Podcast Summary: The Agile Brand with Greg Kihlström®

Episode #835: Qualtrics' Jordan Harper on Using Synthetic Panels to Get Real Insight
Date: March 27, 2026
Guest: Jordan Harper, Principal AI Thought Leader, EDGE Center of Excellence, Qualtrics
Location: Recorded at Qualtrics X4 Summit, Seattle

Main Theme and Purpose

In this episode, host Greg Kihlström discusses a transformative approach to customer insights with Jordan Harper from Qualtrics, focusing on the use of synthetic panels—AI-trained models that simulate audience feedback. The conversation explores how AI-powered synthetic research can augment or even surpass traditional methods for understanding customers, mitigate common survey biases, and unlock experimental possibilities for market researchers—all while addressing skepticism and trust within organizations.

Key Discussion Points & Insights

1. From Traditional Research to Synthetic Panels

Timestamps: [00:55]-[05:00]

• Harper shares his multidisciplinary background (physics, nuclear engineering, marketing technology) and his current role driving AI innovation at Qualtrics.
• Qualtrics is “fully embracing” AI, integrating advanced technology across platforms to enhance customer experience and facilitate actionable insights.
• The conversation introduces synthetic panels—AI models trained on real survey data to simulate human survey responses, addressing issues of fatigue, bias, and scale present in traditional research.

Quote:
"We're integrating AI technology and tooling into pretty much all of our platforms…to help support customers, make their lives easier…get better insights and action from the data."
— Jordan Harper [02:58]

2. Metaphor: Mirror vs. Lens in Research

Timestamps: [05:01]-[08:05]

• Human feedback is like a “mirror”—reflective but often limited by survey fatigue and cognitive shortcuts.
• Synthetic panels act as a “lens”—letting organizations analyze vast, nuanced historical datasets for deeper, actionable insights without overburdening customers.
• Technology has evolved data collection efficiency but mostly made people more accessible, not fundamentally changed insight-gathering—synthetic approaches provide a true leap forward.

Quote:
"What I think is interesting about AI is…it’s actually a new way of leveraging the data and the insight…that we’ve gathered over years and years."
— Jordan Harper [07:15]

3. Synthetic Feedback as “More Honest” Data

Timestamps: [08:05]-[11:42]

• LLMs are “honest” in the sense that they're not affected by self-presentation or social desirability, common in human responses.
• Harper gives an example of how question priming creates consistent bias in human answers, but synthetic panels are largely immune, showing minimal variance as compared to their prior answers.
• Synthetic models remember prior answers for consistency (e.g., if a simulated respondent is “from a rural area,” subsequent answers align) but don’t let emotions or priming alter underlying sentiment.

Quote:
"We know as researchers that humans do not always tell the truth…what we've seen with synthetic a lot of the time is that it doesn't really fall for those same kinds of self-reflection type tics."
— Jordan Harper [09:00]

4. The Changing Role of Market Researchers

Timestamps: [13:20]-[16:51]

• Synthetic panels free up researchers to be more experimental: you can test questions in multiple ways and only present the best ones to real audiences, minimizing fatigue.
• Harper shares a travel industry case: synthetic panels revealed that the “solo traveler” question was ambiguously interpreted, highlighting places for improvement before fielding with real customers.
• While AI surfaces “signals” or anomalies, human expertise is still crucial for interpretation and action.

Quote:
"It expands the ability to do more research, to become more experimental, to test survey design before you put it out to customers."
— Jordan Harper [13:45]

5. Building Organizational Trust in AI-driven Insights

Timestamps: [16:51]-[19:57]

• The key to internal adoption: validation and experimentation side-by-side with human data. Show cases where synthetic and human results match (which is most of the time) and call attention to anomalies for scrutiny.
• Transparency about similarities and differences is critical for building trust among stakeholders.
• Standard LLMs (like ChatGPT, Gemini) tend to give too-consistent answers, lacking the nuanced “messiness” of true human responses. Qualtrics’ AI achieves a distribution much closer to real survey results.

Quote:
"If you ask a general LLM…quite often might get the top answer correct. But…it's 100%. And we know that humans are a little messier than that…Our model…was able to grasp the nuance of human messiness."
— Jordan Harper [18:00]

6. The Future: Predictive, Proactive Market Research

Timestamps: [19:57]-[21:53]

• Market research can become more predictive—simulating reactions to new features or products before launch, without risking customer fatigue or accidental leaks.
• Synthetic panels could permit more proactive, hypothesis-driven research, supplementing but not fully replacing human feedback.

Quote:
"You could, you could use synthetic to ask those kind of questions. You can do more predictive research that would enable you to make decisions that were driven by data in advance."
— Jordan Harper [20:55]

7. Personal Reflections and Advice for Staying Agile

Timestamps: [22:02]-[22:51]

• The X4 Summit highlight for Harper: reconnecting with community and seeing speaker Priya Parker.
• Harper’s advice for staying agile: maintain curiosity, move toward emerging tech, and focus on how progress can improve outcomes.

Quote:
"Always staying curious, always trying to keep an eye on what's next and what's moving…Ask what it can do to make things better rather than…avoid it being a problem."
— Jordan Harper [22:32]

Notable Quotes & Memorable Moments

• On AI Integration:
  “What was really clear to me from speaking to everyone at Qualtrics was this was being fully embraced and lent into, but in the right way.”
  — Jordan Harper [02:58]

• On Research Evolution Metaphor:
  “It felt to me like building this big array of mirrors of customers inside organizations… AI… is like a lens to peer into that big data set and extract the useful things from it.”
  — Jordan Harper [07:02]

• On Bias in Human vs. Synthetic Responses:
  “With humans… priming created a clear pattern of bias, but with synthetic, [there was] almost no pattern… That’s really interesting.”
  — Jordan Harper [10:05]

Key Segment Timestamps

• [00:55] – Setting up the synthetic panels topic
• [02:58] – Qualtrics’ AI vision and integration
• [05:01] – Mirror versus lens metaphor in research
• [08:05] – The “honesty” of synthetic responses
• [13:38] – How synthetic research transforms researcher roles
• [16:51] – Overcoming organizational skepticism
• [19:57] – The predictive future of research
• [22:02] – Personal highlights and advice

Overall Tone

The conversation is candid, curious, optimistic, and pragmatic—recognizing both the promise of AI-driven research and the irreplaceable need for human judgment.

Summary Takeaway

Synthetic panels, powered by AI models authentically trained on human data, represent a breakthrough for marketers and researchers: streamlining the collection of actionable insights, reducing survey fatigue, diminishing bias, and enabling an agile, experimental approach to customer understanding—all while preserving the essential interpretive role of the human expert.