The Digital CX Podcast: Driving digital customer success and outcomes in the age of A.I.

There are a handful of genuinely useful ideas scattered through the episode — tracking developer commits to keep courseware current, the time-redistribution framing for AI-assisted work, and voice-first prompting — but they are surrounded by dense filler, mutual affirmation, and meandering tangents that dilute the signal considerably.

“most of my work, like 60, 70%, is in the planning and the setup and like crafting the framework to make the work happen. And then the work happens, and then I spend probably 20% of my time reviewing the output of the work.”

“you look where your developers are pushing their commits and have an AI agent or something tracking that because then you are getting what is different, whether they've told you or not.”

The episode recycles several widely-circulating AI takes (culture over technology, leapfrogging tools, entrepreneurial mindset) with modest added texture; the director/IC/manager time-split framing and the commit-tracking-for-courseware tip are the closest things to genuinely fresh angles, but neither is deeply developed.

“this whole AI revolution is about culture change, it's about getting people comfortable with it”

“I assign work almost like a director level, I watch it happen for like 10% of my IC level, and I have 20% like the manager who's checking off”

Eric Mistry is a credible hands-on practitioner — currently doing AI transformation work inside Zapier with real client exposure — but he is a mid-level individual contributor, not a founder or executive who has scaled something significant, and his claims rest largely on personal experimentation rather than organisational outcomes.

“in September I got invited by Zapier to basically do my dream job, which is AI transformation, yeah, but also through a teaching and learning lens”

“one part of my job is working directly with clients, like actually sitting in with their company, talking to them”

The episode scores above average on specificity because Eric names concrete tools (Cursor, Claude Code, Codex, Clueso, Lovable, Whisperflow, Chrome DevTools MCP, Tailwind CSS, PARTA) and shares a few first-person anecdotes with numbers, but there is zero business-outcome data, no customer metrics, and no broader evidence base beyond personal workflow experiments.

“I want to build a mobile game that helps me explain concepts of Zapier... I put it through the new Chat GPT in February. One shot.”

“I spent like 20 bucks of credits, which is a lot of credits. Um failed.”

The host asks mostly open setup questions and responds with frequent agreement rather than probing follow-ups; the good/better/best framework question is the sharpest moment in the episode, but there is no productive challenge or disagreement anywhere in the conversation, and several threads are opened and dropped without resolution.

“like, okay, you're a your basic, you're you're kicking the tires on vibe coding stuff and trying to get into it. Like, what does that look like? What is level two and what does level three look like?”

“Yeah, totally.”