CDO Magazine Podcast Series · 2026-08-17 · 7 min
Key moments - from our scoring
Substance score
45 / 100
Five dimensions, 20 points each
The shift from structured to unstructured data management represents a fundamental challenge that has persisted since the personal computer revolution enabled free-text data entry in the 1970s. Kyle DuPont argues that AI now unlocks the "80% unlock" - the ability to make sense of unstructured files that have historically been opaque black holes in enterprise systems. Data X-Ray positions itself as a foundational metadata intelligence layer that answers the real question business users ask: not "I need data intelligence" but "I need to find the files that matter for my job." This applies across personas: compliance and privacy officers seeking exposed files for regulatory fulfillment, business analysts building bid proposals for defense contractors, and other roles requiring rapid, intelligent file discovery. Rather than a cleanup tool, the platform functions as an intelligence mesh that provisions metadata to AI agents, enabling them to match user queries to relevant documents. The extensible architecture allows organizations to build custom modules on the metadata foundation, turning unstructured data governance from a security and protection problem into a value-creation problem.
AI enables organizations to move beyond file protection and finally understand the content and relevance of unstructured documents at scale, creating metadata intelligence that connects specific users to the files they actually need for their job rather than requiring manual analysis of entire repositories.
Data intelligence refers to understanding the content of files themselves, while metadata intelligence - Data X-Ray's focus - creates a foundational layer that tracks file relevance, exposure, and applicability to specific user workflows, queries, and business outcomes.
The platform enables compliance and privacy professionals to quickly understand which files are exposed or relevant to specific regulatory requirements or subject access requests, rather than manually searching through unstructured data repositories.
The rebranding reflects a strategic pivot to capitalize on AI's newfound ability to extract intelligence from unstructured data; the original company predated the AI revolution and focused on file protection rather than intelligence provisioning.
Our reviewer’s read on each dimension, with quotes from the episode.
The episode covers the legitimate problem of unstructured data intelligence and positions AI as an enabler, but relies heavily on broad conceptual framing (the 1970s 'original sin', the 80% unlock) rather than concrete operational insights. While the defense contractor bid-proposal example is useful, most claims remain at the architectural or strategic level without deep tactical depth that would materially advance a CDO's thinking.
we're now finally at a place with AI where we can actually not only understand the structured portion, which has been the traditional kind of data management focus, but also the unstructured portion
it's the kind of 80% unlock that we have, uh, that we have available to us now with these new techniques that AI provides us
The framing of unstructured data as 'the black hole' and the idea that AI enables new value extraction are established narratives by 2024. The defense contractor use case is moderately concrete but not particularly novel - the pattern of AI unlocking metadata and relevance is well-trodden. The first-principles breakdown of personas (compliance vs. business analyst) is sensible but not counterintuitive or contrarian.
if you think about the original sin of files, it was, you know, the moment in the, in the 70s where you gave people the ability to enter free text into a computer
they don't ever ask themselves like I need data intelligence. They ask themselves like I need metadata to be able to do the thing that I'm trying to do
Kyle DuPont is the founder and CEO of Data X-Ray with ~8 years of operating history and a background in financial services and technology. This is legitimate practitioner credibility, though his background is primarily as a vendor/founder rather than a buyer-side operator or executive managing large data organizations. He has relevant experience but is speaking from a vendor perspective on a niche product category.
I actually spent about uh, 10 years, uh, in Japan working in uh, various aspects of financial services and technology
we're both kind of tech geeks, um, have been around the data world for a really long time
The episode includes one concrete client example (defense contractor using AI for bid proposals) and a quantified mention ('a billion files that are unstructured'), but lacks Named companies (other than passing reference), financial metrics, timelines, ROI figures, or measurable outcomes. Most claims remain conceptual; the single use case is illustrative but not detailed with hard results or before/after metrics.
one of our clients is a big defense contractor and they want to be able to not only understand uh, what file is relevant, they actually want to be able to take that file and add it to an intelligence kind of mesh essentially that allows agents to help them win bid proposals for, you know, in their case, Department of Defense
we have a billion files that are, that are unstructured
The host asks open-ended setup questions but rarely pushes back, challenges, or digs into specifics. Questions are mostly invitational (e.g., 'tell me a little bit about yourself') or serve as segues rather than sharp probes. The interviewer does attempt to reframe (cleanup crew vs. guardrail) and shows domain knowledge, but there are no follow-ups to vague answers, no requests for concrete metrics, and no productive tension - the interview reads as a friendly introduction rather than a rigorous interrogation.
Why don't you tell the audience a little bit about yourself
So tell me a little bit why you're so passionate about Data X Ray and data intelligence
Computed from the transcript - who did the talking, and the words that came up most.
Kyle DuPont, Founder and CEO of Data X-Ray, speaks with Maribeth Achterberg, CDO Magazine Editorial Board Member, about how AI is changing the value of unstructured data. The conversation focuses on metadata, file relevance, and how enterprises can help users and agents find the information needed for specific workflows.
Transcribed and scored by The B2B Podcast Index.
Speaker A: Foreign. Hello and welcome to the CDO magazine interview series. I'm Mary Beth Hochtenberg, a CDO magazine executive and an editorial board member and founder and CEO of Verity Digital Advisory. Today I have the pleasure of speaking with Kyle Dupont, CEO and founder of Data X Ray. Welcome Kyle. Why don't you tell the audience a little bit about yourself.
Speaker B: Thanks very much, Mary Beth. Really pleased to be here. Where to start? Um, you know, we're going to start with my adult life. Um, I actually met my co founder and uh, at Georgia Tech where we met with both the schools. So we're both kind of tech geeks, um, have been around the data world for a really long time. Um, my co founder Alistair ended up taking the more academic group. But I actually spent about uh, 10 years, uh, in Japan working in uh, various aspects of financial services and technology. Um, we ended up starting uh, a Hololen Data X ray in London, uh, where we spent about five years. And now I am, um, based in Atlanta where I have uh, my family and me, uh, here in Atlanta.
Speaker A: Uh, it's so interesting, it's always interesting to understand where people got their start and why they get into the things that they do. And you founded Ohalo and now you're rebranding it Data X Ray. So tell me a little bit why you're so passionate about Data X Ray and data intelligence.
Speaker B: I think we're just like at a really cool spot um, in the world right now. Right. Like if you think about the original sin of files, it was, you know, the moment in the, in the 70s where you gave people the ability to enter free text into a computer. Uh, in the personal computer revolution, you know, you ended up with, with just a ton of data being entered into computers. And it just kind of is this black hole that, you know, you just forget about what's in your data all the time. And fast forward, you know, 50 years, we're now finally at a place with AI where we can actually not only understand the structured portion, which has been the traditional kind of data management focus, but also the unstructured portion, uh, where we can actually uh, understand and parse and make use of all the different intelligence that we've built over the years, uh, with, with AI. So I'm really excited about it because of that. Um, it's, it's the kind of 80% unlock that we have, uh, that we have available to us now with these new techniques that AI provides us.
Speaker A: Yeah, honestly, I lived a lot of that, probably 35 years of that, 50 years. And so I completely understand what you're talking about of this and the unstructured data. Um, so how do you think AI creates a creator need for the value that your platform provides?
Speaker B: Well, uh, to answer that question, I probably have to go back a few years, um, because we've been a company for a while now, about eight years old at this point. Um, and so we come from a pedigree that's before the AI, uh revolution. And you kind of have to think about what unstructured data was about at that point. And it was about how is this file being protected really? Nobody actually could usefully analyze the file. And so it was about, let's just protect the file and have people analyze it. Now we have these new techniques with AI, uh, that allow you to just get so much more value out of it. Um, you know, as an example, like one of our clients is a big defense contractor and they want to be able to not only understand uh, what file is relevant, they actually want to be able to take that file and add it to an intelligence kind of mesh essentially that allows agents to help them win bid proposals for, you know, in their case, Department of Defense. Uh, we have kind of several stories like this, but it's really about how do we actually provision that data and that metadata to agents. They know what files are relevant to what user queries and can help them build the end user business value that uh, these, these end users are seeking.
Speaker A: Yeah, so, and I, I can relate to that because you know a great deal my career was trying to set and wrap up that reference data and metadata so you could understand the file, but now you can actually get in and read the file itself. But I wonder, so do you think about data intelligence as um, kind of a cleanup crew or ah, in a compliance guardrail. How do you think these capabilities create value for your customers?
Speaker B: Well let's actually take first principles approach. So I am uh, a number of different business users. So think about just a generic compliance, uh, Persona or privacy person. And then like separately from that, let's think about like a business analyst user. I can use that example, I just talked about it, about the bid people, uh, that are trying to build these proposals. Um, and let's say you want to use AI to help you out. Like I want to use AI if I'm a privacy person to help me understand what files are exposed so I can, you know, do something about it or maybe fulfill a subject access request or be compliant with the ABC acronym that you have. You're trying to be compliant with at whatever point, um, or uh, I'm the person trying to build the bid and I need to know what files are. So like, the question is not what files. Um, they don't ever ask themselves like I need data intelligence. They ask themselves like I need metadata to be able to do the thing that I'm trying to do. So the apiary question is like, how do we actually give them the metadata that they need? And so how we conceptualize data X ray is really as a foundational metadata layer upon which we have now different modules that do different end user workflows or values. Um, and we're super happy, you know, to allow the other, the users to create their own modules as well. So it's about extensibility of that metadata layer into these modules that we uh, that we're really uh, double clicking, um, on Gotcha.
Speaker A: So it's less about cleaning up the data, it's more about exposure to and organizing it so that you can get to where the value is correct.
Speaker B: Yeah, yeah. Like even if 80% of your data is unstructured, we have a billion files that are, that are unstructured. Well, like for any one user, it's not true that they need the 80% of the data. They probably need like a handful of files that actually get their job done. Right. So like, how do you go from a billion files to the handful of files that somebody actually needs?
Speaker A: Kyle, before we part, is there anything else you'd like to, um, say?
Speaker B: No, I just really appreciated the conversation today. Mary Beth, it's been great talking to you and uh, I hope everybody, uh, if they have any questions on unstructured data or how to, how to structure it, how to understand what's in it, please give us a ring. We're always happy to discuss. So, uh, really appreciate the time and let's structure unstructured data.
Speaker A: Yeah, absolutely. Well, thank you Kyle, for joining me today. And for more interviews and insights, Please please visit CDomagazine Tech. Have a great day.
Other episodes covering the same guests and topics, from across The B2B Podcast Index.
Why AI Pilots Fail: How to Escape AI Pilot Purgatory and Scale Enterprise AI with Ronnie Kwesi ColemanUsing AI at Work · on AI agents92 / 100
More Agents Than Employees: How Zapier Disrupted Itself Before AI CouldTalking AI · on AI agents90 / 100
PMs Are Fintech: How Column is Rethinking Property Management BankingThe Profitable Property Management Podcast · on AI agents89 / 100
Inside an AI-powered personal dashboard | Matthew TiemannPossible · on AI agents83 / 100
90% of AI prototypes never reach production (w/ Temporal's Samar Abbas) | AI BasicsThis Week in Startups · on AI agents81 / 100
From Billable Hours to Business Outcomes: How AI Is Rewriting the Rules of IT ConsultingEvolving the Enterprise · on AI agents78 / 100
Building Trusted Self-Serve Analytics for Better Commercial Decisions with Ashutosh Katiyar51 / 100
AI Agents Join the Workforce: Muralidhar Vemulapalli on Why They Need Accountability 
Why Better Data Beats Bigger AI Models: Enterprise AI Executive Rohit Chauhan
AI Governance Makes Innovation Scalable in Trust: Insights from Muralidhar Vemulapalli
No Use Case, No Value: Mindfuel CEO Nadiem von Heydebrand on Proving AI ROI