Your AI strategy is doomed if you’re still obsessing over unstructured data 

Australian businesses are still wasting millions on AI tools that promise to solve decades-old data problems.

Worse still, it means they’re missing the fundamental structural shift that could deliver an actual competitive advantage. The recent State of Martech 2025 report captures this perfectly: 108 pages of handwringing claiming “unstructured data” is what’s holding businesses back from AI success. It’s a persistent, convenient narrative, and it’s blocking businesses from becoming the real gamechangers AI can facilitate. 

The fixation on unstructured data isn’t new. For over a decade, we’ve been promised that the next wave of technology would finally help us wrangle emails, call transcripts, social posts, and customer feedback into submission. Like the persistent low hum of a fluorescent light in an office no one quite brings themselves to renovate, this problem murmurs on. The vendors, as ever, assure us they are on the brink of a breakthrough. 

The persistence of this narrative betrays something more revealing than the problem itself. It hints at a larger structural blind spot in how businesses are framing the role of AI. The real story isn’t that unstructured data is hard to manage – this has been true since the first business stored a customer comment in a CRM. The real story is that organisations remain stubbornly committed to solving this problem in vertical silos. 

Unstructured vs structured data

Consider how most AI initiatives begin: a marketing team invests in AI-powered customer analytics. A sales team experiments with conversational AI. Operations automates workflow management. Each effort begins in earnest, but rarely do these systems speak to one another. The AI becomes another tool in another stack, optimising local outputs while leaving the larger organisational machinery untouched. Data remains fragmented, insight remains partial, and leadership continues to mistake tactical automation for strategic transformation. 

Look inside any major Australian enterprise today and you’ll see this pattern repeated endlessly. Marketing builds martech stacks. Sales deploys conversational AI. Operations automates workflows. Each function protects its territory while the board ticks the “digital transformation” box. AI gets strapped onto each, like a modular accessory promising departmental productivity. But when AI is boxed into functional lanes, it inherits the very fragmentation that makes unstructured data unwieldy in the first place. 

The more consequential movement underway is less headline-worthy, perhaps because it resists productisation. Instead, it is a structural shift toward horizontal integration: AI across functions, disciplines and hierarchies as connective tissue for the organisation itself. 

This shift challenges the familiar managerial comfort zones. It demands CMOs stop thinking of AI as a marketing solution. It demands CIOs treat AI less as infrastructure to be provisioned and more as intelligence to be operationalised across the enterprise. It demands boards stop funding isolated proof-of-concepts and start asking how AI can unify the enterprise’s nervous system, bringing together customer signals, operational data and strategic intent into a coherent architecture. 

This is no small demand. Horizontal integration asks organisations to do the politically inconvenient work of dismantling silos, rethinking data governance and renegotiating ownership of what AI is allowed to touch. It asks leaders to admit that no single function owns the future and that meaningful progress requires collaboration across teams and mental models. 

David Brudenell, Decidr

Organisations will continue to pour investment into localised AI tools, optimising isolated parts of the machine while the machine itself remains disjointed. Vendors will continue to promise that the next release, the next API, the next fine-tuned model will finally unlock the value of unstructured data. And next year’s report will diagnose the same problem, once again. 

There are no shortcuts but there is a choice, and it is stark: continue treating AI as a departmental accessory that gets bolted onto existing structures while unstructured data remains fragmented; or recognise AI as a catalyst for organisational redesign that eliminates fragmentation. The very same fragmentation making unstructured data so problematic in the first place. 

Businesses can keep panning the same creek, hoping for different results. Or they can look upstream, where the real shift has already begun. 

Related Content

F.Y.I. Decidr partners with eBev to revolutionise hospitality supplier operations and sourcing with AI

F.Y.I. Decidr and Amazon Web Services enter AI partnership for SMEs