At Knowledge 26, it was impossible to escape AI. Every keynote, breakout session, roadmap discussion, and vendor booth somehow came back to AI Agents, orchestration, governance, or automation.
That was expected. However, what I did not expect was how often CMDB came up.
I’ll skip the major announcements like Otto and Armis here; they’ve already been covered by others. I’ll focus on my observations.
At Knowledge25, AI created urgency around data quality. At K26, the conversation has shifted. Data quality is no longer about keeping the CMDB clean. It is becoming the foundation on which enterprise AI depends. (K25 Reflections piece)
Most vendors in the expo area communicated something related to AI, regardless of product or service. For better or worse, we decided to stick with our years-old slogan “Improve Data Quality in ServiceNow.” Sticking with it felt almost old-fashioned amid all the AI messaging, but, interestingly, quite a few people told us it was refreshing.
It’s not that I don’t think AI is fundamentally changing things – I do. I use AI every day in my work and life. It’s not that, as a company, we wouldn’t be a part of the movement – we most certainly are, and as a company, we also use AI every single day to improve our work.
Walking around K26, it was easy to feel as if enterprise AI were already fully operational everywhere. But the reality most companies face is very different. Outside the showcase examples, many organizations are still trying to solve foundational CMDB and governance problems.
And AI doesn’t work around bad data. It amplifies it.
In fact, Gartner analyst Roxane Edjlali recently predicted that through 2026, organizations will abandon 60% of AI projects unsupported by AI-ready data. (Gartner)
It’s about trust
We’ve talked about Improving Data Quality in ServiceNow ever since we created the first versions of Data Content Manager, but we should really be talking about making ServiceNow data trustworthy. That’s more than ensuring relationships and attributes in a database are correct.
While AI Control Tower will govern the AI agents in ServiceNow, who will govern the data the agents rely on, and how? That’s the question, and that’s where Data Content Manager fits the picture. It brings the structure and process you need to reliably and consistently manage your data quality on the platform.
Last year, I think it was Mark Bodman who mentioned he didn’t think he would ever see the day CMDB would be mentioned in a keynote speech. Well, not only that, but there was a CMDB song!
Now, in K26, CMDB was mentioned in most presentations as a necessary data source for AI Agents. And where does Control Tower store its configuration? The CMDB.
The Data Model continues to matter
Now, aligning with CSDM seems to be self-evident. At this point, CSDM alignment almost feels assumed. Everybody is expected to follow. And that’s great, everybody benefits.
However, in my experience, most companies are nowhere near aligned. There’s a lot of work to be done to achieve data quality high enough for functional AI to be realistically possible.
Again, this is what Data Content Manager does. It was built specifically to operationalize trustworthy data management at scale, whether that means CSDM alignment, governance, auditing, accountability, or engaging the right people across the organization.
Trustworthiness
Trustworthiness is about governance, accountability, and running data management as a scalable, consistent process rather than patchwork fixes here and there.
To achieve a trusted level of data, people need to be engaged in doing the work. No team can do it alone. A large organization might have hundreds or thousands of people who need to contribute but may not know how. They are, for example, business application owners who couldn't care less about data models or the CMDB, but whose input is still critical.
There’s no question about whether they need to be engaged; they do. The real challenge is how to engage them consistently, at scale. That’s exactly the kind of operational problem we built DCM to solve.
When you combine a systematic process, clear data requirements, automated auditing, CSDM compliance, and effortless engagement across the organization, you naturally increase the trustworthiness of your data.
This is likely to be the differentiator between AI success and AI failure.
The Landscape
In my years of running Data Content Manager, I’ve never seen so much attention being paid to data quality in ServiceNow. It’s great, we’ve seen the impact data quality improvement can bring to companies even without AI, let alone with it.
There is a lot happening in the data quality space. You only need to look at the ServiceNow store, and you’ll find all kinds of new entrants attacking the data quality problem from different angles. Also, ServiceNow itself is obviously working on the topic.
While there will be a lot more noise, the fundamentals remain. You need a process, and you need purpose-built tooling to support that process. The tool needs to be market-proven, have a solid foundation and team behind it, and evolve over time to meet new needs rather than locking you into expensive point solutions and then having to customize the customizations.
The more I listened at K26, the clearer it became that AI is not reducing the importance of operational discipline. It’s exposing where it never existed. The organizations that succeed with enterprise AI will not necessarily be the ones with the most AI. They’ll be the ones with the most trustworthy operational data.
