What does it actually take to turn AI from an experiment into measurable growth?
In this episode, I caught up with Jason Riback, President of MediaMint, to unpack what “agentic growth services” really mean in practice. MediaMint works with leading organizations across media, entertainment, retail, and technology to scale front-office operations across marketing, sales, media, and data. But this conversation was less about buzzwords and more about execution.
Jason shared his journey from engineering at the University of Michigan to McKinsey, and then into the heart of the San Francisco startup ecosystem. That blend of operational rigor and startup agility clearly shapes how he thinks about growth today. For him, AI is only valuable when it produces definable improvements in real workflows. That means fewer manual handoffs, fewer errors, faster cycle times, and better output quality. Otherwise, it is just another system sitting on the shelf.
We spent time breaking down the gap between data-driven marketing in theory and decision-making in reality. Reporting is always retrospective, Jason reminded me. The real challenge is using insights in near real time to influence spend allocation, targeting, and optimization before a campaign ends. That requires governance, clean data, and clear accountability. Without those foundations, organizations risk operational exposure and opaque decision logic that no one can confidently explain.
One of the most thoughtful parts of our discussion centered on human oversight. Jason was clear that while AI can technically retrain models and adjust guardrails on its own, handing over full autonomy creates a black box problem. Enterprises need the right governance layer, where recommended changes are reviewed and approved against clear outcomes. Automation should feel invisible within the workflow, not like another dashboard demanding attention.
We also explored MediaMint’s Intelligent Assistant platform, MIA. What stood out to me was the pragmatic approach. Rather than offering a one-size-fits-all tool, MediaMint customizes AI agents around each client’s tech stack, data connectors, and workflow steps. That flexibility is essential because no two marketing organizations operate the same way. The goal is applied agentic execution embedded into daily workloads, not theoretical AI capability.
Finally, we turned to the people side of transformation. As automation becomes embedded in front-office operations, roles will inevitably shift. Jason believes teams will move away from repetitive execution and toward managing, interpreting, and optimizing AI-driven processes. That shift demands AI literacy, cross-functional alignment between marketing, tech, and finance, and shared agreement on what good looks like. Accountability does not disappear simply because an agent executes a step.
If you are wrestling with how to apply AI inside marketing and revenue operations without creating new risks or unnecessary complexity, this episode offers a grounded perspective. It is a conversation about discipline, governance, and measurable outcomes, not hype.
What would change in your organization if automation genuinely reduced cycle times by 50 percent while improving quality and transparency at the same time?
Neil: , [00:00:00] thank you for joining me on the podcast today.
Could you tell everyone listening a little about who you are and what you do?
Jason: Sure thing. First off, Neil, great to be here. Thanks for having me. Um, my name's Jason Riback, president of MediaMint which is an AI powered growth services company. We've been in business for just over 15 years.
Started back in two thou, late 2010. And actually at the time I was one of the first, I was the first client, um, helped, started the company back in the day, really trying to focus on how do we grow and scale businesses. And I think what it's evolved into is something really great where we kind of get to sit with.
Business leaders on a day-to-day basis and think about their biggest challenges and how they accelerate their revenue and growth services side of their business. Um, in terms of my background, I born and raised in Detroit, Michigan. Went to University of Michigan, did a master's in engineering there as well.
And then started my career at McKinsey and Company. I was there for about. Six [00:01:00] years really focused on operations transformation across a wide variety of industries and, and supporting those clients globally. Then I decided to kind of make my way out to San Francisco because I had a bit of the, the startup bug and wanted to get into technology.
I jumped first into media technology and then never looked back. Started a company called. Tribal fusion that was a mid-stage startup at the time which was great because it gave visibility into the entire end-to-end business and also allowed me to have you know, real impact right away. Move from there into more of the pure place startup.
It helped start a company called Home run.com. Um, it was in the daily deal media services. Space. And as we were growing from 10,000 users to 20 million users, we had a real problem in how do we actually build a, a long-term sustainable business? And in order to do that my colleague from Tribal Fusion a bci was just trying to start [00:02:00] a company out in India.
And that's where we had the genesis for Medium. It became kind of the, the team that was able to kind of help us build and scale grow. That startup. So it, it turned out to be a great relationship. Over the years I was advising the company and then back in 2017, I came into join the Journey.
And we've since grown that company from, we were around 150 employees to now just around 3000 employees globally with eight delivery centers around the world. Expanding to, to more you know, as we speak serving over 250. Clients. And really, again, I think the, the excitement and a bit that is surreal is really seeing the growth and development of the, the team around us.
And so we've, you know, really focused on, you know, leading with you know, our core values of how do we create an environment for people to flourish, focus on, you know, pursuit of excellence for our clients and [00:03:00] continuous improvement. Um, making sure that we always have the highest level of integrity and respect above all, across those teams.
And then, you know, if we do all those things right, we can really deliver on you know, what we see as true customer success. So it's it's been a great journey so far. And I think now as we're shifting into an agent supported environment, it, it's really a. A a turning point where I think we're gonna see incredible acceleration going forward.
Neil: So much I love about your backstory and one of the things that stands out to me is that you, you go with problem first solving the problem first tech second, and too often these days I, I hear it the other way around where people go, AI or tech first. And when I was doing a little research on you, I read that you described media intelligent assistant or.
MIA as an operating model for modern marketing operations. But before we dig a little bit deeper into [00:04:00] that, what problem inside enterprise marketing were you actually aiming to solve with this framework?
Jason: So I, I, I think the, to the point you were saying, really thinking about what are the problems that we're trying to, to solve mm-hmm.
And so, you know, understanding the entire workflow from end to end, right? You wanna identify where can you leverage the right systems and tools to improve that process. And so it's not just about having a a strategy on, okay, we're gonna implement these tools, but it's at the execution of which tools to implement at what stage of the process.
And so you know, when we think about it, we wanna make sure that the the, the. Assistance that we may be developing, right? It becomes true applied ag agentic where we're actually using it to, to drive definitive and definable improvements in a process. What are those KPIs? How can we actually take [00:05:00] a system and either reduce the process times by, you know, 30, 50, 70%.
Also focus on how do we actually improve the quality of the outputs that were delivered. And so rather than just having a a black box, we really wanna make sure that, you know, there's clear clear improvements that we can measure and demonstrate that the value of the tools and technology that we're bringing are actually getting the results that you need.
Neil: A lot of marketing teams, they're currently drowning in tools, dashboards, and alert fatigue and so many other things there. So how does intelligent automation through medium intelligent assistant, how is this reducing friction instead of just merely adding another layer of complexity on top? Tell me more about what you're doing here.
Jason: Yeah, well, you know, we definitely don't wanna introduce just another dashboard. Yeah. We wanna really embed into the existing workflows. So. [00:06:00] If again, automation is reducing the steps you know, pacing, quality reconciliation, campaign setup then it's just becoming noise. So the goal is to tighten the feedback loop have fewer man manual handoffs, fewer errors, faster adjustments without losing accountability.
Automation should feel invisible in the workflow, not like another system that you have to manage.
Neil: And we often hear about data-driven marketing a big buzzword right now, and, and yet still in many organizations, many teams still struggle to move from reporting to action. So how, how are you helping organizations turn that performance data into real decisions around targeting, spend, allocation, optimization, et cetera?
Jason: Yeah. You know, look, any reporting is gonna be retrospective.
Neil: Yeah.
Jason: And how do you look at that more in real time versus at the end of a campaign and execute against that in real time? So, you know, the execution requires [00:07:00] change. So we focus on, you know, where there's operational revenue intelligence meaning the data's not just visual.
Visual, I'm sorry, visible. But structured so we can have decisions quickly. And that could mean, you know, can we reallocate spend mid-flight? Can we adjust yield levers? Can we catch discrepancies before the end of a billing cycle? So how do we take the information that we have in real time and make the proper adjustments to drive better outcomes and, and actions.
Neil: And AI everywhere is in MarTech conversations everywhere right now. But equally there is that big ROI question mark in not just marketing, but any tech project at the moment. So in your experience at Medium, where is AI genuinely accelerating marketing operations and, and where is it still overhyped? I would imagine you've, you've seen a few examples of both there, but what excites you and, and what is still overhyped would you say?[00:08:00]
Jason: I, I think definitely with the implementation of ai, there's you probably hear from lots of folks so many initiatives that, that some have found, some have worked, some have failed, right? Many have failed. So I think where works really, really well is when there's structured data repeatable workflows and.
In those, you can see meaningful efficiency and accuracy gains because, you know, you're starting with, you know, you're, where you're ending up. I think the overhype is right now on totally autonomy without context. Marketing environments are financially sensitive and cross-functional. Removing human judgment entirely is not really realistic right now.
Augmentation is where the value is. You know, I will give an example. We've had a couple of clients who've said, Hey, can now the AI. That we implement, can it start training itself on the next model, improving it and going forward? And the answer is [00:09:00] technically yes. Theoretically yes. It can, can take those insights and give you guidance on what to do, and you could set it up where it
it can make those corrections to your guard rails. Now the problem is then becomes the black box. Where is it? Making the right adjustments. This is where we think, you know, human interaction, the right governance is required to ensure that the recommended changes or the insights that you're having are approved to what you want the outcome to be in met flow.
Neil: When doing a little research on you, before you sat down with me today, I was reading how you emphasize governance, clean data, and most importantly of all, arguably is human oversight areas are very passionate about there. So what rest do enterprises face when they may be guilty of rushing into ai, into their marketing workflows without those foundations that, that you mentioned are in place?
You've probably, again, you've probably seen this happen a fair few times as well.
Jason: They're [00:10:00] definitely focused. We're definitely focused on trying to make sure that as we set these up as I mentioned, we have the right guard rails and governance in place for any of these implementations. I think. Where it starts is the data, right?
So if the data is not reliable, if you have garbage in, garbage out, you need to make sure that as you're pulling from all these different sources and we have these different connectors to, to understand what is happening across these different systems. We wanna make sure that AI can make the right decision.
So if data isn't reliable, AI is. Going to confidently make the wrong decision at scale, and that's the real risk. So we've seen pilots stall because companies treat agents like plug and play tools. Without a structured playbook, access controls and defined ownership, you don't get scalable value.
You get that operational risk.
Neil: I suspect for many people listening, they'll be aware of that gap between marketing strategy at the board level [00:11:00] and execution inside regional or channel teams. So how do you see structured automation models improving visibility and most importantly, decision velocity across those global operations?
'cause this, this feels quite a big opportunity too.
Jason: Yeah. You know, strategy breaks down when execution varies by the region or the team. A structured model standardizes the way workflows run. Not to eliminate flexibility, but to create consistency. So when automation and playbooks are aligned, leadership gets real visibility and regional teams spend less time reinventing processes and more time executing.
Neil: Of course, if we look a bit deeper at the MarTech stacks, they can have often become expensive experiments. So what investment discipline do you think enterprise leaders should be applying when maybe evaluating new AI and new automation tools to ensure that they, they get that measurable business impact and that [00:12:00] ROI that is under the microscope right now?
Jason: Yeah. You know, I think, I think you need to start with keeping it simple. Start with real operational bottlenecks tied to revenue. If you can't measure and the yield, the cycle time, the error reduction over the throughput, it may not be the right place to deploy AI yet. You also want to get your teams bought in to the outcomes that they can drive with ai.
So having those early successes really means a lot to the adoption of the agent solutions and tools that you want to put in place. So discipline means embedding automation into these live workflows, validating the results, and not just collecting a whole set of new tools off the shelf.
Neil: And we've talked a little around technology and AI and tools today, but of course at the heart of everything here, what's gonna make it a success is people and intelligent automation will change roles [00:13:00] and some responsibilities.
So how should leaders listening be thinking about upskilling teams, aligning marketing with tech and finance, and reinforcing accountability as automation. Inevitably becomes embedded in day-to-day execution From a people standpoint, what do you see here? What would you advise?
Jason: You know, I think that the roles are gonna shift from doing just the ongoing repetitive work to, to managing it.
So it's really how do you think of delegating the tasks, elevating yourself, focusing more on strategy, growth product. You know, your clients, those relationships, I think that's really where the, the value lies. And in order to get there, it really requires some AI literacy so a stronger cross-functional alignment.
So how am I engaging with the teams around me and using AI again to gimme that space, to do that marketing, tech and finance. They all have to agree [00:14:00] on what good looks like and that accountability doesn't go away. Even if an agent executes a step a human still has to own that up.
Neil: Um, we have talked a lot around the medium in intelligent assistant platform, or MIA for people that are listening and hearing about this 'cause it is a relatively new release.
What excites you about it now and where it's heading and what you're working on For anybody listening, what excites you about this and how would you sell it to somebody?
Jason: I think what's most exciting is, um, we've taken a very. Pragmatic approach. Yeah. And how we've been building it out, how we've been sharing it with clients.
And effectively we've built a foundation layer that is, is built on the 15 plus years experience we have in the industry doing all of these different functions, all these different workflows. So that was our starting point. The, what we realized is you actually need to truly customize the AI agents for each client.
[00:15:00] Everybody has different. Tech stack. So everybody has different data segmentation and and, and different data sets that they need to pull into, let's say that data lake, right? Um, so there's no one size fits all even of the platforms. So having the right data connectors, customizing certain data connectors, especially if someone has proprietary systems or tools and then addressing it against, you know, very specific.
Functions and workflows. That's where we, we find it most exciting. So each time we implement it while again, there may be some common structure and base layer, the actual implementation for a workflow, again, it could be on the media planning function, the ad operations function, campaign management and optimization function.
Each of those have different steps. We can. Build it around the client's workflow so it becomes a a truly productive tool for them. And [00:16:00] then when we can layer on the meeting in team's ability to kind of be the the execution layer as well, right. We now have, you know, we, our team now had that many to leverage the tool.
It becomes truly applied ag agentic solutions within those workflows. So from my standpoint it's not just, again, handing a client a tool and saying, put this data and here's the data that you get out, and I hope you use it the right way. It's really, your systems and tools are constantly evolving. Here's the tool that works today.
And then as we need to change the inputs, as we need to change the. The guardrails from the things that the, as we learn we can do that in real time with the cloud, right? Again, with the right governance. So I think it is we've seen a lot of success with the first set of implementations we've done because we're on the journey with the cloud and we're making sure that, [00:17:00] that it's not just another tool on the shelf.
It has to be used on a daily basis.
Neil: Love that, and I think it's a powerful moment to end on. But before I do let you go, for anybody listening, wanting to understand a little bit more about what separate, what separates AI hype from actual operational success, learn more about MediaMint and everything we discuss today.
Where do you like to point everyone listening?
Jason: Yeah. I think the best place to, to find out more about MediaMint is just come to our website MediaMint.com. You could also find me on LinkedIn. You know, we regularly share insights on how marketing, operations and intelligent automation are evolving. We're always happy to continue the conversation with you know, with anyone and, and really trying to lead the, the evolution of the space.
Neil: Awesome. Well, I will add links to both the website and your LinkedIn profile, I, everyone listening to go check that out. Learn a little bit more about what's happening here. MIA [00:18:00] is relatively new, a few months, but there's so much going on there and it's gonna continue to evolve this year, I'm sure. So check that out.
Check back in with me, let me know what your thoughts were, how you, how it might work, how you are working, et cetera. But more than anything, Jason, thank you for shining a light on this today and starting the conversation. Really appreciate it.