Maia by Make

Make finally released Maia, an agentic coworker that lives directly on top of your scenarios. After using it to optimize a scenario burning 40% of my credits, I'm convinced it's their biggest release of the year. In this video I walk through how I actually use Maia day to day: not for building from scratch, but for asking questions, getting architectural advice, and letting it implement and debug changes while I watch. I take a scenario a coworker built, ask Maia to explain it in plain English, then have it propose and implement a real optimization. Then I test it against a change I'd already made myself that morning, and Maia found the more efficient path. What I cover: What Maia is and why it beats managing Make via Claude/ChatGPT + MCP (domain knowledge vs. hallucinated module configs) Launching Maia inside any scenario: continue where you left off, improve current scenario, fix errors Asking Maia to explain an inherited HubSpot "mark deal unresponsive after 21 days" scenario module by module Reading Maia's thinking trace and its skills as it pulls module configs and blueprints Identifying the real bottleneck: 60 emails processed per run, and why the filters weren't the lever Maia's follow-up action buttons, "implement all," and its self-generated to-do list Second use case: conditional email copy for Ireland (non-alcoholic drink) without extra operations Why Maia's formula-inside-Gmail approach beat my Make Functions version and removed a whole step Running the scenario from chat, Maia noticing the run and reviewing the execution logs itself Maia debugging a failed HubSpot module mid-run, deleting modules, and iterating to version 4 Tools/features touched: Maia (open beta) — explaining, optimizing, implementing and debugging scenarios in-canvas Make scenario version control — rolling through v2, v3, v4 safely Make Functions — my original approach, later simplified away HubSpot, Gmail, Airtable — the modules being optimized -------- 00:00 What Maia actually is 01:02 Launching Maia in a scenario 01:41 Maia vs Claude and MCP 02:08 Finding the credit-heavy scenario 02:51 Asking Maia to explain it 04:10 Spotting the real inefficiency 05:07 Asking for optimization advice 05:21 How Maia's skills work 06:20 Suggested actions and implement all 07:05 Second case: Ireland email logic 09:06 Formula inside the Gmail module 10:34 Version control and test run 11:34 Back to the big scenario 12:38 Maia debugs a failed run 13:52 Version four and takeaways --------