Episode Overview

Podcast: Behind the Numbers: an EMARKETER Podcast
Episode: AI-Driven Media Management, with Gigi and Amazon Ads (Part 1)
Date: December 18, 2025
Host: Jeremy Goldman (Senior Director of Content, eMarketer)
Guest: Adam Epstein (Co-founder and CEO of Gigi)

Main Theme:
This episode explores the evolution of media management in the advertising industry, focusing on how agentic AI is transforming campaign workflows, agency efficiency, and the roles of human advertisers. Adam Epstein discusses the origin story of Gigi, a platform designed to automate and augment media management tasks, particularly within Amazon DSP, and reflects on the broader impact and adoption of AI within agencies.

Key Discussion Points and Insights

1. Gigi’s Origin & Strategic Pivot

• Initial Premise:
  • Gigi was founded on the belief that CTV (connected TV) advertising would increasingly be concentrated within large tech and media companies, leading to closed ‘walled gardens’.
  • Adam’s expectation: "They were incentivized and had the means and were going to become increasingly walled gardens like Amazon and YouTube..." (Adam, 01:24)
• Market Reality and Pivot:
  • Contrary to expectations, major platforms opened up, allowing buying through various DSPs (Demand Side Platforms), notably benefiting Amazon DSP.
  • Gigi shifted focus: Instead of just CTV, customers requested "a full funnel solution," forcing Gigi to reevaluate and become "AI first" or "AI native."
  • “We knew that we needed to be AI first or AI native. And we didn't really know what that meant at the time, so we totally took a step back." (Adam, 02:45)

2. Identifying Rote Media Management Workflows

• Observation:
  • Enterprise media managers perform numerous repetitive and manual tasks; over a million people in such roles across agencies and brands.
• Research Process:
  • The Gigi team spent time observing day-to-day activities of media managers, asking: "What are the buttons that you press in the Amazon DSP that you press the most?" (Adam, 04:40)
  • Specific method: Analyzing the “history tab” in Amazon DSP to determine most-used functions.
• Actionable Insight:
  • Conversion of common actions into “tool calling”—turning frequent tasks into actionable API endpoints that Gigi’s AI could perform for the user.
  • Example use case: “Updating budget flights to extend a campaign for a month." (Adam, 05:33)

3. Agentic AI in Action: Human in the Loop

• Tech Philosophy:
  • Gigi is designed to augment—not replace—human strategy, handling the repetitive tasks so practitioners can focus on strategic value.
• User Experience:
  • "Our customers can just say to Gigi, 'Hey, can you update these budgets?'…Gigi would press all those buttons on behalf of their customer, say, 'This is done.' And the customer would accept that plan from Gigi." (Adam, 06:05)
• Primary Goal:
  • Free media managers from mundane tasks, returning them to higher-value strategic activities.
  • "No one signed up to press a million buttons to extend a flight." (Adam, 07:17)

4. Human Oversight and AI Accountability

• Human Reviews AI Work:
  • Jeremy notes a key insight: "Everybody's thinking about like AI is going to be reviewing the human's work...no, actually the human is reviewing the AI's work." (Jeremy, 08:00)
• Design Decision:
  • Gigi is never permitted to change a bid, budget, flight, or campaign without human approval.
  • Adam frames Gigi as “a really fast intern”:
    • "The mechanism...is such that you assign Gigi work. Gigi says, here's what I'm presenting to you...Here's a natural language rationale of the work that I've done...And now it's ultimately up for you to press accept." (Adam, 09:07)

5. Implications for Agency Business Models

• Agency Efficiency and Margins:
  • Adam predicts agentic AI will be "a massive tailwind for agencies," helping them serve more clients without proportionally increasing staff—potentially transforming operating margins and business valuation. (Adam, 11:53)
• Long-term Vision:
  • Two North Stars for Gigi:
    1. Positively change operating margins.
    2. Positively change how people work: "How can people be more strategic? How can they spend less time pressing buttons...and more time delivering and honing in on the right message at the right time?" (Adam, 13:42)

6. Where AI-Driven Media Management Stands (Adoption Curve)

• Currently Early Adoption:
  • Despite traction, Adam views the industry and Gigi’s solution as "unbelievably early" in the broader adoption cycle. (Adam, 14:22)
• Biggest Challenge Now:
  • Not if the technology works, but whether organizations can adapt their processes and change management to become AI-native:
    • "The test isn't the tech itself, but...the change management required for an internal company to say, 'I want to be AI native...and re-architect our org...in an AI first manner.'" (Adam, 15:23)

7. Change Management and Overcoming Skepticism

• Resistance to Change:
  • Many agencies have legacy processes (e.g., redundant QA steps) that may not make sense in an AI-driven workflow.
  • Example: One agency wanted to download and re-upload bulk sheets even after AI had automated campaign setup—described as "almost like printing out a fax to then print out an email." (Adam, 18:08)
• Real Human Issues:
  • Adam emphasizes respecting these concerns and helping agencies rethink their processes to fully leverage AI’s potential, while still maintaining safeguards for quality.

Notable Quotes & Memorable Moments

• On the Purpose of Gigi’s Pivot:

  • “We knew that we needed to be AI first or AI native. And we didn't really know what that meant at the time, so we totally took a step back.”
    — Adam Epstein (02:45)

• On AI’s Role in Agency Work:

  • "No one signed up to press a million buttons to extend a flight. No one signed up for, on a quarterly basis, updating creative tags for a week. So new campaigns can run..."
    — Adam Epstein (07:17)

• On Human + AI Collaboration:

  • "Everybody's thinking about like AI is going to be reviewing the humans' work...no, actually the human is reviewing the AI's work. I thought that was a really interesting way of framing it."
    — Jeremy Goldman (08:00)

• On the Design Philosophy:

  • "We've very intentionally made Gigi work so that Gigi does not change a bid, budget, flight or campaign without human direction...That's actually how...hiring a junior employee...They need to gain trust."
    — Adam Epstein (09:07)

• On Industry Transformation:

  • "If used successfully, an agency's operating margin should fundamentally change. And the degree to which agencies might be valued on a EBITDA multiple might fundamentally change as well, and they might get rewarded by being valued on a software multiple."
    — Adam Epstein (12:30)

• On Change Management:

  • "The test isn't the tech itself, but the test is the change management required for an internal company to say, like, I want to be AI native..."
    — Adam Epstein (15:23)

• On Outdated QA Processes:

  • "Downloading a bulk sheet to upload it into the DSP after Gigi-ically builds the campaigns is almost like printing out a fax to then print out an email."
    — Adam Epstein (18:27)

Timestamps for Important Segments

• Gigi’s founding and pivot to AI: 01:24–04:16
• Researching media management pain points & agentic workflows: 04:40–06:37
• The 'human in the loop' approach to AI media management: 06:37–08:15
• Allowing user oversight, AI as 'junior employee': 08:15–11:16
• Impact on agency business models & margins: 11:53–13:57
• Gigi and AI adoption curve: 13:57–16:31
• Navigating skepticism and change management in agencies: 16:31–19:57

Conclusion

This episode offers a detailed and candid exploration of how agentic AI like Gigi can automate rote media buying tasks, what it takes to build effective human-AI collaboration, and why the real battle for AI adoption lies in change management and adapting agency workflows. Adam Epstein’s insights are grounded in real industry examples, balancing excitement for AI-powered efficiency with respect for the complexities of human organization and process change.

Listeners gain actionable perspectives on what it means to be truly “AI native” in advertising, framed in a conversational, approachable style.