Checkgrow

Operations

AI Agents

Pre-built specialized AI workers that generate LinkedIn posts, emails, ads, articles, and more — grounded in your company knowledge.

Overview

AI Agents are pre-built specialized AI workers — for LinkedIn posts, emails, ads, blog articles, outreach, OKRs, and more. Each agent runs a brief-driven chat: you describe what you need, it generates multiple options grounded in your company knowledge.

Every run is saved as delivery history. You can edit, approve, translate, share, export to PDF, chain results into another agent, or convert them to Tasks. Which agents you see depends on your per-user agent permissions.

Briefs are grounded in your Knowledge Center

Agents pull from your Buyer Personas, Products, Use Cases, and scraped pages — so a two-line brief still produces on-brand, company-specific output.

Image #1 — AI Agents list with agent cards, category badges, token counts and Open Agent buttons

Who needs it

AI Agents suit anyone who produces recurring content or outreach:

  • Marketers — generate LinkedIn posts, ads, and blog articles in multiple options at once.
  • Sales teams — draft outreach emails tailored to your Buyer Personas.
  • Managers — draft OKRs and structured plans from a short brief.
  • Reviewers — approve deliveries in bulk and hand them to the team as Tasks.

Credits per run

Each agent run charges credits. Every agent card shows its token count, and the AI Assistant confirms with you before running an agent on your behalf.

How to use it

  1. Find your agentOpen AI Agents from the sidebar. Use "Search agents..." or the Filter Agents popover (by Category and Channel) in the secondary sidebar, then click Open Agent on the card you want.
  2. Write the briefType into "Tell me what you'd like to create...". Add context with Attach Files, Scraped Pages, the Web toggle, or Merge to reuse previous responses as context.
  3. GenerateClick Generate Now. The button shows "Generating..." while the agent produces multiple options as result cards.
  4. Refine an optionClick a result card and describe the change in "Ask to refine this option...". Select several cards to refine them together ("Ask to refine selected options..."). You can also edit fields inline, like options, translate, or upload media.
  5. Approve and hand offSwitch to Table view — all deliveries, tick the rows you want, and click Approve (n). Approved deliveries flow to Tasks for execution. On a single option, use Action → Generate Task.
Image #2 — An open agent response with the generated option content, Sign Agent and New Request controls

Working with results

Every run is kept as delivery history on the agent page — here is what each part of it does.

Result cards

Each option supports inline field editing, like, an "Approved" badge, delete, media upload, and a translate dropdown. The Generate More card adds more options. The per-option Action menu offers Generate Task, Send to Integration, and Generate PDF.

Deliveries table

The Table view — all deliveries lists every option across responses. Use Filter by response, Customise columns, per-row Approve toggles, bulk Approve (n), and bulk like or delete.

Chaining agents with Sign Agent

Click Sign Agent to open the "Sign AI Agent" dialog — "Select an agent to generate content using this response as context". On success you'll see the toast "Content Generated". Use it to turn a blog article into LinkedIn posts, for example.

Responses, history, and archive

Toggle the secondary sidebar to Responses and switch between History and Archive, with per-agent search. On an agent page you can Share, export to PDF, or Archive / Unarchive; New Request starts a fresh brief.

Models

The agent list header has a model dropdown: Claude Sonnet 4.6, Claude Haiku 4.5, GPT-4o, GPT-4o mini, Gemini 2.5 Pro, and Gemini 2.5 Flash. Click Refresh to reload the list.
Image #3 — Deliveries table listing every option across responses with per-row Approve buttons

Tasks — where approved deliveries land for execution

AI Assistant — find and run agents from chat

Buyer Personas — audience knowledge that grounds every brief

Products — product knowledge agents draw on