Podcast Summary: "Using AI at Work"

Episode 92: Using AI-Powered Synthetic Audiences For Smarter Marketing with Justin Brooke
Host: Chris Daigle
Guest: Justin Brooke (Founder of AdSkills.com and Agentskills AI)
Date: February 23, 2026

Episode Overview

This episode explores how AI-powered "synthetic audiences"—or virtual focus groups—are revolutionizing the marketing landscape. Chris Daigle interviews veteran marketer and AI practitioner Justin Brooke, diving deeply into practical applications, technical workflows, results, and the shifting culture around AI adoption in marketing. Key takeaways include the creation and use of synthetic audiences, the tech stack behind these advancements, and the broader implications of agentic AI beyond just advertising.

Main Discussion Points & Insights

1. Genesis of Synthetic Audiences

• Traditional marketing relies on customer Personas (ICP/avatar), refined through trial, error, and real-world feedback.
• Justin realized that large language models (LLMs) are "giant pretenders"—expert at adopting roles—so he began leveraging them as synthetic customers for feedback and messaging ([08:45]).

2. How Synthetic Audience Workflows Operate

• Simple Process:

  1. Feed ad copy or marketing material into the system.
  2. Have the AI "pretend" to be the Persona and provide feedback.
  3. AI acts as a copywriter, improving the copy based on feedback.
  4. Introduce a synthetic data analyst to score and select the best versions ([10:00]-[15:00]).
  5. Repeat and iterate for consistent gains.

• Evolution: Expanded from single Persona processes to passing messages through groups of diverse synthetic Personas—essentially "virtual focus groups" ([12:45]).

• Tangible Results:

  • Improved engagement metrics (ads, social posts).
  • Notable case studies show 4-19x ROAS (Return on Ad Spend) in client projects ([14:30]).

“Now all of a sudden, their ads are becoming way more profitable than anything they were. The way I’ve been selling this—because people don’t know, it’s so new—I just say ‘Hey, give me an ad. I'll run it through my magic...’ And it just works better.”
— Justin Brooke ([15:20])

3. Industry Momentum & Validation

• Early adopter skepticism is rapidly subsiding as major organizations and validation sources—Harvard, Stanford, Colgate, New York Times—demonstrate the accuracy of synthetic audiences ([16:00]).
• “Virtual focus group” is more descriptive, but “synthetic audiences” is gaining as the industry term ([16:46]).

4. Economic and Workflow Advantages

• Traditional focus groups are slow, expensive, and largely unscalable. Synthetic audiences can offer high-quality feedback in real time for cents per run ([17:32]-[18:18]).
• Transformational for any business investing in advertising or messaging—not just Fortune 500 companies ([18:18]).

5. Best Practices for Building Effective Personas/ICPs

• Depth Over Breadth:
  • Avoid superficial, one-sentence Personas.
  • Ideal synthetic Persona is 1,200+ words, built from interviews, surveys, support logs, or rich behavioral data ([22:21]-[23:30]).
• Key Persona Elements:
  • Demographics: Age, gender, education, earnings.
  • Psychographics: Motivations, preferences, books, stores.
  • Lifestyle, pain points, empathy map (what the customer sees, hears, thinks, says, feels—[24:21]).
• Good input = trustworthy, repeatable output.

“I’d rather have three 1,200-word Personas than 500 one-sentence personas. You’re just not going to get the accuracy otherwise.”
— Justin Brooke ([23:50])

6. Impact on Marketing Culture

• Varying responses from traditional marketers:
  • Some “Mad Men” types want to avoid feedback (“they want their sales pitch to win”).
  • Many dismiss synthetic audiences as “toys” due to lack of understanding.
  • A small group tries it, sees ROI, and shifts workflows immediately ([19:50]-[21:09]).
• Synthetic audience validation is becoming analogous to tracking—absolutely essential, soon to be standard ([21:09]).

7. Tech Stack & Implementation

• Multiple options for all levels of technical proficiency:
  • Low/No Code: Mind Studio, N8N, Make, Claude projects.
  • Advanced/Agentic: Crewai, LangChain for full agents ([30:13]).
• Cost: As low as $0.19 per iteration ([18:18], [30:10]).
• Non-marketing use cases emerging: sales training, HR conversations, startup pitch preparations—any setting where simulated diverse feedback is valuable ([31:38]-[34:00]).

8. Beyond Marketing—Agentic AI

• Justin's foray into OpenClaw (an open-source agent framework).
• Practical Uses: Reporting, custom dashboards, automated support, data entry, streamlined marketing iteration ([43:07]).
• Standing caveats: Security and privacy, especially when running agentic AI with deep system access. Cloud-based instances (e.g., DigitalOcean) mitigate risks ([35:17]-[39:50]).

“At the end of the day, [the agent’s] brain is still the same AIs we're using today…But the plus one is it makes it so you don’t have to learn how to use them better.”
— Justin Brooke ([38:29])

9. Advice for Leaders and Practitioners

• Don’t get overwhelmed. Focus on your domain.
• “Pick your lane” and go deep rather than trying to master every AI trend ([49:55]).
• No coding background needed; transformative value lies in changing how you think and seeing opportunities exact to your business ([47:11]).

Notable Quotes & Memorable Moments

• On the future of marketing:

  “I believe 2027, 2028, this is step one of marketing… before you go deploy it against your in-house list or spend ad dollars, you’re going to run it through your little virtual focus group, synthetic audience.”
  — Justin Brooke ([18:18])

• On the accuracy & speed of synthetic audiences:

  “New York Times…created what they call their digital twin of the beta test group…they’re finding that this digital twin audience is 92% as accurate as their human beta testers were."
  — Justin Brooke ([27:13])

• On the need for robust Persona data:

  “You can’t just give it a one-sentence ‘pretend to be this person’…when we’re creating a Persona, it’s 1,000-1,200 words. When Stanford and Harvard did this, they were doing two-hour-long interviews…”
  — Justin Brooke ([22:21])

• On the AI learning curve:

  “I believe in sharing the ideas instead of hoarding the ideas. As I shared [the workflow], people gave me this idea, that idea, and I just started collecting them…”
  — Justin Brooke ([28:48])

• On what’s coming and how to keep up:

  “For me, I’m just keeping track of AI for marketing. What’s your thing? Pick a couple of models. I pick one lane—and that’s how I can filter it and bring value.”
  — Justin Brooke ([49:55])

Timestamps for Key Segments

| Time | Segment / Quote | |-----------|-----------------------------------------------------------------------------------------------| | 03:13 | Justin’s core takeaway for listeners: Virtual focus groups/synthetic audiences are real & vital| | 08:45 | How Justin started using AI as synthetic customer feedback | | 12:45 | Expansion from “pretend customer” to “virtual focus groups” | | 14:30 | Real-world results: campaign ROAS improvement, client case studies | | 16:00 | Harvard, Colgate, Stanford adopt and validate the method | | 18:18 | “This is step one of marketing by 2027/28…” | | 19:50 | Traditional marketer reactions: skepticism, adoption bubbles | | 21:09 | Synthetic audience validation: The “tracking” of the front end | | 22:21 | Building robust ICP/Persona for accuracy | | 24:21 | Persona details: Demographics, psychographics, empathy mapping | | 27:13 | New York Times “digital twin” case study (92% accuracy) | | 30:13 | Tech stack options for all skill levels | | 31:38 | Expanding synthetic audiences beyond marketing: sales, HR, board prep | | 35:17 | OpenClaw/Agentic AI exploration and lessons | | 39:57 | Real-world agentic AI workflow for business operations | | 43:07 | Practical uses: dashboards, support, data entry | | 47:11 | Fast evolution—what’s “early” now will be sticks & rocks soon | | 49:55 | “Pick your lane”—filtering AI information overload for effectiveness |

Where to Follow and Learn More from Justin Brooke

• Twitter/X: @AgentSkills
• Web: AgentSkills AI
• LinkedIn/Google: Search "Justin Brooke"

“There’s not a lot of people that I follow on X that I don’t have to run through a filter of, you know, the BS sniffer. And that’s one thing I appreciate about Justin…”
— Chris Daigle ([49:07])

Closing Insights

• Synthetic audiences are rapidly moving from “toy” to essential best practice.
• AI-driven workflows, including agentic AI, can unlock value far outside marketing—think sales, HR, leadership, and more.
• Practical, non-technical adoption is possible today; focus on your domain’s use cases.
• The pace is blistering—learn, adapt, iterate, and don’t get overwhelmed by “everything AI”; select and master your relevant area.

Recommended Listening for Marketers, Executives, and Business Leaders Looking to Harness Practical AI Today.