The Data Foundation Problem No One Talks About Before Buying AI

Every major workplace and real estate technology vendor now has an AI story. Platforms that spent years competing on dashboards and integrations are repositioning around intelligence, automation, and insight. Buyers are being asked to evaluate AI-powered tools, allocate budget for AI-powered tools, and in some cases build the business case for AI-powered tools, often at the same time.

Almost nobody is asking the question that matters most before any of that: is our data actually ready for it?

AI doesn't fix bad data. It amplifies it.

There's a seductive assumption built into how AI is being marketed right now: that if you layer enough intelligence on top of a data problem, the problem disappears. It doesn't.

"AI isn't magically going to solve those foundations," says Nate Colle, Head of Professional Services and Operations at Metrikus. "If that underlying data is inconsistent, if it's disconnected, it's disjointed, if it's incorrect, AI is really just going to amplify those problems rather than fix anything."

This is the risk most organisations aren't discussing. Fragmented, inconsistent, or untrustworthy data doesn't become reliable when AI processes it faster. It becomes unreliable faster. The wrong answer surfaces more quickly, with more confidence, dressed up in the language of intelligence.

The pattern is consistent across AI deployments in data-heavy industries. Gartner research suggests that through 2025, 80% of AI projects will remain alchemy, run by wizards whose methods are not understood or explainable, with poor data quality cited as one of the primary reasons AI investments fail to deliver. For organisations making significant technology investments, this is not a theoretical risk. It's a predictable outcome for those that haven't addressed the foundation first.

"If that underlying data is inconsistent, if it's disconnected, it's disjointed, if it's incorrect, AI is really just going to amplify those problems rather than fix anything." 
- Nate Colle, Head of Professional Services and Operations, Metrikus

The trust problem that kills adoption quietly

When a system produces answers people don't trust, they stop using it. This is rarely announced. It happens gradually: a leader questions a figure, a team finds a discrepancy, someone pulls the original spreadsheet to cross-check. Within weeks, the AI-powered tool becomes something the organisation invested in but doesn't rely on.

The silent killer of AI adoption is not technical complexity. It's the erosion of trust that follows when the outputs are inconsistent.

Trust in building intelligence has a specific meaning in practice. It requires connected systems describing the same building in the same language. It requires occupancy measured consistently across floors and sites. It requires data that reconciles. Not four slightly different answers to the same question depending on which platform you ask.

McKinsey research consistently finds that organisations with strong data foundations are significantly more likely to report above-average financial performance — and that the gap between data leaders and laggards is widening, not closing. The organisations getting the most value from AI are overwhelmingly those that invested in clean, connected data infrastructure before the AI conversation started.

What good foundations actually look like

The organisations that are AI-ready share a few characteristics that have less to do with their technology vendors and more to do with the decisions they made before they started evaluating technology.

Their building data comes from connected systems, not parallel ones. Space names are consistent across platforms. Occupancy is measured the same way across sites. There is a single source of truth for the estate. Not six different dashboards maintained by six different teams, each of which tells a slightly different story.

When that foundation is in place, AI can do something genuinely useful: it can help leaders test what the future could look like before they commit to it.

"The future of building intelligence isn't just about reporting," says Colle. "It's about helping organisations test what the future could look like." Consolidating from five floors to three. Modelling the impact of a three-day mandate on peak capacity. Understanding how a change in booking policy would affect availability across a twenty-building estate. These decisions can be explored from data that already exists, but only if that data can be trusted.

"The future of building intelligence isn't just about reporting. It's about helping organisations test what the future could look like." 
- Nate Colle, Head of Professional Services and Operations, Metrikus

The gap between organisations is widening now

Here is what the next three years look like for organisations that have their data foundations in place: AI surfaces insights continuously, without anyone having to go looking for them. Patterns emerge across thirty or fifty buildings that no analyst could have identified manually. Similar buildings behaving in very different ways. Occupancy trends shifting by region before they show up in a report. Space layouts that consistently outperform others. Leaders make estate decisions with confidence because they trust the data underpinning them.

For organisations without those foundations, the picture is different. Teams spend time validating information rather than acting on it. Conflicting dashboards get reconciled by hand. AI outputs get questioned and cross-checked rather than used. The investment in AI technology doesn't deliver. Not because the technology is wrong, but because the infrastructure underneath it isn't ready.

"It's no longer going to be about who has the most data," says Colle. "The difference is who has the data they can trust. Because AI is not going to be a competitive advantage in itself. The trusted data is. AI is simply going to allow organisations to realise that advantage much faster."

Most organisations already have more data than they can process. That has been true for years. The question AI makes urgent is not how much data an organisation has. It's how much of that data they can actually rely on.

"It's no longer going to be about who has the most data. The difference is who has the data they can trust. AI is not going to be a competitive advantage in itself. The trusted data is." 
- Nate Colle, Head of Professional Services and Operations, Metrikus

What the most capable organisations will look like in five years

The organisations that get this right will look meaningfully different from those that don't. AI will sit on top of trusted, real-time intelligence across the entire estate. Insights will surface continuously. Emerging patterns will be flagged before they become operational problems. And the leaders responsible for the workplace, not just the analysts and technical specialists, will have direct access to the intelligence they need to make decisions faster.

"Organisations that succeed won't have more data," says Colle. "They'll be the ones that have turned the data they have into faster decisions."

That outcome is attainable. But it starts with a decision that has nothing to do with which AI platform to buy. It starts with getting the foundation right: unified, connected, and trusted across the estate.

The organisations asking that question now are the ones who will be positioned to move fastest when AI delivers on its promise. The ones that skip it will find out why it matters, eventually, at significantly greater cost.

See how Metrikus gives you a trusted data foundation across your entire estate. Talk to our team