Transformers Podcast - The Business Case for Multimodal AI and Automation
Introduction
In the latest episode of IBM’s Think Transformers Podcast, host Ann Funai, CIO and VP Business Platform Transformation, IBM, sits down with Mark Polyak, Chief Product and Technology Officer at MINT.ai, to explore how multimodal AI and intelligent automation are redefining the future of work, marketing, and human innovation. Below, we unpack each chapter of this compelling conversation, from the foundations of trustworthy data to the rise of agentic AI.
All Chapters
00:00 – From conflict zones to campaigns: Using data to get closer to truth
04:52 – Why data integrity matters: Guardrails, transparency and trust in AI systems
06:35 – Beyond descriptive: Unlocking prescriptive analytics for smarter business decisions
12:34 – Adapt or stagnate: The rise of multimodal AI and intelligent automation
15:33 – AI and the workforce: How technology is reshaping labor trends
21:20 – 3 ways AI is changing marketing: Automation, creativity and ROI
25:00 – Most AI solutions are failing: The role of trusted partners in driving real results
28:55 – The “glue” role: Why business translators are essential to success with AI
31:44 – Responsible innovation in action: Building multiagentic AI with guardrails
36:20 – Spotting AI hype: How to tell real value from empty promises
39:55 – Team of Teams: Leadership lessons from high-security environments
From Conflict Zones to Campaigns: Using Data to Get Closer to Truth
Mark Polyak’s path into marketing technology wasn’t linear. Before joining MINT.ai, he spent years in the world of counterterrorism — integrating satellite images, drone footage, and communication data to uncover insights in high-stakes environments.
He describes that time as an early lesson in how truth depends on perspective. “The higher the risk of a mistake,” he says, “the more you want transparency and auditability.” That philosophy followed him into the corporate world. Today, instead of analyzing conflict zones, he helps brands navigate the noisy, data-saturated world of modern marketing. But the goal remains the same: to get as close to truth as possible by bringing multiple signals together.
Why Data Integrity Matters: Guardrails, Transparency and Trust in AI Systems
If there’s one thing Mark insists on, it’s that data quality always comes first. “If the data is wrong,” he says, “it doesn’t matter what kind of AI you apply on top.” At MINT.ai, this means building systems where privacy and precision go hand in hand. The team works within GDPR and CCPA frameworks, uses anonymized datasets, and often creates synthetic data to enhance granularity. The outcome is more than technical, it’s ethical. Reliable data builds reliable systems. And in AI, trust is everything.
Beyond Descriptive: Unlocking Prescriptive Analytics for Smarter Business Decisions
Analytics used to tell us what happened. Now it can tell us what to do next. Mark calls this the shift to prescriptive analytics — where systems not only describe and predict, but also advise. It’s a vision that extends beyond business dashboards. Imagine your smartwatch or fitness app not just tracking your performance but understanding how you respond, adapting your training or recovery plan in real time. We’re entering an era, he says, where the data we create daily becomes our most personal feedback loop — one that could improve not just efficiency, but wellbeing.
Adapt or Stagnate: The Rise of Multimodal AI and Intelligent Automation
“Adapt or stagnate,” Mark says simply. That’s the new reality of work. Multimodal AI — systems that can process text, visuals, audio, and sensory data together — is already reshaping industries from logistics to manufacturing. Robots that can see, hear, and respond will handle complex tasks once limited to humans. He points to examples like Amazon, where automation could soon account for more than 70% of operations. It’s not about replacing people, but redefining how human intelligence and machine capability work side by side.
AI and the Workforce: How Technology Is Reshaping Labor Trends
Mark references a pair of studies that capture this transformation. The first, from MIT, shows that as automation increases, tasks requiring abstract thinking — creativity, systems logic, strategic planning — rise by nearly 70%. Routine and manual work, by contrast, declines sharply. The second, from Microsoft, found that mid-level and senior roles are growing even as entry-level positions shrink. AI, he explains, is empowering experienced workers to do more with less, but it’s also reshaping what “entry-level” even means. The result is a workforce that looks less like a pyramid and more like a diamond — broader in the middle, focused on expertise and adaptability.
3 Ways AI Is Changing Marketing: Automation, Creativity and ROI
In marketing, Mark sees AI’s impact everywhere:
- First, in automation — where repetitive workflows are being streamlined, freeing up teams to focus on strategy and storytelling.
- Second, in analytics — where real-time data helps brands make smarter budget and channel decisions.
- And third, in creativity — where generative tools make it possible for small teams to produce high-quality content faster than ever.
“It’s not about machines replacing creative people,” he says. “It’s about giving creative people more space to think.”
Most AI Solutions Are Failing: The Role of Trusted Partners in Driving Real Results
Despite the promise, not all AI projects succeed. In fact, most don’t. Mark cites a study showing that only 5% of AI initiatives deliver measurable ROI. The reasons are familiar — messy data, outdated workflows, and unrealistic expectations. But there’s a solution: trusted partnerships. That’s why MINT.ai joined forces with IBM. By combining IBM’s strength in data engineering with MINT’s deep understanding of media operations, the partnership aims to deliver not experiments, but outcomes. “We bring the best of data science, engineering, and media knowledge together,” Mark explains.
The “Glue” Role: Why Business Translators Are Essential to Success with AI
Every AI project involves people who think differently — data scientists, engineers, strategists, creatives. The real challenge is helping them understand each other. Mark calls this the “glue role”: the people who translate between disciplines and make collaboration possible. They define what success looks like, align teams around measurable outcomes, and bridge technical and business goals. “Without translators,” he says, “you get great models that solve the wrong problems.” It’s a reminder that even in a machine-driven age, communication is still the most human advantage.
Responsible Innovation in Action: Building Multiagentic AI with Guardrails
At MINT.ai, innovation always comes with boundaries. Mark describes the company’s multiagentic AI system — a network of specialized agents managed by a supervisory layer that ensures every decision aligns with the original goal and can be audited afterward.
Every step, every adjustment, every outcome is traceable. In industries handling millions in advertising spend, that transparency isn’t optional — it’s essential.
For Mark, responsible AI means building systems that are smart, explainable, and safe. Not just efficient, but accountable.
Spotting AI Hype: How to Tell Real Value from Empty Promises
When Mark hears someone claim that AI can “replace humans,” he tunes out. Real progress, he says, happens when people define realistic expectations — understanding what AI can do well and what still requires human creativity and judgment. The best systems know their limits. “Find the use cases where 70–90% accuracy makes a difference,” he advises. “That’s where AI can really deliver value.” Hype fades fast, but practical innovation endures.
Team of Teams: Leadership Lessons from High-Security Environments
Mark ends with a story about leadership. Drawing from his time in the military and intelligence world, he shares the philosophy of General Stanley McChrystal’s Team of Teams: hire the right people, give them clarity, and then get out of their way. Trust and autonomy, he says, are the most powerful forces in both high-security environments and high-tech industries. “If you trust the right people for the job,” he adds, “you don’t need to micromanage. You just need to empower.”
Listen to the Full Episode
The full conversation with Ann Funai and Mark Polyak is also available on your favorite platform:
The opinions expressed in this podcast are solely those of the participants and do not necessarily reflect the views of IBM or any other organization or entity.
.png)
%20MINT_622%20Advanced%20Deep%20Dive%20(blog).png)
