Podcast Summary: The Marketing Millennials –

"The Biggest Mistakes Marketers Make with AI (And How to Fix It)"

Daniel Murray with Phil Lakin, Head of Enterprise Innovation at Zapier

Episode 350 | September 19, 2025

Overview

In this episode, host Daniel Murray interviews Phil Lakin, Head of Enterprise Innovation at Zapier, about the most common mistakes marketers make when adopting AI, practical solutions, actionable quick wins, and the evolving skillset required to thrive in the AI-powered marketing ops landscape.

Lakin, with extensive hands-on experience in no-code automation, community building, and embedding AI into enterprise workflows, shares real-world examples, tactical guidance, and a fresh perspective on how marketers can actually leverage AI for impact—without falling for hype or resorting to expensive point solutions.

Key Discussion Points & Insights

1. Phil Lakin’s Background & Zapier’s AI Evolution

• [01:14] Lakin explains his path from founder of No Code Ops to leading enterprise innovation at Zapier.
• Transitioned Zapier from being “the go-to tool for automation” to an “AI-first approach,” including the AI Transformation package, which embeds teams at companies to help them go AI native in one year.

2. How AI Works in Zapier—and Common Misconceptions

• [03:30] Breakdown of Zapier’s AI integration:

  • AI Features: Copilots for building zaps, system-building, error resolution.
  • AI in workflows: “AI step” that can take form submissions and generate personalized SDR outreach. This automates low-value work and ensures learnings are centralized.

    “Not every SDR needs their own GPT… all getting the best version of a first draft and that is super powerful.”
    — Phil Lakin, [05:18]

  • AI Agents: Context-aware, less deterministic automations that can access knowledge bases, make inferences, and act more like empowered interns than simple workflows.

• AI ≠ Magic Eraser:

  “People think…you can just throw everything at an AI or at a chat or at an agent and just be like, cool, it’s just going to fix things. It’s not.”
  — Phil Lakin, [07:20]

• Critical Framework:

  • Know when to use a workflow, an AI step, or an agent.
  • Always battle-test agent outputs—like onboarding a new intern.

3. The Biggest AI Mistakes Marketers Make

i. Lack of Process Clarity

• [09:07]

  “You have to be able to do the process manually and know what looks good manually before you start automating the process… It’s the same as AI.”
  — Daniel Murray

ii. All-or-Nothing AI Thinking

• [10:29]

  “They think…when AI does something it means AI has to do everything. And that’s not true. You can have AI be a part of a human-in-the-loop process.”
  — Phil Lakin

• Practical Example: AI grades inbound leads; high scores go to AEs, low scores are sequenced, “maybe” scores alert humans for final judgment ([11:00]).

iii. Falling for Expensive Point Solutions

• [15:34]

  “The biggest mistake is just…seeing something and being like we want to buy that, we have to integrate it…versus what if I just gave you the building blocks…for 5% of the cost.”
  — Phil Lakin

• Cost & Control:

  • Platform consolidation using orchestration tools can replace multiple point solutions—less cost, more customization—even considering the maintenance tradeoff.

4. Quick Wins & Practical AI Uses for Marketers

i. Prompts: Context Is King

• [19:06]

  “Overload with context. Always put in a lot of context… The more specific you can get with AI, the better you’re gonna get an output for.”
  — Phil Lakin

• Prompting Guidance:

  • Tell AI your goals, style, desired research (e.g., “Act as if you’re X and Y working together”), first give all relevant info, then iterate.

• Daniel’s Tip:

  “Too many people are not spending enough time on the prompt… Just like an intern, the more information you give them, the better job they would do.”
  — Daniel Murray, [20:26]

ii. Automating Routine, Repetitive Tasks

• [22:44]
  • Identify repetitive ChatGPT tasks (e.g., weekly reports, email rewrites).
  • Use Zapier to trigger AI steps automatically and send outputs for approval or delivery—no more copy-pasting between platforms.

iii. Human-in-the-loop Workflows

• Combine AI scoring/drafting with human oversight at key decision points (qualified leads, message approval).
• Leverage AI for “boring” but necessary work (e.g., documentation, research summaries).

5. Community, Learning & Evolving Marketer Skillsets

i. Where to Learn/Find Inspiration

• [25:02]
  • noCodeOps — free Discord run by Zapier, for ops/automation professionals.
  • Operations Nation, Generalist World.
  • Curate TikTok/LinkedIn feeds by seeking out AI/marketing innovation content.

  “I’m learning so much stuff from just curating my own feeds.”
  — Phil Lakin

ii. Generalists Are Rising

• Marketers must blend ops, strategy, content, automation, and data (“secret weapon” mindset).
• “OPS people have to become strategists, strategists have to become OPS people…the first one to do both is gonna win.”
  — Phil Lakin, [27:12]

iii. Becoming a Manager—of Humans & AI

• Managing a team of AI agents requires classic people management skills—reporting, assigning, giving feedback—as tools get easier.

  “The folks that are going to be most well suited... are going to get a little uncomfortable... learn the language of the OP stuff... go to an AI bootcamp...”
  — Phil Lakin, [27:12]

• A new emerging role is someone whose job is agent management.

6. Agent Examples and “Secret Weapons”

i. AI Information Summaries

• [36:36] Lakin created a Zapier agent that digests Slack/Drive updates, filters by job-relevant criteria, and relays only what’s relevant.

  "I built a Zapier agent on top of Glean... every week just tell me the things from last week that are updates across all these relevant things that I need to know about, and then... I'll have it read it out loud to me."
  — Phil Lakin, [37:50]

ii. Documentation, ADHD & “Neuro-Spicy” Productivity

• Dictate documentation out loud, let AI transcribe, organize, and format.

  “Literally talking out loud to documentation is the best way to create it.”
  — Phil Lakin, [40:15]

Notable Quotes

• On AI’s real benefit:

  “95% of all this stuff existed before AI. It’s just AI made it super popular.”
  — Phil Lakin, [08:56]

• Approach to prompting AI:

  “If you’re confused about how to write a great prompt...just ask AI, what do you need for me to improve this prompt?”
  — Phil Lakin, [21:13]

• Role evolution:

  “Middle management has become a superpower, which is so funny, because I think middle management was always what I've tried to prove, that we're doing the work...”
  — Daniel Murray, [33:21]

• On marketer generalists:

  “As long as you know the right questions to ask, you’re good.”
  — Phil Lakin, [35:52]

Marketing “Hill to Die On” – Phil Lakin

[41:59]

“In today’s world, any marketer not religiously creating content on their personal brand about their company to learn what works and doesn’t…is not doing their job to the best they could be…”

• Microtesting content on social is the fastest way to learn what resonates.
• One single post (built atop multiple tests) led to $6M in qualified funnel and $500k+ closed revenue at Zapier.

Key Timestamps

• 01:14 — Phil’s background and AI in Zapier
• 03:30 — How Zapier uses AI and biggest AI mistakes
• 09:07 — Importance of knowing process before automation
• 10:29 — All-or-nothing mistake with AI in marketing
• 15:34 — Don’t default to AI point solutions
• 19:06 — How to prompt AI for better marketing outputs
• 22:44 — Quick win: Automating repetitive AI tasks
• 25:02 — Best communities & learning resources
• 27:12 — Generalists, marketers’ evolving roles
• 36:36 — Phil’s AI agent for info summarization
• 41:59 — Lakin’s marketing hill to die on: Personal content testing

Where to Connect with Phil Lakin

• LinkedIn: Philip Lakin (with one L!)
• Website: philiplakin.com (sign up for SMS drops on AI/marketing/opportunities)
• TikTok: Growing presence
• SMS Platform: Uses SlickText for concise, high-engagement info drops

Memorable Moments

• Phil’s agent that auto-summarizes all project updates and reads them aloud, a game-changer for remote teams and “neuro spicy” professionals ([37:50])
• Advocacy for replacing newsletters with SMS drip; “I hate writing email newsletters and I hate reading them” ([47:41])
• Lively, friendly debate on who’s better suited to manage AI teams—AI nerds or people managers ([30:10–33:01])

Tone

Conversational, practical, often witty. Daniel Murray’s “no BS” style elicits tactical details and story-driven answers from Lakin. The advice is candid and built on lived experience.

TL;DR

• AI in marketing isn’t set-and-forget; success requires process, QA, and human-in-the-loop workflows.
• Avoid expensive point solutions—learn automation basics, consolidate where possible.
• Always overload prompts with context—treat AI like a new intern.
• Empower marketers to be more generalist, blending technical and strategic skill sets.
• Micro-experimenting through personal branded content is the fastest growth lever.
• Document, maintain, and communicate—AI won’t do it all for you, but it will do it with you.

Highly recommended episode for any marketer struggling with where to begin, how to scale AI’s real value, or those itching for practical, non-hyped advice on AI orchestration in the real world.