Research

Only 34% of Financial Services Content Is Connected, Iron Mountain Finds

Iron Mountain's new research found only 34% of financial services content is connected across systems, while 96% of institutions still prep data for AI by hand.

Iron Mountain highlights data gaps in financial AI
Iron Mountain highlights data gaps in financial AIAI-generated
By Elena Vasquez5 min read

Why it matters

  • Only 34% of financial services content is connected across systems, according to Iron Mountain research published 17 November 2026.
  • 96% of financial institutions need manual input to prepare unstructured data for AI workflows.
  • 89% of financial services and insurance firms said legacy enterprise content management costs directly limit AI and modernisation investment.
  • Iron Mountain's InSight DXP update achieved more than 96% extraction accuracy on complex financial instruments, rising above 99% with human review.
  • The Vanson Bourne survey covered 400 IT and data leaders, including 134 from financial services and insurance, across six countries.

Only 34% of financial services content is connected across systems. That single figure sits at the centre of new research released by Iron Mountain on 17 November 2026, and it frames a problem the industry has been quietly paying for years: banks and insurers are investing heavily in AI while their underlying information remains fragmented, unclassified and expensive to prepare.

The numbers get starker from there. Iron Mountain's study found that 96% of financial institutions need manual input to prepare unstructured data for AI workflows. Nearly all of them, in other words, still rely on people to sort, tag and organise content before any model can touch it. And 84% of respondents said highly skilled IT and data science staff are being pulled away from strategic work to handle routine data tasks — a quiet misallocation of some of the sector's most expensive talent.

Iron Mountain has a name for this burden. It calls it the "information tax" on AI adoption in financial services: the labour costs, delays and lost investment capacity that pile up when workers spend their time preparing disconnected content rather than using it directly in business processes.

What does the survey actually say?

The research, conducted by Vanson Bourne on Iron Mountain's behalf, covered 400 IT and data leaders across the US, UK, Germany, Brazil, India and Australia. Respondents came from banking and financial services, insurance, healthcare and life sciences, technology, manufacturing and the public sector. Within that group, the report analysed a dedicated subset of 134 financial services and insurance decision-makers. Every respondent worked at an organisation with 1,000 or more employees, so the findings describe large enterprises rather than the broader market.

The picture the data paints is one of genuine tension. Financial services and insurance organisations reported average gains 32% higher in IT efficiency and 21% higher in AI readiness than their cross-industry peers. These are not technology laggards. On several measures, the sector outperforms everyone else.

Yet the same firms reported that older content systems are actively holding them back. A striking 89% said the cost of legacy enterprise content management systems directly limits their ability to invest in AI and modernisation. The money going to maintain ageing repositories is money not going to new capability. And 45% of respondents identified data classification and tagging as their primary manual burden — the single biggest choke point in the data supply chain.

Put plainly: firms are improving technology performance while still struggling with older content systems and disconnected information spread across multiple platforms.

Why does this matter for banks and insurers?

The stakes here are structural. Banks and insurers are among the biggest investors in AI anywhere in the economy, but they also operate in heavily regulated environments and sit on enormous stores of documents, records and unstructured information. That material is difficult to classify and connect at the best of times, and regulation makes getting it wrong expensive.

The consequence is operational drag. When firms try to deploy AI tools across lending, compliance, servicing and back-office functions, fragmented information forces staff to step in — organising, tagging and validating material before systems can use it. The AI layer may be fast, but the human preparation layer underneath it is not. That gap between model capability and data readiness is where the "information tax" gets paid, every day, in salary hours and delayed projects.

Varun Gupta, senior vice president of global product and solutions marketing at Iron Mountain, put the company's position bluntly in a statement released alongside the research.

"Financial services firms are investing heavily in AI, but paying a steep operational tax to feed those models trusted, reliable information. Connecting that information, preserving its business context and embedding governance into workflows are essential to scaling AI with confidence. To unlock AI's full potential, financial institutions need a unified content intelligence layer that automates governance and structures dark data at ingestion so models and agents can act on reliable data. In today's AI era, your underlying data foundation is just as critical as the AI technology itself," Gupta said.

The last sentence is the thesis of the whole release, and of the sector's current predicament: the quality of an AI deployment in finance may now depend less on which model a bank buys than on the state of the content underneath it.

What is Iron Mountain shipping alongside the research?

Unsurprisingly for a vendor-commissioned study, the research arrives paired with a product. Iron Mountain has introduced a new release of InSight DXP, its content platform for managing information and workflows, and it is pitching the update directly at the problems its own survey documents.

The update is intended to connect business systems more closely with enterprise information and reduce the work involved in turning fragmented content into structured data for AI use. The release focuses on automated content intelligence and governance. The software is designed to preserve business context around information and strengthen oversight across workflows — a meaningful claim in a regulated sector where knowing where a document came from, and who approved it, is not optional.

Iron Mountain has tied the release to document-heavy processes such as trade operations, mortgages and commercial lending, and it has published accuracy figures to back the pitch. The company said InSight DXP has achieved more than 96% extraction accuracy on complex financial instruments, and more than 99% accuracy when human review is included. Those numbers matter because extraction accuracy on complex instruments — the dense contractual material that dominates trade finance and lending — has historically been the hardest problem in financial document processing.

The timing is deliberate. The latest release of InSight DXP is scheduled to be available worldwide this month, putting the product in market at the same moment the research lands.

What comes next?

The broader takeaway for financial institutions is that the AI conversation is shifting downstream — from models to data plumbing. Iron Mountain's survey suggests the sector's AI readiness advantage is real, with that 21% edge over cross-industry peers, but so is the constraint, with 89% of firms naming legacy content costs as a direct brake on investment. The vendors that win the next phase of enterprise AI in finance may be the ones that solve the 34% connectivity problem rather than the ones selling another model. If the "information tax" shrinks, it will be because content gets structured at ingestion, governed automatically, and connected before anyone — human or machine — has to ask where it lives.

Source: itbrief.com.au

Share this article:

More from Elena Vasquez

Elena Vasquez

Show full bio

Market editor covering technology and product news in artificial intelligence and machine learning at AI In Context.

708 articles

Related articles

  1. Iron Mountain: Only 34% of Financial Services Data Is Connected
  2. Iron Mountain says only 34% of financial data is AI-ready
  3. Novo launches AI banking tools with approval-based controls
  4. Connected financial data matters more than the AI model
  5. Only 8% of Australian Firms Have Data Ready for AI at Scale

« Previous articleNext article »