What a misconfigured CMS and a frontier AI model tell us about the data gaps quietly sitting inside your revenue stack.
In late March 2026, Anthropic, the AI safety company behind the Claude model family, made headlines for all the wrong reasons. A configuration error in its content management system left approximately 3,000 unpublished internal assets in a publicly accessible data store. Among them: a draft blog post describing a new model, Claude Mythos, which Anthropic itself describes as a step change in AI capability and the most powerful system it has ever built.
The model, internally codenamed Capybara, would sit above Anthropic’s existing Opus tier. The leaked documents described it as “currently far ahead of any other AI model in cyber capabilities” and warned that it could help attackers outpace defenders.
Anthropic confirmed the leak was real. The cause: human error in a CMS configuration. Not a sophisticated breach. Not a targeted attack. A misconfigured data store.
If that’s a gap in a frontier AI lab, it’s almost certainly a gap somewhere in your revenue stack too.
This Is Not an Anthropic Story. It’s a Data Governance Story.
It is tempting to read the Mythos leak as a curiosity; an embarrassing moment for a high-profile AI company, quickly contained. But the underlying failure is one that plays out quietly, and repeatedly, inside mid-market B2B businesses every day.
The gap is not between having data and not having it. Most businesses have more data than they know what to do with. The gap is between data that exists and data that is understood: where it lives, who can access it, what state it is in, and what the consequences are when something breaks.
In Anthropic’s case, the consequence was reputational. For your business, the consequences are typically commercial: missed revenue, inaccurate forecasting, broken customer journeys, and marketing spend that cannot be properly attributed or optimised.
The Revenue Stack Has the Same Vulnerabilities
At Centralise, we work with mid-market B2B businesses to build what we call revenue infrastructure: the systems, data, and processes that connect marketing activity to pipeline, pipeline to revenue, and revenue to growth. In that work, we encounter the same categories of failure, again and again.
Data that was configured correctly at point of setup but has drifted as the business scaled. Permissions and ownership that were never formally defined. CRM records that contain contact data no one has audited in two years. Automation workflows that fire based on fields whose values can no longer be trusted. Attribution models that assume clean data, sitting on top of data that is anything but.
None of this is the result of negligence. It is the result of growth: new tools added without decommissioning old ones, new team members inheriting configurations they did not build, and a pace of change that consistently outstrips the capacity to document and govern.
The result is a revenue stack that functions... until it doesn’t. And when it breaks, it rarely announces itself cleanly. It surfaces as unexplained churn, inconsistent reporting, or a campaign that performed well by one measure and poorly by three others.
What Data Enablement Actually Means
Data enablement is a phrase that risks becoming jargon. At its core, it means something specific: building the foundations that allow your data to be trusted, acted on, and connected across the business.
For most of our clients, that involves four things:
- Data structure and taxonomy: ensuring that the objects, properties, and relationships in your CRM reflect how your business actually operates, not how it operated when HubSpot was first configured.
- Data quality and hygiene: identifying and resolving the records, fields, and values that are duplicated, incomplete, or inconsistent, and building the processes to prevent recurrence.
- Governance and ownership: defining who is responsible for what, how changes are documented, and what the standards are for data entering and moving through the system.
- Reporting maturity: building the infrastructure that connects activity data to outcome data, so that decisions can be made on the basis of what is true, not what is assumed.
HubSpot is the platform through which we deliver this. As an Elite HubSpot Partner, we work across the full suite, CRM, Marketing Hub, Sales Hub, Operations Hub, with the specific goal of making the platform a reliable source of commercial truth, not simply a tool that stores data and sends emails.
The Parallel with AI Capability
There is a broader point here that the Mythos story illustrates well. The AI capability conversation has accelerated dramatically. Businesses are being asked to consider AI-powered forecasting, AI-driven personalisation, AI-assisted prospecting. The models available are, genuinely, remarkable.
But AI does not improve bad data. It operationalises it. An AI model trained on or operating against a CRM with structural inconsistencies, incomplete records, and undefined ownership will produce outputs that reflect those problems at scale: faster, and with greater confidence than a human analyst would.
The businesses best placed to benefit from the current generation of AI tools are not necessarily those with the largest budgets or the most sophisticated technology stacks. They are the ones that have done the foundational work: clean data, governed systems, reliable reporting.
You don’t need Claude Mythos to benefit from AI. You need data that is ready for it.
The question most businesses are asking right now is “which AI tool should we be using?” It is already the wrong question. The more useful one is: what does our data, our operations, and our revenue infrastructure look like in a world where AI capability continues to compound at this pace?
The businesses best placed for that wave are not necessarily the ones using the most advanced models today. They are the ones building the foundations, clean data, reliable systems, clear ownership, that allow them to move decisively when the next step change arrives. Because it will. Mythos will not be the last model described as unprecedented.
You do not outrun the tide. You build for it.
Join Us: Data Enablement in Practice
On 23 April 2026, Centralise is hosting an in-person event on Data Enablement in Practice: AI, Automation & Reporting Maturity.
The event is designed for marketing, revenue, and operations leaders at mid-market B2B businesses who are working through exactly these questions. How do we get our data into a state we can trust? How do we build reporting that reflects commercial reality? How do we lay the foundations that make AI a genuine advantage rather than another system sitting on top of broken infrastructure?
The format is practical and applied. Centralise practitioners and external speakers will work through real use cases across the Centralise, Optimise, and Scale stages of revenue infrastructure maturity. There will be time for structured discussion and honest conversation about where businesses actually are, not where they aspire to be.
The Claude Mythos story will pass. The data governance challenges it illuminates will not.
If the questions it raises are ones your business is grappling with, we would welcome your attendance.
Register your place at the April event → centralise.co.uk/events
Centralise is an Elite HubSpot Partner and revenue infrastructure consultancy for mid-market B2B businesses.
