ChatGPT Apps + AgentKit (MCP) open a new PLG distribution channel. Learn how SaaS teams can ship inside ChatGPT, win discovery, and avoid shallow copilots.
What just dropped:
Apps inside ChatGPT (Apps SDK built on MCP) You can now ship real apps directly inside ChatGPT with your own backend and interactive UI. Discovery happens in-chat, with a public directory and monetization coming soon. Early partners include Canva, Coursera, Spotify, Zillow, Figma, and Booking/Expedia.
AgentKit
A first-party stack to design, test, evaluate, and deploy task-performing agents. Includes a visual workflow builder, file search, connectors, evals, guardrails, and ChatKit (an embeddable chat UI). Helps teams go from prototype to production faster.
So why Product Managers in SalesTech and Prospecting should care?
New PLG distribution channel
You can put an “ICP Builder”, “Lead Finder”, or “Enrichment Wizard” directly inside ChatGPT and get contextual discovery. Think of it as a new app store with a conversational UX and a funnel of hundreds of millions of weekly users!!!
Reliable agentic workflows
Prospecting is a chain of tasks: profile parsing → ICP firmographic match → intent → dedupe → CRM writeback → outreach draft. AgentKit gives you the scaffolding for orchestration, evals, guardrails, and connectors so your team focuses on signal quality and scoring, not low-level plumbing.
Faster in-product copilots
ChatKit lets you embed a copilot in hours, not sprints. Ideal for Account Research Copilots, List QA Agents, or Playbook Builders that call your enrichment or re-ranking services.
Standards
Everything is built on MCP. Expose your private tools (crawler, match-merge, scoring) once, and reuse them both inside ChatGPT Apps and in your own UI.
Happy to help your SaaS business with AI transformation and PLG growth from strategy to implementation, roadmap design, and team enablement. If you’re exploring how to integrate AI copilots or PLG loops in your SaaS, let’s connect.
