eva is asking you to confirm the run in the Evergrowth panel.
The research lands on the account record, not in a rep's own chat. Launch agent lists the account agents, each with a hint on when to use it, and every launch shows its estimated credits before the rep confirms.
A rep asks for a new play, and Eva works out the steps: she reads the guides and existing plays, then writes it into the team's Play Agents. Generating it on persona-fit contacts is a separate launch, at 1 credit per play per contact.
RevOps writes the criteria once, so every rep's accounts are judged the same way. Each answer carries its reasoning and its sources when found, a rep can rate it with a thumbs up or down, and RevOps maps the fields to your CRM.
Publish them and the agents that use them work from the new version for every rep. By default, only the Admin and Sales Ops teams can change them; reps read them.
Clay's column comes from Clay's pages and docs, Evergrowth's from the product itself. Both checked 2 Oct 2026.
| Dimension | ||
|---|---|---|
| Who runs it | ||
| Core model | A data and workflow platformGTM data and workflow platform for GTM engineers, with a chat layer for reps | A workspace of digital colleaguesWorkspace of specialized digital colleagues, built for teams |
| Who orchestrates | GTM engineers and RevOpsGTM engineers and RevOps build and orchestrate; reps use search and approved Functions from their AI tool | Ops orchestrates, every rep runsOps builds the training and orchestrates the agents; every rep runs them from the record or by asking Eva |
| What RevOps owns | Every play, permissions, budgetsBuilding each play as a Function, plus rep permissions and credit budgets | The training center and who edits whatThe training center, and who can edit what, set per team |
| What a rep can do | ||
| Where reps work | Only in their AI toolIn their AI tool, through Clay MCP: Clay's Sales Rep role cannot open the Clay workspace | Workspace, Slack, Teams or AI toolIn the workspace, Slack, Teams or their AI tool through MCP, with the same Eva in each |
| In a rep's AI tool | Search, lookups and draftsSearch, lookups, drafts and the Functions Ops enables | Eva, running the agentsEva, who runs the agents and writes plays, as she does in the workspace |
| A new play | Ops builds it as a FunctionBuilt by Ops as a Function, then enabled for MCP by an admin | RevOps writes it, or EvaWritten by RevOps, or by Eva when a rep asks; by default reps can create plays |
| Data, CRM and automation | ||
| Research | AI columns builders tuneAI columns and Claygent agents, built and tuned by builders | Agents set up once, reasoning shownQualification and research agents RevOps sets up once; each answer shows its reasoning |
| Contact data | Waterfalls across 150+ databasesWaterfalls across 150+ databases, built in Clay; reps can ask for emails and phones from their AI tool | Persona-fit contacts, billed on findsContact Finder picks persona-fit contacts; the Email/Phone Finder bills only what it finds |
| CRM | Ops controls write-backsOps controls CRM write-backs centrally; the default rep tools include no CRM write | 2-way sync with 6 CRMs2-way sync with 6 CRMs; findings write back as EG fields, emails and phones where you map them |
| Always-on runs | Built and run by OpsAudiences, schedules and webhooks, built and run by Ops; Workflows is in open beta | Scheduled, or on a saved viewScheduled workflows, daily, weekly or monthly, or when a record enters a saved view |
Evergrowth's column: the product as it ships, checked 2 Oct 2026.
ARIS rolled the agents out across its regions. These 4 changes come from its customer story.
“Seeing reps go into an AI system and start running their own agentic workflows, that's something I wouldn't have imagined two years ago.”
If your reps only need search, lookups, drafts and the Functions Ops builds, Clay MCP covers that. Evergrowth is for teams that want every rep running qualification, research and plays on their own accounts, from agents RevOps has trained once.
Not quite. Clay MCP gives reps search, lookups, drafts and the Functions Ops enables. Evergrowth puts Eva in those tools and in Slack and Microsoft Teams: the same Eva as in the workspace, running the same agents.
Yes, for what a builder does best: data workflows in Clay, and in Evergrowth the training center, research signals and team permissions. What changes is the queue: by default, reps no longer wait for plays on signals you already track.
Yes, through your CRM. Evergrowth syncs accounts and contacts with 6 CRMs, including HubSpot and Salesforce, and writes findings back as EG fields. Emails and phones go to the fields you map, and inbound sync runs on a schedule.
No. Every plan includes unlimited users, and you pay for the credits the agents use, from €747 a month.
Your CRM data, research outputs and agent activity are never used to train or fine-tune models. Not by us, not by our providers.
Each workspace is logically isolated, so your data is never accessible to another customer.
If you cancel, your data is returned or deleted, your choice. If you make no request, all personal data is permanently deleted within 30 calendar days.
The detail, the sub-processors and the DPA are on Security and trust.
Under the MSA, everything you upload and everything the agents produce from it is yours.
A live walkthrough on your ICP: agents qualifying, researching, finding contacts and writing plays. Bring your GTM engineer.
Request a demo