Building AI-Ready Customer Data with Tealium CEO Jeff Lunsford

Artificial intelligence is only as good as the data behind it. In this episode, we sit down with Jeff Lunsford, CEO of Tealium, to discuss why customer data has become one of the most strategic assets for enterprises embracing AI.

As organizations race to deploy AI applications, digital assistants, predictive analytics, and agentic workflows, many discover that fragmented, outdated, or poorly governed customer data becomes the biggest obstacle—not the AI model itself. Jeff shares how enterprises can move beyond traditional Customer Data Platforms (CDPs) to create real-time customer intelligence that powers meaningful AI outcomes.

During our conversation, we explored how the customer data landscape has evolved from the early days of tag management into today's world of real-time data orchestration, AI activation, and predictive decisioning. Jeff explains where Tealium fits within the modern enterprise architecture alongside data warehouses, cloud platforms, reverse ETL, and customer engagement systems.

We also discuss the importance of creating real-time customer context, enabling AI systems to make faster, more intelligent decisions while maintaining strong governance, privacy, consent management, and regulatory compliance. Jeff provides a practical overview of AIStream and explains how organizations can deliver AI-ready data to applications, models, and autonomous agents in real time.

The conversation also explores:

• Why data quality—not AI models—is often the biggest barrier to successful AI deployments
• The role of real-time customer context in improving personalization and customer experiences
• Predictive intelligence and AI-driven decisioning
• AI at the edge and real-time activation
• Building trusted AI through strong governance, privacy, and consent management
• Partner ecosystems spanning cloud providers, data platforms, and AI technologies
• Emerging trends including Model Context Protocol (MCP) and agentic AI workflows
• Practical advice for CIOs, CMOs, CDOs, and CEOs preparing their organizations for the next generation of AI

Jeff also shares career advice for students entering the workforce, discussing the skills that will remain valuable as AI continues to reshape nearly every industry.

Whether you're leading AI strategy, modernizing your customer data architecture, or simply trying to understand how AI creates business value beyond the model itself, this episode offers practical insights into one of the most important foundations of enterprise AI: trusted, real-time customer data.

Topics Covered

• Tealium overview and enterprise strategy
• Customer Data Platforms (CDPs)
• Real-time customer data and context
• Data orchestration and activation
• AI readiness
• AIStream
• Predictive intelligence
• AI decisioning
• Customer experience personalization
• Privacy, consent, and governance
• Data quality for AI
• Agentic AI and MCP
• Enterprise AI strategy
• AI careers and future workforce

If you enjoyed this episode, be sure to subscribe to The Macro AI Podcast, leave a review, and share it with colleagues interested in AI, enterprise architecture, customer data, and digital transformation.

About your AI Guides

Gary Sloper

https://www.linkedin.com/in/gsloper/

Scott Bryan

https://www.linkedin.com/in/scottjbryan/

Macro AI Website: 

https://www.macroaipodcast.com/

Macro AI LinkedIn Page:  

https://www.linkedin.com/company/macro-ai-podcast/

Gary's Free AI Readiness Assessment:

https://macronetservices.com/events/the-comprehensive-guide-to-ai-readiness

Scott's Content & Blog

https://www.macronomics.ai/blog

Transcript

00:00
Welcome to the Macro AI Podcast,  where your expert guides Gary Sloper and Scott Bryan navigate the ever-evolving world of artificial intelligence.  Step into the future with us  as we uncover how AI is revolutionizing the global business landscape  from nimble startups to Fortune 500 giants.  Whether you're a seasoned executive,  an ambitious entrepreneur,

00:27
or simply eager to harness AI's potential,  we've got you covered.  Expect actionable insights,  conversations with industry trailblazers  and service providers,  and proven strategies to keep you ahead in a world being shaped rapidly by innovation.  Gary and Scott are here to decode the complexities of AI  and to bring forward ideas that can transform cutting-edge technology  into real-world business success.

00:57
So join us,  let's explore, learn  and lead together.  Welcome back to the Macro AI podcast. I'm Gary Sloper. My co-host Scott Bryan is off today  working with a client. So you're left with me  here on the show and I'll do my best to fill his large shoes  today for this episode. We're very excited to welcome Jeff Lunsford, the CEO of Telium. Telium is a company that sits right in the middle of one of the biggest issues in enterprise AI, customer data. We've talked about data quite a bit on this show.

01:26
clean data, not clean data.  And a lot of companies are experimenting with artificial intelligence and you may be one of them. But the real question is whether they have the trusted real time governed data needed to make AI useful in actual business workflows. That is where Telium comes in. They help companies collect, unify, govern and activate customer data across marketing, customer experience, analytics and AI systems. So today we're going to

01:56
talk with Jeff about what Tealium does, why customer data has been so important in the AI era and how business leaders should really think about preparing their organization for more real-time AI-driven customer experiences. So with that, Jeff, welcome to the show. Thank you, Gary. Great to be here. I really appreciate it. We know you're a busy guy. as I kind of just did here on the intro, I think for our listeners, Jeff, who

02:25
who may not know Telium well, could you maybe start with the CEO  level view of the company, what you're doing today,  who you serve and why it's especially relevant as enterprises move deeper into artificial intelligence? Sure, absolutely. So the history of Telium is the founders and I worked together at a company called Website Story, which was the world's first real-time web analytics platform. And back in the early  aughts,

02:55
people were processing batch log files to do day old analytics. And we invented the way to do real time pixel JavaScript combo to give you real time instrumentation in the cockpit of your business. And then roll the tape forward to 2011, we saw the same need emerging with customer 360 data. So as software was eating the world,  our thesis was there are going to be more and more independent software packages that contribute

03:24
part of the customer experience across all channels. And that each of those software packages would have a slice of the customer experience, but not all of it. So someone needed to build a neutral platform for the industry to collect in real time, all of the data from all those different software packages, resolve them around identity, track things like consent and privacy compliance, and enrich those identities based on that.

03:51
those behaviors and then pump them back out to all the very same systems, all of which could benefit from that customer 360 view. So that was 15 years ago.  And today, Telium is a global company. work with 10 of the 20 largest banks in the world, six of the 10 largest telcos, three of the five largest airlines, all these big companies ah who have this, the very same  customer data problem  and opportunity, irrespective of industry.

04:18
It doesn't matter if you're retail, pharma, healthcare provider, travel, uh you name it, financial services, everyone has this uh issue, this data fragmentation issue, but it's also a great opportunity, right? Because uh now with AI taking over the room  and taking all the budgets, most of the enterprise priorities, uh everyone wants to... uh

04:46
Obviously use AI to optimize their business. Most boards of directors are challenging their management teams to do two things with AI.  One, grow the company more efficiently, find efficiencies using AI, uh optimize your processes, streamline your processes,  improve per head productivity and all that fun stuff.  In between the lines, that means  manage out some jobs basically. And then, um

05:12
Number two is figure out how to influence the top line with AI, right? Grow faster.  And that  all involves customer facing AI solutions. And if you're going to build a customer facing AI solution, great AI needs great data. Great customer facing AI needs great customer data. So,  Telium has over the last  three years  really found ourselves  working with all of our customers.

05:41
in taking this real-time trusted  data uh set that we have about their customers. And it's only first-party data. Telium, think of us as almost like software you would run on-prem, right? We're not a  network or a marketplace or anything like that. So a big bank's  data that they use Telium software to resolve and  enrich is their own data for their use. And so that data

06:08
gets used by AI powered fraud models, right? AI powered content recommendation engines, AI powered pricing models, AI fueled audiences. And then when you then send audiences to the typical digital media advertising endpoints like a Google or a Met or a TikTok or Pinterest or Snap, guess what's happening behind their API, more AI, right?

06:39
So, Telium is a software company, cloud-based software company, works with biggest companies in world, helps them collect that customer data, their first-party data. And what we preach is to companies, your first-party data is your most valuable strategic asset. capturing the value of that and then harnessing it to the power of AI is  where Telium is here to help. It's interesting you mention that because we've stressed data, clean data often.

07:08
on the show and we come across a lot of clients and I'm sure you, you know, I think you're alluding to this where they may be an established organization. We have five, 10 years worth of data. We should be great. And, and when individuals or even us, you know, separately myself or Scott are working with them, I'm going to find out that the data is not, it's very fragmented. It's in places that are not accessible and there's a lot of other issues there. So I think that's really

07:35
kind of the interesting point because  what you were saying where there's this board, uh could be pressure at times to make some changes. Sometimes they don't understand that, we'd like to move fast, but we're only as fast right now as our clean data to be able to make this work.  Absolutely. So what we typically find is, cause we kind of grew up in MarTech, you know, is our very first product.

08:03
in 2011 was a tag management solution for organizing website data. And then we  built further back in the enterprise and across all the delivery channels. But so we typically grew up in Martech inside a big enterprise. And when by the time we meet the AI teams, they typically stood up a project  and  they have built some kind of agentic offering.

08:27
And they weren't aware that Telium existed over here with this real-time data profile. And so they went back into the data warehouse and they've got these batch data pipelines streaming 24-hour stale data that, you know, is sometimes  has errors embedded in it anyways.  And then when they plug into the real-time source, it's like, holy cow, our results just got 30 % better. Well,  you know,  that's intuitively obvious to...

08:54
almost anyone, right? And we've now learned across the MarTech, AdTech, and CRM ecosystem, the real-time data is highly,  much, much more predictive, much higher signal ah than even 24-hour  old data, much less  three-year-old data or five-year-old data. But we don't replace that historical data, we complement it. And I think you mentioned, Gary, you had some history in the CDN business. uh

09:23
You know, we think of what we do is a high performance cache that sits  above the  cloud data warehouse or in-house data warehouse, but below all the customer experience systems.  And when you're collecting data and enriching profiles and creating experiences and doing all that in real time, you you need this sort of high performance cache because you can't wait for the query from the CDW to come back and give you this data. And then you have to run a workload.

09:53
So what we do is we keep these profiles persistent and basically a hot cache that you can hit. And so now it's not only outbound agentic offerings. almost every company in the world now has 10 to 20 % of their inbound traffic is agents. I mean, even pre-AI, 50 % of inbound traffic was bots, right? And we've been filtering that out. But now these agents,

10:20
That's good traffic. We don't want to filter them out because it may be  Susie's agent on her behalf coming to book  a ticket on my airline.  so  being established at the data layer the way Tullium is, we have the ability to  speak to that inbound agent in JSON-LD as an example, which is the language  agents like  to talk  and um basically expose data in real time when they hit it.

10:49
And we have an API where you could get a real customer 360 profile,  know, kind of sub a hundred milliseconds out of that hot cache. So now you've given the agent the most, the freshest context possible. And whether it's an outbound agent, offering or an inbound agent shopper or agent browser, you know, ah the action is happening right there at the data layer.  So it's just a really cool time. And there's, you know, the new protocols and MCP, ADA, A2P, all that kind of stuff.

11:18
Um, at the end of the day, it's all about great structured, filtered privacy compliant data. Yeah. And that's, that's really good to hear because you kind of answered a question before I even thought about asking it, which is we, you know, we know a lot of executives here, terms like CDP and data warehousing and reverse ETL, tag management, data orchestration, and hearing how. Thelium fits in there and, and, and the gap that it can.

11:46
probably fill for a lot of these organizations is really cool.  Speaking of gaps, where do you see the  biggest customer data readiness gap or gaps when enterprises try to AI into production? Yeah,  great question. um

12:08
You know, I think it's  what I would say is time. The biggest gap is time because I will go in and talk to a customer and they'll be bringing in 30 day old data  into,  you know, some kind of a model for some purpose.  And

12:28
To them, that's the data that was offered to them when they said, where do we store our customer profiles? In a large bank, have what in the old days we call it the CIF, the customer information file. so to them, that's okay. They don't realize that that data warehouse gets fed with three other batch pipelines that run weekly. Yeah. biggest gap is really time.

12:58
And then the other thing is, so we just fundamentally believe and we believe for 15 years, Atelium, and even before that, a website story and at Limelight that, you know, real time  is the most elegant architecture. It's the most efficient architecture  because data at rest is expensive. And then when you have data at rest and you want to bring two data at rest,

13:22
two databases at rest together into a join, guess what that is? That's a workload. Guess what that costs? Money.  And so, so what we do  is we collect data, all the event data in real time. We resolve it around a visitor profile in real time. We enrich it in real time, and then we'll send it to all the other systems of action uh and the systems of storage and analysis uh in real time. And so,

13:50
rather than a cascading series of, you know, I call it a batch backward looking hairball. That's a technical term. Rather than having a cascade. remember that one. Yeah. Rather than having that, you've got this elegant real-time architecture and it's how all computers are designed by the way. You know, it's not like a new idea. CDNs have caches for a reason, right? Right. Right. And so this is just really the idea of customer data combined with the concept in a CDN and streaming data.

14:18
applied to  all customer experiences plus AI. Does that make sense? Yeah, it makes absolute sense. So do you think we'll see more organizations  look to re-architect how they're storing, capturing their data and delivering it?  Is this a, maybe a good result  of the whole AI movement where organizations can pause and re-architect? Like, do we think that would happen?

14:46
Yes, we're seeing a wave of it  and it's,  but slash however, in the enterprise, you know, almost nothing ever goes away. Things just get layered on top. Yeah, just the tech that just piles. Yeah. So we decided long ago, hey, we're not going to go in and try to replace a data warehouse. We're also not going to replace a marketing cloud. We're going to be this neutral company that sits in the middle and integrate connects it all. uh

15:14
And that's been our, we're the only company in the industry that I'm aware of that's really stayed neutral. Like we don't have email, we don't have mobile, like we don't do all the experience stuff. There's 30 vendors upstacked that can do that and all the ad tech vendors and we're not a CDW. So,  you know, the, the new idea of a customer uh data lake, ah I think is, has taken hold and you've obviously got Snowflake and Databricks ascending plus then uh the Amazon.

15:43
you know, and GCP and Azure offerings. And I think almost all big companies are going through some effort to create a consolidated customer data lake, and then some kind of way to deal with the real time data, which of course we would believe we would hope the telium would be this for everybody, but it's a competitive market. so, but you know, architecturally, this should be, you should have this real time capability and

16:13
effectively, it's like a  cache for data collection, processing enrichment and consent status known, right? You got to capture consent and propagate it down stack and out to all the systems. And so  what we saw starting about nine months ago was a wave of new RFPs coming out. And most of them were titled CDPRFP. But then when you

16:36
kind of got into the rationale for why they were doing an RFP. It was because they're trying to do uh customer facing AI and they need to get their data act in gear. they've, and a CDP is to them a way that the industry has now trained them that they could get their customer data well organized. Does that make sense? Yeah, it makes complete sense. And, you know, I, and the reason why I was asking that question is I feel like we're seeing this across  the different uh aspects of,  of, you know, whether it's the app.

17:06
tier, now database tier, talking about data. We're seeing that on the physical layer. Uh, because to your point where, where you're, where you're kind of re-architecting that data layer. Now, if you're in that, Hey, I'm not going in completely GPU as a service. I'm going to have hybrid and kind like we were talking about earlier, the latency that needs to, you know, be removed  to talk to these environments has to be like really negligible.

17:32
And a lot of organizations I'm seeing, they're doing all these right things and they're, you know, the industry is so focused on electricity and the data center costs, but what they're forgetting is that layer one through three connection ah is  these systems are being built on what they needed for data transfer, you know, in 2020  terms. So I  think we'll almost see like kind of like a bottoms up approach to re-architect everything, which

18:01
I don't know if all of the C-suites out there or more importantly, the investors and the board members understand the amount of um capital may be required to do a lot of this.  It'll be interesting.  It will be interesting. I think what you just described is, you know, we would talk about the workloads are actually going to move to the edge.

18:26
And, you know, right now they're kind of centralized workloads and they're expensive. You know, when people get those,  uh, those invoices, whether it's a token invoice from, you know, Anthropic or a workload invoice, compute invoice from your favorite, pick your favorite CDW like, Whoa, that was expensive.  so that you're starting to see open source models and put aside the Chinese risk for a minute, but open source models.

18:55
that can run locally on a server bought from Best Buy that can give you 90 % of as good of an outcome as a state of the art model out on the Pareto curve is the geek like to talk about, but 90 % of today's Pareto curve model is better than 90 days ago's Pareto curve model, right? So if you know.

19:20
We're all getting really spoiled. so if you just think about most business applications, you know, I'm not going to, I'll tell you a teal at a tealium story. had our cyber security team using models and. know, uh, doing a whole bunch of cool stuff. Then we got a really big token bill, um, you know, five digits and, know, for the company, our size, like, wait, what happened there? You know, it was all well intentioned. just, just kind of.

19:49
popped up and that's happening. And you you heard Uber, you know,  after two months blew their annual budget, so on and so forth. So, you know, I think that everyone rational is going to rationalize this is huge value, but also, yeah, we all operate, you know, disciplined businesses. And so you will see processing get pushed to the edge and maybe it'll be the llama open source.

20:14
or the Gemma open source, you know, because it's American or whatever, but you you're going to see that stuff and it may not even run locally. I may just ask my team, Hey, go put it, stand it up on AWS, but that way, you know, we don't have to pay the API fees of the state of the art,  uh, you know, from open AI or anthropic or whoever, you know, ever working with.  um,  it's going to be interesting. I mean, I've never seen things, I've never seen a cost curve go down like

20:43
Performance goes up 10 X, cost curve goes down 10 X. It's like a hundred X change  in value per year that we're seeing right now. It's crazy. Yeah. And I can only imagine if I  put my CFO hat on trying to keep up with that change, as you're mentioning, you know, with those costs just  fluctuate so much. um So  talk to me a little bit about, uh you know, the Telium AI stream  in simple terms, uh you know,

21:13
at least for our listeners and what problem it solves, because I think there's probably some listeners out there that would really like to understand that a little bit more. And I found it very interesting just in my own research beforehand. So maybe you could talk a little bit about that and the use case that you're also seeing around, you know, using it. Sure. So, you know, for AI builders that are building an app or a website, you know,

21:43
that is an experience in an app, they need to have a data layer and good customer data. And so we wanted to solve that problem for them rather than have them go knock on the data warehouse doors and get that 24 to seven day stale data. And so we just said, let's use this exact same technology platform that we've been using for marketers.

22:09
and customer experience professionals and media professionals and let's  let AI builders grab a hold of it.  And so we've actually stood up like an MCP server where, you know, Claude can use Telium to build a data layer and do all this cool stuff. so we're like, I mean, almost every software company now is publishing an MCP server. It's kind of table stakes, but you know, the idea is that we want the native AI builders inside the enterprise. uh

22:37
to be able to find telium as a safe way to collect customer data and consent and then create enriched profiles and then do whatever they want in their AI application, right? So like I said earlier, this could be fraud, this could be pricing, this could be entertainment, it could be anything, right? It could be shopping. we want that data to be real time. And so...

23:04
If you think about our retail AI builder, someone's building uh some kind of agentic offering for one of our large retail customers. Well, you want to know what's happened on the mobile device, what's happening in the store, what's happening on the website, like before someone walks in the store or when they walk in the store, you want that experience to be have the context of what they did on their mobile app or even on their PC at home before they came, got in the car. So that's the value of real time. And we just want those. uh

23:33
AI builders inside the enterprise. Like we're not building Thelium for Vibe coders, you know, that's not our business. We sell the enterprise  and we want them to be able to find Thelium in their natural workflow. So let's say a Lang chain builder, you know, we want them to find us in the Lang chain marketplace as an example. We're not there yet, but we're  working, you know, working our way there. are, and we've got an app  in the Snowflake marketplace. We've got apps in Amazon. uh

24:02
So, so forth. that's  the concept of AI stream is basically just our, all of our products were event stream, audience stream, AI streams, concept of real time data. That's interesting. And so, so, um, are you seeing how companies are using predictive intelligence? So AI decisioning and real time activation to improve that customer experience. So think of things like retention, um, marketing efficiency, you know, upsell, cross sell.

24:32
in that service. And if so,  I mean, that would probably also tap into more of that enterprise focus, right? Because especially with AI, you have more companies that are trying to make sure that not only keep the customers that we have, but also continue to cross-sell  within that base. Yeah. Yeah. And we, you know, we talk about customer data and customer data platforms, but the great thing about Telium is, is we help you capture and build profiles for your unknown visitors as well as your known visitors.  Oh, okay.

25:02
uh And so we help you with customer retention. And that's one of the,  when you go back to like that bank example of their CIF, their CIF only has their customers. But hey, if I'm running a home equity campaign or credit card campaign or an SBA lending campaign, you know, I want to target a whole bunch of people, right? So that's another reason why you want this, this cash out here that isn't your CIF. And so um

25:30
The short answer is yes, we're seeing crazy good results.  I would say on the low end, like, and this is three years old now, but when Meta, Google, and all these guys started rolling out their CAPIs, their conversion APIs, we were the launch partners for all those. We were early  adopters, which allowed our customers to be early adopters. And on average, they see a 26 % improvement in return on ad spend.

25:56
Why? Because we can give them a bigger data surface area and then the AI models within the walled gardens of those large media properties can have better data, better targeting, better results. So like the 26 % improvement ROAS is kind of the low end. On the high end, we're like 50 % improvement in conversion rates for a global telco when we started allowing them to do first page personalization.

26:26
because when a customer landed on the website, we have this,  an API called Context API. It used to be called Moments API.  Context API allows you to query your  telium  visitor database in real time, get all that context, and then build a personalized experience. So if you have like, let's say a new account application  or a phone purchase application.

26:52
and you personalize that for that person. It's like I'm a US Navy veteran, so you might give me Navy branding, know, things like that. And

27:01
50 % improvements in conversion rate there. And then, you know, we've seen 300 and 400 % improvements in conversion rate and spoken of publicly at our customer conferences, right? So, yeah, now that AI can crawl content everywhere, like I encourage anyone to just go out go, tell me the coolest Helium published case studies and what the highest ROIs are. And, you know, you'll get some good results. Because most of this stuff is...

27:29
public knowledge and you know, we've, put their presentations out there with their permission. A lot of it, you had to attend the conference to see, but  we have uh conferences in Europe, Asia, and the U S every year. And the content is always only like, you know, 25 % tealium, 25%, you know,  industry experts like yourselves, and at least 50 % is customer content, you know, and they're up there talking about how they use tealium plus

27:59
other technology vendors plus other service vendors in the room to do these crazy, crazy good initiatives that had great ROI. No, that's great to hear. And I love hearing how companies like yourself really value that voice of the customer because we all know we've been in this industry a long time and there's a lot of products and there's a lot of

28:24
focus on trying to acquire new customers. So if you can be more mine, mine actally focused on delivering specific material, like you just mentioned the, the telco uh example. mean, that's a very hard market to acquire new customers. Yeah.  And  to be able to, capture those eyeballs very quickly is, is hard to do. Now you add everything today where SEO has completely gone down and people are going into

28:54
public LLMs like you mentioned before to  ask it questions, whether  what company does this or  what's a great chili recipe, whatever it might be.  A lot of that SEO traffic, and I'm sure your marketing team has seen this as well,  it's changed in the fact that you're  sitting on the  forefront of that data  to help customers make those changes or advances and

29:21
be able to uniquely get to  the right eyeballs ah definitely sets you apart, especially from an enterprise level, right? So easy for a lot of these startups at times to acquire new customers because they've got that one little widget, but to get into the global 1,000, ah it can be a tough challenge because everybody's trying to get their mind share ah and being able to do it with your data is very intriguing.

29:47
Yeah, 100%. So we use Profound. There's a nice commercial for Profound to  our LLM rankings. so like Telium ranks for CDP companies, which is probably 150. We're number one  for AI. We're number one for customer data platform. We're number one for uh data layer. Number one for data orchestration. Number one for data security and compliance.  So we track all these terms that are

30:15
number one for integration, number one for mobile data operations, and number one for tag management. And this is ahead of like, Google's number two for tag management. So even though they have 30 main websites and we have, I don't know, 20,000 across say 700 global companies, our brand recognition when you ask the LLMs is higher just because of 15 years of hard work. getting, as you know, the LLMs,

30:42
ingest all human knowledge, right, but they wait it, they wait it and they wait what someone else said about you more highly unpaid, let's call it earned content, They wait that more highly than paid content. so, but that's a never ending, we have to keep going, right? Of course, you know, I love the concept of the infinite game. You know, so many entrepreneurs, I've been doing this for

31:12
since 1994 when I got out of the Navy. So I guess that's what 32 years ago. you know,  most entrepreneurs think of like a business as an episode,  like, okay, I'm gonna do this, I'm gonna win, then I'm gonna do the next thing. Like, it's not, this is not a finite game. Like if you're building an enterprise technology company, there is no... uh

31:34
There is no end to that game. have to constantly be on the innovation treadmill. You have to constantly listen to your customers, constantly track new emerging tech trends, constantly roll out new products, constantly help you. Every layer of technology is always getting commoditized, right? You're constantly moving up the value chain  and it never ends. When you go public, that's just a timestamp. You get some more money in the bank and the next day you're back to work. And I think that's what's

32:03
unique about your organization as well because you see so many, I use my  air quotes, AI companies  that have popped up that haven't been in business as long as you have. And you've curated  your business model over the last several years. this isn't just, we just popped up, we got some investors and we're trying to solve something through a widget. This is core enterprise focus. And to your point, ah the profound rankings, mean, that's amazing to be able to

32:32
to be at a higher standing than the likes of a Google and some of those other big companies in there. mean, that's hats off to you and all the employees. And as a direct result, your customers continue to value that, especially if they're 50 % of probably the speaking components of your forums that you have globally. That's awesome. Yeah, yeah. It's a lot of hard work and paying attention to the things that, excuse me, that like,

33:00
SOC compliance, HIPAA compliance, um privacy regulations and every privacy regime around the world, right? So we've got nine different AWS regions. They're all triple kit redundant. So then that's the other thing is SLA is like, all the hard stuff that um the startups have in front of them.  Teliom has  this great global platform that's  operating, scaling, doing a great job. um

33:30
But like I said, we can't rest. We have to keep going and keep developing and keep building. Yeah, yeah. Even though if you're number one, you got to make sure that you're looking over your shoulder to stay in that spot. And that's what I love about what you guys are doing. That's great. So congrats. I would say if I'm a CIO or a CMO or CDO listening right now who want to...

33:56
make customer data AI ready over the next 12 to 24 months. So you have tremendous amount of experience, you're advising  and producing for clients today. What should that C-suite focus on for their customer data to be AI ready over the next 12 to 24 months? Or if they're getting that board pressure,  to fast track it. Are there any quick hits that they should really focus on in your opinion? um

34:25
I would start with architecture. again, I think that that, you know, top down real time is the most elegant architecture. um just, where should data actually finally be at rest and then  everywhere else it should just be getting processed and move forward. And so think about architecturally, think about it from a compliance standpoint, because you have to, you know, there, there are actually countries where the executives go to jail if they  violate privacy rights. Right. Serious stuff.

34:55
And um so privacy and compliance are critical under thinking about the best, most high-performance architecture where you can have the freshest context for all things AI that also allows you to  abide by all the regulations. Those would be kind of like the navigational buoys, you know, that I would try to drive the boat in between.  And  AI is not just agentic stuff. It's literally embedded in everything.

35:25
Right. And so you just have to,  you really want to think about not sending stale data into your AI stuff.  Um,  you're going to get suboptimal results. You're to pay a lot of money for workloads that don't add value. You know,  that's like the third navigational buoy is, just efficiency because  you will  see a lot of,  um,  solutions to your point pop up to solve this problem, that problem, that problem.

35:54
And um that kind of goes back up to the overall enterprise architecture. What is your enterprise data architecture?  And um let's not create more silos, right? Let's create more consolidation and let's have one high performance  layer of trust where we have our good trusted data and let's everyone drink from that fire hose, right? Not have a bunch of different point to point integrations, because those are expensive. They tend to be batch.

36:24
So you got delayed data going everywhere. Then you've got to run workloads to remunch that data. it's just, that's the batch backward looking hairball. I'd talk about her. Well, to use your analogy, the navigational buoys, it's your red, right return back in the port, right? To make sure that things are, we're all driving on the same side. That's great. That's really, really good sound advice. So speaking advice, we always ask our guests,

36:54
you know, what advice  would you have for college students now who are thinking about a career in AI or just navigating,  again, getting back uh to the navigational buoys here, how would they navigate a career knowing that AI is making that push? And there is  some job disruption,  but also some  ability to upskill uh folks as well.  any recommendation you give to young workers? For sure. I think, um

37:23
You have to start  building.  And, know, the new saying is you can just do things, right. And it really is true. Right. And so, you know, have the agency to just get your hands dirty. Like, you know, I'm an old software engineer, so I've been playing with it longer than most, but I still play around for fun. And like I was on replica last weekend and had it built an app.

37:48
to teach my granddaughter to read, you know, just for fun, to see how it worked and just fun stuff like that. doesn't have to be, you don't have to write a new accounting system, but, then understand how to stand up agents and manage agents, because I agree with the statement that you are, your job will not be replaced by AI, but you're, you may be replaced by someone using AI. And so that's what we're seeing in our companies is, and I'm involved in

38:16
four different companies right now. And these, these people are, are just sort of rising to the top because they've been the more forward leaning, they're  standing up,  uh, Nemo bots and Nemo clause inside our AWS instance, collaborating with our security people. They're curious. They have agency, they're streamlining workflows. You know, we, have one gentleman who,  uh,  was managing our  SDR outbound team  and

38:46
Um,  that team is now basically one third the size with two times the output. So you're talking about,  and we didn't let any of them go. just, the  SDR has high turnover. So you just, basically six X your output  over about an 18 month period  using automation  and some AI content generation, standardization and things like that. So we're getting a.

39:16
better result, better pike productivity. And, you know, that was really  one guy with his team  and inspired team  under him that leaned in and said, let me go, let me go practice code and figure out how we can do this more efficiently. Got another guy in compliance and legal who stood up a bunch of agents and he's got a lead agent and four or five other agents  and uh ex military guy. he, you know, he basically he's just curious and leaning in and

39:46
automating our RFP response process. So we call that RFX. So when you get a big company sends you an RFP with 650 questions, you know, we used to have a whole team of people who responded to those questions full time. That was their job. Now, CCAI replaced my job.  Those  folks didn't really enjoy responding. Like they now are freed up to do more value added stuff. And this gentleman's fleet of agents can do all the RFP responses.

40:15
are at least 90 % of the questions for us and then humans will scan it. so there are just obvious place. So back to the college kid, like whatever you want to do, whatever your passion is sports or art or whatever, you know, there's some way to start experimenting. These tools are free for a reason. That's one thing I do like about all the, uh, all the large labs are giving free level tier free tier, uh, access. so wherever you are in the world,

40:44
You can learn, you can lean in  and the stuff you're using in that free tier, it may be rate limited, but it's more powerful than what the labs were doing nine months ago. Like, I mean, so you have incredible power at your fingertips and just, you know, if you don't know what to do, guess what? Ask the, ask the model what you should do. Right. Right. it'll give you advice. It's attaching to be, to you to become more efficient and having those agents so you can do more faster.  Um, you know, I think I read something the other day, they,

41:14
They likened it to the early days of, know, Excel didn't replace  CPAs, but it made them more efficient. Right. Yeah, exactly. So, you know, and think, you know, we all kind of owe a little bit of that responsibility in the industry to folks that feel like AI, you know, especially  using a public LLM is cheating and it's not, ah you know, you just have to get rid of some of that stigma and show when it's best to be used and not to be used. You know, there's a lot of other ethical

41:43
reasons not to use some of the tools, right? yeah, I think that's interesting. And I completely agree with your assessment there in terms of just start playing with it, whatever your passion is, because you do learn it quickly. And to your point, being an original application developer, I I remember when I first started learning Pascal, and then I eventually moved to Visual Basic, and I thought Visual Basic was like groundbreaking. Now, I mean, I, you

42:13
I understand and remember all the logic on how to build an application, but had I had these tools back then, it would have been a much different ride.  Yeah. I  love Pascal. Wild dew loops all day long. Well, it's,  it's been, it's been great to have you on what, what, what should we close with, um, around Tealium and anything else you wanted to cover before we, uh, we end today.

42:39
You know, just anyone who's curious, email me at Jeff at tealium.com. We love talking to folks out in the industry, anyone who's curious. And, you know, the sun never sets on Tealium. a global company. So wherever you are, we want to try to help and we'll try not to sell too hard. We like to be consultative and just really understand the problem that our prospects and customers are trying to solve and figure out.

43:09
uh if there's a way we can help. And we find that in doing that and then in listening to those problems, that informs our roadmap, you what to build next. So, yeah. That's great. Thank you for having us, Gary. Great, great conversation. appreciate  it. honors all ours. We appreciate  all that you're doing in the ecosystem. Thank you for your service uh as a Navy veteran. And, you know, I would say to our  listeners and those watching,  you know, please, please check out TELUEM.

43:35
especially if they have some of these forums and meetups  globally. I that's a great way to talk to folks in the organization  outside of just Jeff, because obviously he's just one person and  a busy person as well. But that's a great way to understand how to incorporate a company like Telium into your business.  So thank you again for  listening to the episode today. Please  continue to like and

44:00
Subscribe and  pass our episodes on to your network. Until next time, we'll see you soon. Thank you. Thanks, Jeff.