AI Digital Brings AI and a Platform Agnostic Approach to Brands and Agencies with Mary Gabrielyan

Summary

---

Marketecture Podcast Summary

Episode: AI Digital Brings AI and a Platform Agnostic Approach to Brands and Agencies
Host: Ari Paparo
Guest: Mary Gabrielyan, Chief Strategy Officer, AI Digital
Date: May 4, 2026

---

Episode Overview

This episode features a deep-dive interview with Mary Gabrielyan, Chief Strategy Officer at AI Digital, a pioneer in AI-driven media consultancy. Mary discusses AI Digital’s unique approach to media and supply management, the evolution and integration of artificial intelligence in programmatic advertising, and the challenges agencies and brands face on their AI transformation journeys. The conversation is candid, technical, and accessible—a valuable primer for those looking to understand AI's real impact on advertising and marketing.

---

Key Discussion Points & Insights

1. What is AI Digital?

• Founded eight years ago by Steven Magley, AI Digital was among the first to embed AI, even before it became a trend (00:38).
• Company Focus: Not just a managed service provider, but a "native media consultancy" with three main service lines:
  • Managed Service (Media planning, buying, and execution)
  • Smart Supply (Supply management and curation)
  • AI Labs (Consultancy helping clients with AI transformation and bespoke tools)
• Elevate Platform: The newly relaunched intelligence layer is designed for planning, optimization, reporting, and insight generation (01:47).

---

2. The Elevate Platform: All-in-One Intelligence Hub

• Target Customers: Mid-size agencies and brand-direct marketers in the US and Canada (03:09).
• Platform Agnosticism: Flexibility to select the best data partners, DSPs, and SSPs for clients’ needs (03:17).
• Key Modules:
  • Research: Unified access to multiple competitive and audience analysis tools, providing AI-driven insights (03:48).
  • Advanced Planning: Discovery of audience segments and inventory, with push-to-launch capabilities.
  • Reporting: Real-time campaign visibility, including pacing, MMM (marketing mix modeling), and path-to-conversion reports.
  • Unique Differentiator: Ability to ingest and analyze offline and email data for comprehensive outcome tracking (04:55).

---

3. AI in Supply Path Optimization (SPO)

• The Challenge: Industry jargon around SPO is often "abused"—most vendors offer reactive, not truly AI-driven, optimization (06:03).
• Mary’s Differentiator: Focus on predictive, real-time optimization—choosing the best supply paths before the auction occurs rather than just analyzing post-bid data (06:28).
• Encouragement to Marketers: "Please ask your supply providers what they are really doing and why they're calling it AI..." (07:28)

---

4. Optimization: From Metrics to Outcomes

• Traditional Metrics: CPM, CTR, ROAS remain the basics, but clients are increasingly seeking true "outcome optimization" (08:37).
• Shift to Value-Based Bidding: Smart clients are connecting data signals to actual business results, moving beyond metric chasing (09:03).
• Quote: “They just don’t want metric optimization, they want outcome optimization.” — Mary Gabrielyan (08:54)

---

5. The Future of Curation & Contextual Targeting

• Dynamic Curation: Moving beyond "static deals" to real-time inventory scoring and package creation, adapting to performance dynamically (11:03).
• Contextual AI: Use of LLMs and AI to understand the nuances of content and environment, enabling true brand safety and alignment (12:43).
• Quote: “LLMs can even read the video scene level...another depth of understanding where your ads are running.” — Mary Gabrielyan (12:52)

---

6. Creativity, Change Management, and Human Skills

• Client Attitudes: Range from enthusiasm to anxiety about AI’s impact on creativity and job security (13:16).
• Barriers to Adoption: Underinvestment in staff training, lack of tool proliferation, and fear of job commoditization (13:40).
• Irreplaceable Human Skills:
  • Bodily experience and intuition (15:02)
  • Visual literacy and “taste” (15:34)
  • True empathy and authenticity (15:52)
• Quote: “Authenticity is something that we shouldn’t even be delegating to AI.” — Mary Gabrielyan (16:14)

---

Notable Quotes & Memorable Moments

• On the misuse of “AI” as a label:
  “Everything now is AI powered, including my soda that I’m drinking right now.” — Mary Gabrielyan (06:28)

• On supply path optimization:
  “Most web vendors call their SPO an AI powered, but in reality [it's] just reactive supply path optimization.” — Mary Gabrielyan (06:26)

• On the human element:
  “Sometimes intuition just knows the best and I cannot delegate that in any shape or form to AI.” — Mary Gabrielyan (15:07)

• Podcast lightning round:

  • Competitive Advantage: “I think we’re agnostic and we are ready to go a mile farther, no matter how big or small the client is.” (16:35)
  • Biggest Challenge: “...to keep on scaling the company at the pace that we are scaling, because we are running out of talent.” (16:49)
  • What animal is AI Digital: “Definitely a dog...We are a huge, huge dog company. Also cat company...our company mascot is our CEO’s dog.” (17:19)

---

Timestamps for Major Segments

• 00:38 — Evolution and background of AI Digital
• 01:47 — Overview of AI Digital’s core services and the Elevate platform
• 03:09-03:48 — Who uses Elevate and a walkthrough of its features
• 05:45-07:45 — Deep dive into supply path optimization and the industry's misuse of "AI"
• 08:37-09:57 — The shift from metric to outcome optimization and value-based bidding
• 10:09-12:55 — AI’s role in curation, dynamic deals, and advanced contextual targeting
• 13:16-16:24 — Creativity, change management, and the unique value of human skills
• 16:35 — Lightning round: competitive advantage, challenges, culture

---

Conclusion

This episode offers a comprehensive and candid breakdown of how AI is genuinely transforming ad tech, not only in optimization and planning but also in reimagining human roles and skillsets. Mary Gabrielyan provides both vision and grounded realism, spotlighting the importance of outcome-driven approaches, the need for true technical sophistication, and the irreplaceable value of human intuition and creativity.

---

For more information: Marketecture.tv

Transcript

[A]

Welcome to Market where you get smart fast with in depth interviews of leading technology executives. I am joined today by Mary Gabrielian, who's the Chief Strategy officer for AI Digital. For those of you listening on audio. She just waved. So, Mary, how are you? Thanks for joining me.

[B]

Thank you for having me, Ari. Very nice to be at Architecture podcast as always.

[A]

Yeah. So tell me about AI Digital. It's not a company I'm familiar with, so I want to learn something today.

[B]

So AI Digital was established eight years ago. Our founder, Steven Magley called it AI Digital and it wasn't coincidence we were using AI when AI wasn't as cool.

[A]

Oh, wow. Eight years ago. That's impressive to have an AI name.

[B]

And so we, we, that's why we define ourselves as a native media consultancy because we have AI in our DNA and with like eight years ago it was machine learning and algorithmic optimizations and now it is LLMs Agent gi. So we have built the company and company's DNA around that and we are calling ourselves Media Consultancy because we are not only doing managed service and supply management, but also we have our AI Labs consultancy arm that helps our clients, agency and brand direct with their AI transformation.

[A]

All right, so I want to take some time to learn about what sort of projects you've been working on because you probably have some interesting stories from the front lines. Why don't we just, why don't we start though by just going through your offerings? So you said you have media consultancy supply. Why don't you just run through your key offerings and what they are?

[B]

Yeah, so AI Digital has two core services, Managed Service and Smart Supply. Smart Supply is our supply management curation arm. Basically, we launched our AI Labs last year. It has an incubator that creates repeatable tools for our clients. We have chief AI officer who conducts exact level trainings, who helps to create AI roadmaps and transformation roadmaps for our clients and companies. So those, those three are our main service lines, if I may say so. We have also launched our Elevate, which is our platform, our intelligence layer as we call it. It is helping our clients to plan, optimize, look at their reporting, look at their MMM reports, path to conversion reports. So it's like one stop shop for our clients to get access to their live campaigns in all the stages and throughout the whole life cycle of the campaign. And yeah, we have several secret endeavors that are coming.

[A]

Secret endeavors. You're on a podcast, you're not supposed to talk about secrets.

[B]

I don't want to spoil, but Q4 is going to be very loud and bright for AI digital.

[A]

So tell me about Elevate. So first of all, who are the customers? What are your typical customers?

[B]

Our typical customers are mid size media agency and brand direct companies based in US and Canada.

[A]

Okay.

[B]

And why they are choosing us? Frankly, because we are tech agnostic and we are independent. So we are choosing the best in class technology for our client, whether it's data partners, whether it's DSPs, SSPs. So we are flexible enough to accommodate the client's wants, needs and media briefs, frankly.

[A]

Yeah. So and what is give, give me some color on how they use Elevate. What do you, if you log into Elevate, what do you see? What's exciting going on?

[B]

Yeah. So Elevate was relaunched this year and now it has four very powerful model modules. The first module is on research. So you enter Elevate and you are immediately accessing several competitive analysis tools and audience analysis tools. So instead of paying, let's say 50 grand per year for just one tool that is providing you with audience intel, you have several under one roof that gives you a combined analytics on your target audience. Same with competitive details. After collecting this research, you can inform your strategy based on research. And it is of course AI powered. We are providing like AI insights and conclusions based on it. And then you can push to the next module which is advanced planning and audience segments and inventory discovery. So you choose your inventory, you choose your audience segments, you create the plan which is a tangible plan that you see and you can send it immediately to launch. After that, of course you see your report, of course you see your daily pacing, what's not. And at the end we are also providing my two favorite reports actually, which is mmm. And in our case it's not just media mix modeling, it's marketing mix modeling. Because we have the opportunity to also ingest customers offline data or email data or whatever data they want to see comprehensively what is really driving their business outcomes. Plus we have Path to Conversion report, which is also very exciting to look at. So Elevate has this end to end capturing of what's going on with your campaign.

[A]

All right, well, let's talk about the AI consultant because you're on the front lines hearing what challenges real businesses have. So let's go through some examples. So I think, do you want to start with supply as a, as a challenge?

[B]

Yes. And honestly, talking about supply and AI in this, this combo is my favorite combo because I feel like AI is the most mature decision for supply and I Think this is what we lacked for years and years in the industry.

[A]

Right. Okay.

[B]

And yeah, 100%. And because supply is such a, such a misunderstood, misunderstood topic. And where I see the most of AI empowerment happening is of course with supply path optimization. But I want to be very specific about what this means because this term is frankly being abused right now.

[A]

It is, it means many things.

[B]

Honestly, everything around AI is being abused because accidentally everything now is AI powered, including my soda that I'm drinking right now. So most web vendors call their SPO an AI powered, but in reality just reactive supply path optimization. What I mean by reactive, I mean they look at their historical data. It's. It's like you are looking at the mirror at yourself after the bid, the bidding has happened, right? When I say, for example, in Smart Supply, we're heavily investing in AI SPO right now. We are like in R and D. Because like this shouldn't be a reactive thing. This shouldn't only be based on your historical performance, but it should be predictive and real time before and like it should be a forecasting of which bitstream path will deliver the best outcome before the auction has happened, not after. And this is the tricky nuanced part. And I'm encouraging all our listeners, clients, please ask your supply providers what they are really doing and why they're calling it AI, whether or not it's AR AI or non AI.

[A]

So are you creating, you know, lists of performance websites, apps, et cetera, and then pushing them back into the dsps? Is that kind of the challenge here?

[B]

We are of course doing that, but not only that, that on top of that we are also creating like predictive modeling. Plus, no one is canceling forbidden. Yes. Yeah, no one is canceling math modeling, of course, but AI helps to fill in the gaps where we don't have enough data and to build like the future assumptions of what is really going to contribute to business outcome versus just. Okay, this is cheaper cpm.

[A]

Yeah, that makes sense. So let's talk about optimization. Everyone says they're using AI for optimization, so what are customers asking for? They just want number go up, right? Or is there some more science to this?

[B]

Everyone wants CPM optimization, of course, everyone wants their CDRs to grow, their ROAS to be better, etcetera, etcetera. But I think one thing that I'm seeing among our clients as well is the shift in what we, what we understand when we say optimization. Meaning that they just don't want metric optimization, they want outcome optimization.

[A]

Yeah.

[B]

And what I'm seeing is that we are moving towards this value based AI bidding, value based decision making in terms of whether or not this ad should be optimized or not. So it is the major ship that we are seeing and the companies that are moving towards this value based AI bidding and optimization and like custom signals connecting the data with like true outcomes, they are in long term, they are definitely seeing better results. However, there is still a fair share of clients that are stick to can we get better CPMs.

[A]

Right? Right. Sure.

[B]

Will this change ever? I don't think so. It makes me very happy that this new cohort of clients that think about value based optimization are emerging.

[A]

And once again we're moving from a model that your team is creating to execution in a different system of some kind. So what's going with adcp? Have you been investing in that area?

[B]

Yes. And like I think ad curation is where programmatic is heading and AI is the engine to make it possible at scale. Because as we know, the scale is one of the trickiest parts with curation. Again, this is pretty nuanced and it has its own dark size that are not quite visible to the clients.

[A]

Right.

[B]

For example, the curation packages that are happening all over the place right now, sometimes it's just marketing label like you create a static deal and it's just marketed as something real time, et cetera, et cetera, et cetera. However, what I'm seeing and for example with Smart Supply and some other industry players are also doing this and investing in this, this makes me very happy. Is like true dynamic curation.

[A]

Right.

[B]

By dynamic curation I mean the deal exists, the trader is running it and in real time the inventory is getting rated and the inventory that is performing stays. The inventory that is not performing is being pushed out and replaced with predicted better performing one.

[A]

Yeah.

[B]

So that is where I think that inventory scoring plus creating packages that are truly changing in real time and adapting is where the ad creation AI future lies.

[A]

Right, Right. That's super interesting.

[B]

Can I add one more thing I just remembered because this is literally what I was discussing with our R and D team this morning. It's about contextual AI and the notorious contextual targeting that has been around for ages and ages is defining actually the whole industry. To be frank, LLMs are very good at understanding the semantics of the environments. I think this allows folks like us who are engaged in supply management to provide the better suited and the most relevant, like the most relevant publishers and the most relevant brand safe environments for our clients. Because you are not just relying on I don't know, block lists or IAB categories. You are truly relying on context, on the semantics and on the connotation because right now LLMs can even read the video scene level. So. And it gives you another depth of understanding where your ads, the environments where your ads are running.

[A]

Yeah, you just have to scale it to millions of websites which is, you know, no problem, right?

[B]

Yeah, no problem.

[A]

And then I guess last, you know, I think we had talked beforehand about, you know, creativity and how clients are sort of embracing or scared of AI as it relates to that. What are you seeing when you talk to customers about that whole world?

[B]

Like when I talk to our clients, when we, when we start talking about AI, it's either silence because they're a bit worried and they don't know and they cannot keep up or it's a complete enthusiast because people, people are adopting it as they go. So it is like two polarized reactions that I'm seeing.

[A]

Yep.

[B]

Probably my answer will be twofold. So first what I see is that companies are under investing in their people training and people adoption. For example, they just buy one tool subscription when there are 15 tools that a modern day marketer or media person should have.

[A]

Right.

[B]

But they expect outstanding results from AI. So it is that. Second is that people are really worried about their jobs and how their work day to day stuff will get commoditized. And, and I feel like these two things coexist and they are interdependent. I feel like companies need to invest more in people training in tool adoption and people need to understand that the skill set is being redefined. A big change. There are some things that are still better human, but it is the agility and the flexibility and the adaptiveness that is going to make or break a person. A company.

[A]

Sure.

[B]

And a workforce.

[A]

So what are, what are some examples of things that are much better human?

[B]

Oh, recently I published an article about like some skills that are never going to be okay.

[A]

I want to hear them. I need to, I need to get those skills.

[B]

So I feel like the first, I'm sure you have all of them and even more. But I feel like first and foremost is like bodily experience. For example, AI would never feel the pain, empathy, and will never have the intuition. I'm hiring lots and lots of people every, not every day, but every month. And sometimes intuition just knows the best and I cannot delegate that in any shape or form to AI or another thing. It can be the visual literacy or as it's called, taste. Yeah, you just need to have the taste.

[A]

You need to have taste.

[B]

Yes. To understand whether or not this creative is working, whether or not this document is correct or the storytelling style is good or bad. Another things can be like empathy, like. And true. Like true, true empathy. Not the art. Because artificial.

[A]

They could fake all this stuff.

[B]

Yes, you can fake it, but it's really surprising. But people always know when it's AI. You've probably come up to like LinkedIn post. You read it and you immediately know whether or not it's AI. And I feel like that is something that cannot be replaced and we intuitively understand when it's AI. So authenticity is something that we shouldn't even delegating to AI. And there are a couple more. So I'm not you with that.

[A]

All right, let's. Let's do a quick lightning round. So I'll ask you a couple quick questions. You asked me relatively quick answers. So what is your number one competitive advantage?

[B]

I think we're agnostic and we are ready to go a mile farther, no matter how big or small the client is.

[A]

All right, what is your biggest challenge?

[B]

The biggest challenge is to keep up, keep on scaling the company at the pace that we are scaling because we are running out of talent.

[A]

So if you see Mariette at Possible, make sure to drop her your resume. I think this is actually getting released after Possible, so we'll see. And last, if AI Digital was an animal, what animal would it be?

[B]

Definitely a dog. We are a huge dog company. We have dog clubs and all of that based on breeds, etc. In like, we are a huge, huge dog company. Also cat company where we say we are very pet friendly. And yeah, our company mascot is our CEO's dog. So.

[A]

Okay. Do you have a dog?

[B]

Of course. I have a Vizsla who is a crazy but a good boy.

[A]

All right, Good boy. All right. Well, thank you so much. Mary Gabrielan, the chief strategy officer of AI Digital. Thanks for joining us today.

[B]

Thank you so much, Ari, for having me.