By playing, you agree to uListen's Terms and Privacy Policy and YouTube's Terms of Service.
How I AIHow this PM uses MCPs to automate his meeting prep, CRM updates, and customer feedback synthesis
Summaries, chapters, transcripts, and FAQs on this page are made by uListen, not YouTube.
At a glance
WHAT IT’S REALLY ABOUT
Zapier PM shows MCP connectors for faster, smarter customer workflows
- This episode demystifies MCPs by reframing them as “app integrations for your AI tools,” enabling an AI client to both read knowledge from your apps and take actions inside them.
- Reid demonstrates Zapier’s MCP approach: build custom tool collections from 8,000+ apps/30,000+ actions, then connect a single server URL to tools like Claude, ChatGPT, or Cursor.
- He shares practical workflows: daily meeting research, post-meeting note logging into Coda/HubSpot, and deterministic automations for longer-running data lookups (e.g., Databricks → Gemini → Coda).
- The conversation emphasizes a “virtuous cycle” where support and chatbot transcripts are synthesized into new FAQ entries, keeping internal and external knowledge bases continuously up to date.
IDEAS WORTH REMEMBERING
5 ideasThink of MCPs as “connectors,” not a new abstract framework.
Reid suggests ignoring the term and focusing on outcomes: give your AI access to knowledge in your apps and the ability to take actions in those apps—those are the two core jobs MCPs enable.
Custom tool bundles reduce chaos and improve reliability.
Rather than adding many separate MCPs with overlapping capabilities, Zapier lets you build purpose-specific “collections of tools” (per client or per use case) and restrict them (e.g., only certain Coda docs or Evernote notebooks).
Claude Projects can act as a tool-usage playbook, not just a knowledge vault.
Reid encodes step-by-step instructions for which tools to call, in what sequence, and where fields should map—making multi-tool execution noticeably more accurate and repeatable.
Use agentic MCP interactions for “in-the-moment” work; use deterministic workflows for long-running steps.
MCP tool calls inside chat clients are time-bounded and can struggle with multi-minute research. For slow lookups (e.g., Databricks enrichment), Reid uses a Zapier workflow that runs asynchronously and writes results back to a doc for later consumption.
Automate meeting prep to eliminate ‘showing up cold’ to customer calls.
His workflow looks up the attendee/company, product usage, prior interactions, and internal context (e.g., Slack or Glean), producing a concise daily plan so he enters interviews informed even when leads arrive via referrals/Calendly spread.
WORDS WORTH SAVING
5 quotesDefinitely don’t think about the word. It really just is like app integrations for your AI tools.
— Reid Robinson
The two things we see people wanting to do is… giving their favorite AI tool the access to knowledge that lives in their apps, as well as giving them the ability to actually do things in those apps.
— Reid Robinson
Check out Claude Projects… [you can] provide very detailed instructions… how it should use tools, in which order it should use tools, what data should go where.
— Reid Robinson
If you could run ChatGPT in your sleep, what would you do?
— Reid Robinson
Let’s say you had the perfect team with infinite time… your perfect support team… would look at every support question and go see, do we have the right help desk content here?
— Claire Vo
High quality AI-generated summary created from speaker-labeled transcript.
