In this episode
Mark Abramowitz, Chief Marketing Officer at Dataiku, joins Greg Kihlström at Ai4 to explain why enterprise AI marketing has moved from selling a vision to proving tangible value — and what that shift demands of marketing leaders. Drawing on Dataiku research covering 900 CEOs across eight countries at companies above $500 million in revenue, Abramowitz describes an environment where nearly every executive assumes AI is already being used inside their company without approval, and where the pressure to show measurable results now reaches every level of the organization. The conversation runs from persona-specific positioning across data, agent-building, and agent-governance buyers, through marketing’s accountability to pipeline and revenue, to the three-tier governance model Dataiku’s own marketing team uses to let marketers build agents safely. Abramowitz closes with a prediction: agents will appear on org charts as a managed workforce, funded, reviewed, and evaluated like labor.
Key takeaways
- Proof, not vision, is what cuts through in enterprise AI marketing. Abramowitz argues the fundamentals of marketing haven’t changed — what wins is customer proof and a community of executives and practitioners who have gotten real ROI, the same pattern that drove early cloud and SaaS adoption.
- 96% of CEOs believe their employees are using generative AI without approval. The finding comes from Dataiku research with 900 CEOs across eight countries at companies with $500 million or more in revenue. Abramowitz’s reaction: he’s surprised it isn’t all of them.
- Roughly 80% of those CEOs expect a peer to be ousted over a failed AI strategy or an AI crisis. The pressure to produce measurable AI results runs from the individual marketing manager to the chief executive.
- The agentic shift is about business outcomes, not productivity gains. Efficiency is the starting point, Abramowitz says, but the destination is measurable business results — which is why the proof burden has moved up.
- AI is not one umbrella market, and messaging has to be persona-specific. Dataiku goes to market on three motions — getting control of your data, building and controlling business processes with agents, and managing agentic sprawl — which map to chief data officers and data scientists, CIOs and chief AI officers, and IT governance and observability teams respectively.
- “Not everybody’s data is better just because there’s AI.” Abramowitz pushes back on the assumption that AI automatically improves marketing analytics: the data foundation still has to be fixed first.
- Nobody outside of marketing should be talking about leads. Abramowitz takes pipeline — specifically stage-two or quality pipeline — and marketing’s contribution in whole dollars to the CRO and executive team, and wants marketing to source more than 50% of Dataiku’s pipeline.
- Clean CRM and signal data is the gate on every marketing agent. The old garbage-in/garbage-out problem now runs faster, at scale, and with agents that take action — a materially different risk than a miscalculation in a spreadsheet.
- A Gold/Silver/Bronze model lets every marketer build agents without losing governance. Bronze covers self-serve builds like writing assistants and messaging digital twins; the middle tier gets adopted by marketing analytics and AI engineering; Gold covers multi-departmental systems that write to Salesforce and 6sense and require full AI engineering support.
- Agents will show up on org charts as a managed workforce. If agents are labor, Abramowitz argues, they get funded, hired, reviewed, evaluated, and potentially fired — and HR systems will need a new construct to manage them.
Chapters
- 0:00 — The gap between AI ambition and business reality
- 1:32 — Mark Abramowitz’s path from Salesforce and ServiceNow to Dataiku
- 2:24 — What Dataiku is: an orchestration and governance layer above data platforms and models
- 3:55 — The SoftBank CRM case: giving sellers thousands of hours back
- 4:52 — The research: 900 CEOs, and the 96% who assume unapproved AI use
- 5:35 — Why 80% of CEOs expect a peer to lose their job over AI
- 6:16 — Proof over vision: how this differs from early cloud and SaaS
- 7:48 — Does the market reward precision about what kind of AI you sell?
- 8:51 — Three go-to-market motions and the personas behind each
- 10:27 — Buying committees at $5B+ enterprises and the case for precision
- 11:56 — “Not everybody’s data is better just because there’s AI”
- 12:43 — Why marketing has to move closer to revenue and to product
- 15:06 — What stays in marketing versus what goes to the CEO and CFO
- 16:28 — Stop reporting leads; report quality pipeline in whole dollars
- 17:33 — What CMOs should actually be asking about their data
- 18:18 — The cargo airline optimizing route offers on Dataiku
- 19:37 — Gold, Silver, Bronze: governing marketer-built agents
- 20:57 — The Gold tier: a new campaign operating model
- 22:15 — Agents as labor, and agents on the org chart
- 23:41 — How Mark Abramowitz stays agile
Why the AI market now demands proof instead of vision
Abramowitz’s read on the current moment is that nothing fundamental about marketing has changed — cloud and SaaS were also futuristic, also a new business model, also a new way to buy. What cuts through in a market like that is customer proof and a community of customers, executives, and practitioners who are willing to say publicly that they got real ROI. What’s different in the agentic era is where the value is claimed: efficiency and productivity are the entry point, but Abramowitz argues the real destination is business outcomes, which raises the evidentiary bar on every claim a vendor makes.
What the 96% statistic actually says about enterprise AI
Dataiku’s research covered 900 CEOs in eight countries at companies above $500 million in revenue. The headline finding — that 96% believe employees are using generative AI without approval — reads to Abramowitz as effectively universal, and he’s surprised the number isn’t higher. Paired with it is a second finding: roughly 80% of those CEOs think one of their peers will be pushed out over a failed AI strategy or an AI-related crisis. Together they describe a market where adoption is already ambient and ungoverned, and where the personal stakes for executives are high enough to distort how AI progress gets described.
Why AI is not one market, and messaging can’t treat it as one
Buyers are maturing quickly, Abramowitz says, and Dataiku’s go-to-market reflects that in three distinct motions: getting control of your data, getting control of business processes with agents, and managing agents built on Dataiku and on other agentic platforms. Each motion has its own persona. The data motion — machine learning, pipelines, predictive work — speaks to chief data officers and data scientists. The agent-building motion speaks to CIOs, chief AI officers, and their IT teams. The agentic-sprawl motion is a governance and observability conversation. At $5 billion-plus enterprises the buying committee is broad enough that precision about who hears which message becomes more important, not less.
Marketing’s move from leads to pipeline to revenue
Abramowitz is direct about the reporting line he wants: nobody outside of marketing should be talking about leads. He doesn’t bring lead counts to the chief revenue officer; he brings marketing’s contribution in whole dollars and as a share of total company pipeline, with a target of sourcing more than half of Dataiku’s pipeline from marketing. He has already committed to a pipeline percentage and expects that as marketing’s systems get more sophisticated he’ll eventually commit to a revenue number — a step he says he would not have considered five or ten years ago, when product marketing owned pipeline and sales owned revenue. He frames the same discipline internally: marketing has to market marketing, telling a data-grounded story about what the CEO and CFO actually care about.
Data quality is the gate on every marketing agent
Asked what CMOs should be pressing on, Abramowitz starts underneath the AI layer. Before building an agent to speed up the lead process, confirm the signals feeding it are right, and confirm that lead and opportunity data in the CRM is clean and accurate. The failure mode is familiar — garbage in, garbage out — but the consequences have changed: it now happens faster, at scale, and the agents take action rather than producing a report. A bad number in a spreadsheet causes a miscalculation; a bad signal into an agent causes a decision.
Gold, Silver, Bronze: governance that still lets marketers build
Dataiku’s marketing team runs a three-tier model. Bronze covers agents individual marketers build themselves in natural language — a writing assistant, a simple digital twin to test messaging, a press-release helper. A marketing council can promote some of those to a middle tier where the marketing analytics and AI engineering teams take ownership while the use case stays marketing-oriented. Gold is reserved for agents with broad multi-departmental business impact — currently a new campaign operating model that optimizes digital spend against account signals, builds campaign artifacts beyond basic briefs, and writes to Salesforce and 6sense. Gold requires full IT and AI engineering support. Abramowitz’s constraint on all of it: he doesn’t want to automate for its own sake, he wants more pipeline, faster and cheaper.
What it means to treat agents as labor
Abramowitz’s prediction for the year ahead follows the “agents as a workforce” framing to its conclusion. If agents genuinely are a workforce, they should be treated like labor — funded by the business, hired by the business, reviewed, evaluated, and potentially fired. He expects to see agents on org charts for real, not as a replacement for headcount but as an expansion of capability, and expects HR management systems like Workday to need a new construct for evaluating and managing them alongside people.
FAQ
What did Dataiku’s CEO research find about unapproved AI use? Dataiku interviewed 900 CEOs across eight countries at companies with more than $500 million in revenue. 96% said they believe their employees are using generative AI without approval, and roughly 80% expect one of their peers to be ousted over a failed AI strategy or AI crisis.
How is the current AI market different from early cloud and SaaS adoption? Abramowitz says the marketing fundamentals are unchanged — both were futuristic new markets with new business models — but the way to cut through today is customer proof and a community of practitioners with demonstrated ROI, rather than vision alone. The agentic wave also shifts the promise from productivity gains to business outcomes.
What metrics should a CMO report to the CEO and CFO? Pipeline contribution, expressed in whole dollars and as a share of total company pipeline, plus quality or stage-two pipeline. Abramowitz argues leads should stay inside marketing and never appear in executive conversations.
What should marketers verify before building AI agents on their data? That the underlying data is accurate — CRM lead and opportunity records, and intent signals from platforms like 6sense. Garbage-in/garbage-out now operates at speed and scale, and agents act on the output rather than simply reporting it.
How does Dataiku govern agents built by its own marketing team? Through a three-tier Gold/Silver/Bronze model: Bronze for self-serve agents built by individual marketers, a middle tier owned by marketing analytics and AI engineering, and Gold for multi-departmental systems that require full IT and AI engineering support.
Will AI agents appear on organizational charts? Abramowitz expects so. His argument is that if agents are a workforce they should be treated like labor — funded, hired, reviewed, evaluated, and fired — which will require HR management systems to add a construct for managing agents the way companies manage people.
About Mark Abramowitz
Mark Abramowitz is Chief Marketing Officer at Dataiku, The Platform for AI Success. In his role, he oversees global brand and demand strategy, and is responsible for building awareness of Dataiku as the platform that unifies people, governance, and AI systems to deliver real results at scale.
Mark brings extensive experience scaling enterprise platform businesses. Prior to Dataiku, he served as SVP of Product and Solutions Marketing at ServiceNow, leading product marketing as the company scaled from $6 billion to nearly $11 billion in revenue and spearheading the launch of Now Assist—the fastest-selling product in ServiceNow history. Before ServiceNow, Mark spent 15 years at Salesforce and 5 years as SVP of Product Marketing for Service Cloud.
Mark Abramowitz on LinkedIn: https://www.linkedin.com/in/markabramowitz/
———- Resources ———-
Dataiku: https://www.dataiku.com
The Agile Brand podcast is brought to you by TEKsystems. Learn more here: https://aglbrnd.co/r/2868abd8085a9703
We’re proud to be a media partner for #MAICON26 – Oct. 13-15! Learn how AI can power your marketing and business and help you grow smarter. Use code AGILE150 to save! https://aglbrnd.co/r/7fe458ced0f04658
Reach your customers with Reddit. Spend $500 in ad spend, get $500 back in ad credit! Learn more: https://advertalize.com/r/491818c79fb1873f
Chaser is the only Slack-native project management platform that helps teams turn messages into tracked tasks, automate follow-ups, and maintain team-wide visibility, without adopting another tool. Now integrated with Claude and other GenAI tools. Learn more at trychaser.com and use code AGILEBRAND for a 3-month free trial (normal trial is 14 days).
The most influential minds in software, AI, and engineering leadership will be at WeAreDevelopers World Congress North America, September 23-25 in San Jose. Learn more: https://aglbrnd.co/r/60a7299222a7bcf1
Start building your own apps with Replit and get $20 off. Learn more: https://aglbrnd.co/r/93531742a7625a20
Enjoyed the show? Tell us more at and give us a rating so others can find the show at: https://aglbrnd.co/r/faaed112fc9887f3
Connect with Greg on LinkedIn: https://www.linkedin.com/in/gregkihlstrom
Don’t miss a thing: get the latest episodes, sign up for our newsletter and more: https://aglbrnd.co/r/35ded3ccfb6716ba
Check out The Agile Brand Guide website with articles, insights, and Martechipedia, the wiki for marketing technology: https://www.agilebrandguide.com
The Agile Brand is produced by Missing Link—a Latina-owned strategy-driven, creatively fueled production co-op. From ideation to creation, they craft human connections through intelligent, engaging and informative content. https://www.missinglink.company
Transcript
[00:00:00] Greg Kihlström: Hi, I’m Greg Kihlström, host of The Agile Brand, and here’s a question for you: What happens to a brand when the story it tells about AI moves faster than the results it can actually show? Agility isn’t about telling that story faster. It’s about being able to change what you say repeatedly as the technology changes, without losing the thread of what your company is actually for. Today we’re going to talk about the widening gap between AI ambition and business reality, and what that means for enterprise marketing leaders. Specifically, we’re gonna cover the shift from selling an AI vision to proving tangible value, how AI is pushing marketing into closer alignment with product and revenue, and what it takes to build trust inside the company and out when the outputs of AI systems are hard to explain. To help me discuss this topic, I’d like to welcome Mark Abramowitz,
[00:01:32] Greg Kihlström: Chief Marketing Officer at Dataiku. So, before we dive in, why don’t you give a little background on yourself and Dataiku?
[00:01:39] Mark Abramowitz: Sure. So Mark Abramowitz. I’m the Chief Marketing Officer at Dataiku. I’ve now been here for almost a year, so not new anymore. and quick background on me, I’ve been in enterprise software for a long- more years than I can count, for the most part. I was in, at Salesforce early on, when cloud was still new. I was at ServiceNow for a number of years leading product marketing, during the workflow transformation that was going on as, as well as when ChatGPT hit. It was like the pinnacle of the product marketing opportunity of fastest inno- faster innovations or take to market, more launches, more positioning, more messaging, more sales enablement, and I got a, a like, a great, great,
[00:02:24] Mark Abramowitz: great experience. And then I decided, “HI’ve been working at two of the largest software companies in growth modes, for a long time, and I’m gonna go take the things that I’ve learnt and the relationships that I’ve built and go, go try and be a chief marketing officer.” And so Dataiku is the platform for AI success. Our– We are, an orchestration layer that sits above your data platforms, your language models, and we are, deliver and bring together your people, which are technical builders and non-technical builders. We orchestrate all of the complexity that is in a large regulated industry, which is who we serve. So again, your data platforms, your language models,
[00:03:10] Mark Abramowitz: on-prem as well as in the cloud, and pull that together with a, with built-in governance. So when you build a, an agent or a agentic workflow or a data pipeline, you can trust it because from the ground in, ground up, we’ve got the workflow and the lineage and the, and the visibility to be able to answer the questions on how, how a model, whatever model or type it was, actually made a question. And so we serve some of the largest organizations in the world, one in four of the glo– of the Fortune 500, so companies like LVMH, like Johnson & Johnson, like Pfizer, like SoftBank. SoftBank is a great example. We’ve really helped them on a CRM use case, which
[00:03:55] Mark Abramowitz: was really cool, to help their– to give their sellers thousands of hours back from having to enter data, as well as bring, like, call notes and meeting notes together. So we gave them a sort of a chat interface to it and are saving them thousands of hours and letting them go sell more. And the opportunity at Dataiku is to deliver and– or to help our customers be successful with AI.
[00:04:19] Mark Abramowitz: Great to be here, and great to be here at AI Four.
[00:04:21] Greg Kihlström: Yeah, right, I know, and great job this morning at the keynote.
[00:04:24] Mark Abramowitz: Thank you. Thank you. It was really, really fun opportunity to get up and I think, well, deliver a slightly different style of talk than maybe the, the, the other folks who were giving keynotes this morning. Trying to really focus on some of the first-person research that we’ve done-
[00:04:40] Mark Abramowitz: … and just try and give a provocative point of view.
[00:04:42] Greg Kihlström: Yeah, yeah. The, the– I– To diverge a little bit, like the 96%-
[00:04:48] Mark Abramowitz: Of CEOs
[00:04:49] Greg Kihlström: … stat. Yeah. Can you maybe share, share that real quick?
[00:04:52] Mark Abramowitz: Sure. So earlier this year, we interviewed nine, nine hundred CEOs from eight different coun-countries at fairly large companies, five hundred million and above in revenue. And one of the key stats that came out of it, there were a few, but one is that 96% of CEOs believe that their employees are using generative AI without approval.
[00:05:14] Mark Abramowitz: So that’s pretty much all of them-
[00:05:16] Mark Abramowitz: … if you will.
[00:05:16] Mark Abramowitz: I’d be surpris- I’m surprised it’s not all of them because I think as we know, whether it’s sanctioned or approved or not, like everybody’s doing it, everybody’s using it.
[00:05:26] Mark Abramowitz: I thought one of the other interesting stats out of that report is just the way, CEOs actually see themselves-
[00:05:35] Mark Abramowitz: … that they are under a lot of pressure to deliver on AI results. And I think it was, 80% think one of their peers is gonna get ousted-
[00:05:47] Greg Kihlström: Yeah
[00:05:48] Mark Abramowitz: … if there’s a failed AI strategy or some kind of AI crisis.
[00:05:51] Mark Abramowitz: And so you might be a, an individual marketer, marketing manager working on something, or you might be the CEO, and we’re all under pressure-
[00:06:00] Mark Abramowitz: … to figure out how to get measurable results from AI.
[00:06:02] Greg Kihlström: Yeah, yeah, definitely. d- some, some interesting research there. So yeah, let’s, let’s dive in here and, you know, as you mentioned, your experience, you’ve seen a few shifts in, in your career, as, as you mentioned. Um-
[00:06:16] Greg Kihlström: How is the current demand for what I’ll call proof over vision in AI different from, say, early days of cloud or SaaS adoption?
[00:06:27] Mark Abramowitz: So I’ve been thinking about that question, and one of the, one of the things that I believe is that the fundamentals of marketing are not changing. How you deliver the tools that we have at our disposal is absolutely changing. But I think, ’cause SaaS or cloud also was futuristic. It was a new market, it was a new deploym- it was a new business model for the, for, for us at Salesforce, a new way to buy, as well as a new technology. And so I think then, today, the way you cut through that is with proof, customer proof, and a community of, of, of, uh,
[00:07:12] Mark Abramowitz: customers, executives, and sort of practitioners that are passionate, believe in what you’re doing, and have actually gotten, real ROI out of those solutions. And today it’s about building agents, and it starts with efficiency and productivity, but I think ultimately agents and this agentic move is about business outcomes, rather than productivity gains necessarily. And that was the same back then, is like, it was really about finding, bringing new products to market, finding customers that would adopt them, tell their story, and turn them into heroes.
[00:07:48] Greg Kihlström: Yeah. And you know, I think AI is a– AI, I’ll just say, is a very broad umbrella, right? I mean, there’s, there’s a lot of things that, that fit under that, everything from agents that can, you know, operate end to end to deliver something to many, many other things. How much of that distinction does marketing actually– does your marketing actually make? And, you know, does the market reward precision in talking about, you know, different types of AI, or is it, is everybody still just buying this, like, umbrella basically?
[00:08:23] Mark Abramowitz: So, I think buyers are, like the market, are maturing and becoming more sophisticated really quickly.
[00:08:33] Mark Abramowitz: And so I think, from our experience at Dataiku is we kind of play in three– o-on top of this, like bringing people and orchestration and governance together, there’s kind of three, th-three ways we go to market. one is kinda how do you get control of your data?
[00:08:51] Mark Abramowitz: Number two is how do you, how do you build and get control of your business processes with agents? And then how do you manage agents built on Dataiku and built on any of the other agentic platforms?
[00:09:03] Mark Abramowitz: And so we already see in our go-to-market, to your point, there’s variations. The data world has been machine learning and pipelines-
[00:09:11] Mark Abramowitz: … and predictive, which is– and getting your data ready. That still matters a lot, and those buyers, whether it be the chief data officer or data scientist, like you have to have very specific messaging for those types of personas. for the agentic process world or building agents, you may be talking more to the CIO or chief AI officer and their teams in IT, and that’s newer, but quickly becoming, I think, less about picking a platform and stack to focusing on business outcomes.
[00:09:42] Mark Abramowitz: And then on the agentic sprawl, which is here, I believe, it’s also more of an IT, governance observability. So when you put all that together, what it forces us to do is to be very persona specific in the way that we bring our positioning and messaging. So– And we wanna land, we can land in each one of those three areas. so we target the data teams, the AI and CIO teams, and the sort of IT operational teams. What’s interesting is the large enterprises that we sell to, $5 billion plus companies, is the breadth of the buying committee, which actually makes it even more important to be very precise about who we’re talking
[00:10:27] Mark Abramowitz: to about what. and so I think it’s, it’s a result of the shift in the breadth of AI, but also we have a real opportunity to use AI in that marketing-
[00:10:39] Mark Abramowitz: … to get very specific about targeting, not necessarily building all that content, but at least more sophisticated in, like, knowing that you are someone interested in data or interested in agents or interested in managing or governing those agents.
[00:10:54] Greg Kihlström: Yeah. And maybe a parallel, I’ll say, trend, is just marketing has evolved quite a bit over the last few years. As you know, it was always kind of a marketing wants to do something kind of, I’ll just say eye roll, sort of like, “Well, what’s the data to back that up?” Or as data gets better, marketing gets more accurate, and, I think marketing is able to prove ROI better, but there’s also a lot more pressure on marketing to be closer to things like revenue, to trust, even to product reality. What is, first of all, you know, is do you think that’s a, that’s a fair assessment? And where does, you know,
[00:11:39] Greg Kihlström: w- where does marketing in your realm, like, kind of fit within that framework? Like, are you– Are, is there, is there– Are you being held to greater results? Are marketers in general being he-held to greater results just because the data is better and, and there’s more opportunity for the data?
[00:11:56] Mark Abramowitz: Maybe my first comment would be not everybody’s data is better just because there’s AI.
[00:12:00] Greg Kihlström: Oh, fair, fair point. Fair point. Yeah.
[00:12:02] Mark Abramowitz: And so, I think, as we, like, work with customers like Roche or LVMH or Johnson & Johnson, some of which are in some, marketing and sales-oriented use cases-
[00:12:15] Mark Abramowitz: … it still goes back to the how good is your data.
[00:12:20] Mark Abramowitz: And so there’s always work to be done, um- We are a very data-centric company, but there are all these systems, right? For CRM, for, for ABM, for, call recording, for and for. And so I think there’s still a real… Like, you still have to get your data. I mean, you gotta get your data right-
[00:12:41] Mark Abramowitz: … before you’re gonna get real marketing analytics out of it.
[00:12:43] Mark Abramowitz: But what I think, and maybe this is part of my product marketing background, is I’ve always believed it needs to be about the, the business outcome, which is revenue, and it needs to be closer to the product. So I think I would… The way I look at it is I think it’s a real opportunity right now, ’cause I think that is true as you look at the AI native companies and some of the– and all of us that are going through the tech transformation in tech.
[00:13:12] Mark Abramowitz: It’s positioning, it’s messaging, and I think it’s, revenue, and some, some of it’s in the product, but the products are more complicated now than I think ever.
[00:13:23] Mark Abramowitz: And so brand still matters a lot. but I feel very comfortable in my role as CMO at Dataiku that I’m very aligned with our chief revenue officer. I’ve committed to, you know, what percentage of pipeline we’re gonna contribute from marketing, and I feel great or fine, maybe not great, but fine-
[00:13:42] Mark Abramowitz: … about being held accountable to that. and I think one day as the systems we build in marketing become more so- more sophisticated, I, I’m, I will eventually get to a, I wanna commit to a revenue number.
[00:13:54] Mark Abramowitz: Which 10 years ago, five years ago, or less, I would never have thought that marketing should connect. Like, what I’ve… Growing up in the large B2B enterprise software companies-
[00:14:07] Mark Abramowitz: … it’s, well, product marketing is about pipeline. sales is about revenue. But I think-
[00:14:13] Mark Abramowitz: … those processes and those silos are getting broken down, whether it’s by technology or relationship. And, I want marketing to be a significant contributor to the business success-
[00:14:26] Mark Abramowitz: … not just the marketing success. It starts with brand and pipeline, but I think eventually it’s revenue. And I also see my role as part of the Dataiku transformation to drive product innovation. Like I w- some of that’s because of my background, but it’s really exciting for me to be part of the roadmap and strategy discussions and being an influencer-
[00:14:48] Mark Abramowitz: … over what we should build, what’s the right product market fit for new products that we’re gonna go take out. And so I love the let’s be more revenue or business oriented and let’s be more product oriented.
[00:15:06] Greg Kihlström: What I’m seeing a lot i- is just that there’s a lot of pressure for CMOs and, and others to contribute more directly to revenue. But to your point, if the data’s not there, there, there’s an expectation that is unmet or unable to be met in the way that maybe a CEO, CFO, COO are, are expecting. So I mean, all the more reason for companies like Dataiku, I guess, as well, is, is, you know, to support that, right?
[00:15:35] Mark Abramowitz: Yeah. And but, so yes, and I think what, what, the way I look at that is the way marketing… There’s a lot of things we report on in marketing-
[00:15:43] Mark Abramowitz: … that stay in marketing.
[00:15:45] Mark Abramowitz: Rather than things that I would bring up to our executive team of CEO, CFO, or chief revenue officer.
[00:15:52] Mark Abramowitz: And so I think marketing still has to do a good job of marketing marketing within the company, and it’s hard to do when you’re caught up in the go to market, the technology, the transformation. But I think you got, we gotta apply many of those same principles that we take to marketing to prospects and customers of-
[00:16:12] Mark Abramowitz: … think about the CEO, think about the CRO, what do they care about, and how do you, tell a story that’s grounded in data that is aligned to w- not just what they wanna hear, but the things that they’re, that they want to talk about. Um-
[00:16:28] Mark Abramowitz: … so, like leads, for example, I don’t think anybody outside of marketing should be talking about leads. I, I don’t wanna talk about leads. I wanna talk about pipeline.
[00:16:37] Mark Abramowitz: Stage two or, like, quality pipeline. And so similarly with the sales organization, like, I don’t go to the CRO and say how many leads we generated. I go and talk about the contribution of marketing in whole dollars, but also over the overall pipeline for the company ’cause I wanna be relevant. I wanna be over 50% of the pipeline that Dataiku is, is generating, I want to come from marketing.
[00:17:00] Greg Kihlström: Yeah. Yeah. So, from the marketer’s perspective, so, you know, I would imagine a lot of your customers are not necessarily the CMO, but they’re probably sitting around… They’re, they’re at the table, but not necessarily the direct buyer. But, you know, I wonder, we have a lot of CMOs, SVPs, marketing listening to this show. What should they be asking, you know, even, even if it’s somebody else’s line item or something like that? Like, what, what should a CMO be asking when, you know, when, when they’re talking about some of the services that you offer?
[00:17:33] Mark Abramowitz: So we actually do have, some marketing, marketing use cases in our customer base. we actually have a global airline that uses us to, optimize route offers. So it’s a cargo airline, and they have many data sources sitting on different platforms, and they have built some models to, help them prioritize, like, what to offer one of their customers as an optimized route and at a particular price. And so I think as a marketer, maybe I’ll, I’ll frame it this way, as a chief marketing officer, l- who, like most of us, we got all the– like CIOs, we got all the tech we need, right? Like a Dataiku. We have Salesforce,
[00:18:18] Mark Abramowitz: we have-
[00:18:19] Greg Kihlström: Right. Right
[00:18:19] Mark Abramowitz: … we have, Sixth Sense, we have Gong, we have, things I can’t, I can’t remember right now.
[00:18:26] Mark Abramowitz: And for us, I think one of the starting points is that all that data sits in Snowflake. And so what marketers should be asking for is, how do we make sure that the data is, accur- or is, is right before you start building agents on top of it? So if you’re gonna go build, an agent to, like, to, increase the velocity of your lead proc- your lead process, well, are you sure the signals from it, from, 6sense are right? Are you… Even more, even harder, is, is your op- is your lead and opportunity data in your CRM clean and accurate? So even though what we do on top of it is so much more sophisticated, I think the, the core
[00:19:11] Mark Abramowitz: is still, do you have trusted data that you’re then gonna build on? ‘Cause if you don’t, your agents are just gonna deliver, like, it’s the same thing as it always was for reports or dashboards, like garbage in and garbage out. Now it just happens much faster and at scale, and, like, these agents take action, right? Very different to putting that data in a spreadsheet where the worst thing that could happen is maybe a miscalculation.
[00:19:37] Mark Abramowitz: But now these agents, take action. And so I also believe governance is super important-
[00:19:48] Mark Abramowitz: … even for a marketing team. That as you… Which we’re trying to do at Dataiku, is I want all of my marketers to build agents, but I want them to do it in a way that, is governed and safe. And so we’ve actually built a very simple three-tier governance model, basically Gold, Silver, and Bronze. Bronze agents are things, you go into Dataiku, you have some agent-building capabilities, natural language. What do you– You wanna create a writing assistant, or you wanna create, a simple digital twin to test your messaging, or you wanna, write better press releases. Some of it is… And some are more process-oriented. They are empowered to go build those. And then we have a, a, a, a council in marketing to say, “Okay, some of those Bronze,
[00:20:33] Mark Abramowitz: we might want the marketing analytics team and our AI engineering team to take over,” but they’re still really marketing oriented, and so there’s, like, this middle tier. But then the Gold are the ones that have real broad multi-departmental business impact.
[00:20:49] Mark Abramowitz: So for us, like, the Gold agents that we’re working on now are really about building a new campaign operating model.
[00:20:57] Mark Abramowitz: We are very account-based in our go-to market, so how do we use signals? How do we, how do we optimize our digital spend based on signals of who’s showing interest? But then how do we also, make sure the campaigns get built with not just basic briefs, but helping us build some of the artifacts that you would need to c- to go and run a campaign? And that is, touches lots of different systems. It writes to Salesforce, it writes to 6sense, it generates content. So that is something we, we put in the Gold level that requires, like, IT or full AI engineering support to go and build, but it’s all built on Dataiku, and it’s all part of getting the marketers to be literate,
[00:21:43] Mark Abramowitz: but also to have AI or marketing have a real business impact. Like, I don’t wanna automate stuff with AI just to automate stuff. I want to deliver more pipeline faster, cheaper, better. I know it’s a little cliché, but, like, those are the things that are driving the things that are my priorities right now.
[00:22:00] Greg Kihlström: Yeah. Yeah. Well, well Mark, thanks so much for joining. two last questions for you as we wrap up here. So if, when we’re back at AI For next year, if we’re having this conversation then, what is one thing we would definitely be talking about?
[00:22:15] Mark Abramowitz: So I did a keynote this morning, as you know with one of my colleagues at Dataiku, and one of the, the core premises was, everywhere you go, including here at AI For, but podcasts and other keynotes, is you hear people talking about agents as a workforce. And so one of the, the things we sort of explored here today was, well, what happens when you take that to its natural conclusion? And that natural conclusion is, if agents are a workforce, we should treat them like labor. And when you treat them like labor, that means they are much, they’re, they are sort of funded by the business. They are hired by the business. They are reviewed, they are evaluated, they are fired, potentially, by the business.
[00:22:58] Greg Kihlström: Right.
[00:22:59] Mark Abramowitz: And so I think that, next year at AI For, we’re gonna see agents on org charts for real, and that they will be, um… Not that there will be less people on the org chart, but I think it’s going to expand the capabilities in marketing in particular, but also in lots of other org, other parts of the company that you will start to see Workday or the HR management systems will have to come up with a new, a new component or a new construct, which is, you know, how do– Again, how do I evaluate, how do I review, how do I manage my agents like I manage my people?
[00:23:34] Greg Kihlström: Yeah. Love it. Well, last question for you. what do you do to stay agile in your role, and how do you find a way to do it consistently?
[00:23:41] Mark Abramowitz: Yeah. it’s a good question. So I think it’s part of outside of work. So, I think exercise is really important, so that is a, an unlock for me to find time to have the brain turn off and actually just focus on not family and not work, but the moment at hand, which gives some, some refresh, if you will. and then from a… So that creates, that requires commitment but then it creates head space to think about what’s next, and allocate time every week to do some non-tactical, non… Like, some bigger thinking and I also think being hands-on with these tools is really important. And so is there something to vibe code? Is there something to explore? How do you try find time to do, build some things with Claude code? and for me, those things go together of you need to take a break, and for me, a lot of that is, is exercise or travel to free up space in your head to go and apply it, outside of the day-to-day things that we’re all wrapped up in.
