Data was once the “new oil”. But artificial intelligence, and above all generative AI, look to be eclipsing data analysis projects for enterprises.
Gartner, for example, predicts that spending on AI will reach a staggering US$2.52tn this year. That figure includes infrastructure, such as data centres, and spending by hyperscalers and AI vendors, not just enterprises using AI tools. Nonetheless, it represents a 44% increase in 2025, according to the analyst firm.
This dwarfs investment in business intelligence. Analyst firms put the size of the global BI market at around $40bn annually, and the wider data and analytics market at $175bn. A sizable sum, but nowhere near the spending expected on AI.
But much AI spending leans heavily on prior investments in data and analytics. Enterprises will need to spend more and invest more time in their data assets if they are to make the most of AI and avoid its pitfalls.
Solid foundations
Business intelligence and AI share more than the need for good-quality data. Predictive analytics, machine learning, and even robotic process automation (RPA) systems have a lot in common with today’s AI tools, even if they lack user-friendly chatbot interfaces.
As Carlie Idoine, VP analyst, told an audience at Gartner’s recent Data and Analytics summit, AI is not going to replace BI. Instead, it will extend its capabilities. “I don't think AI is going to kill BI,” she said. “AI and BI are going to learn to work together. It’s not AI or BI; they are better together.”
As Idoine puts it, business intelligence is not a tool, but a discipline. The fundamentals of business intelligence, including metrics, data lineage, observability, and governance, are all equally important for AI projects.
Many of the lessons learned around BI and advanced analytics are now being relearned for AI. Data quality is as essential for predictive analytics and machine learning as it is for LLMs.
Poor quality and fragmented data continue to hold back enterprise decision-making. But companies that invest in the foundations of good data science see a clear return on that investment. Research, again from Gartner, found that poor quality data costs businesses on average $12.9m a year. That figure is from 2020, so it does not factor in the cost of poor decisions or reputational damage caused by AI using poor data.
“BI and AI both depend on trusted, connected, and well-governed data. Whether organisations are pulling reports on cashflow, forecasting demand through BI tools or using AI agents to automate and execute decisions, they need confidence in the quality, lineage, and context of the data powering those decisions,” Levent Ergin, chief strategist for agentic AI, regulatory compliance, and sustainability at Informatica by Salesforce, told IT Pro.
“A single, trusted data foundation helps businesses deliver more consistent reporting, faster insights, and better decision-making across the enterprise. Without trusted context around the data, organisations risk fragmented analytics, conflicting reports, and reduced confidence in business intelligence insights.”
Questions of authority
But business intelligence and data analytics are about more than tidying up data sources.
BI provides what Gartner’s Idoine describes as “analytic authority” and governance to organisations. Done well, BI brings together enterprises’ systems of record in a single version of the truth, and predictive analytics helps businesses to make decisions around that data source.
AI will not replace that, or at least, not soon.
“We have millions of people every day that are in there using their analytics to create their dashboards to do their queries,” Gartner’s Idoine says. “You can't change that kind of thing overnight. You might be able to put in an LLM, and somebody can ask a question, but that's not going to run the reports that we really run our business on. It doesn't quite get us there.”
Over time, BI and artificial intelligence could converge. Already, organisations are making use of LLMs’ ability to work with natural language so employees can interrogate data without the need to resort to code.
Drawing insights and making projections from data warehouses and data lakes remains a specialist task, even with the current generation of highly sophisticated BI tools.
“When we’re looking at the future, users are really looking to solve complex questions,” Julian Sun, VP and team manager at Gartner, told the summit. “They are looking for better decision-making processes, where AI can use all the context saved all in different places, which will eventually give you these context-rich insights. We no longer need a separate BI layer; it’s one platform.”
Nonetheless, Sun agrees that BI and AI will continue to work together, not least because BI delivers consistent results. AI is notoriously non-deterministic, and so could give different answers to the same question.
But AI has the advantage when it comes to creating a personalized experience, and drawing on domain knowledge across the business, Sun says. He points to areas such as supply chain analytics, vital to day-to-day operations, but which previously needed expert analysts to process that information.
Joint forces
Industry vendors with a foot in both camps agree. “There is naturally overlap between analytics and AI, but I see analytics as the groundwork rather than a competing capability,” Ash Gawthorp, co-founder and CTO at Ten10, an automation and data consultancy, told IT Pro.
“Most organizations are not short of data, but they often struggle to trust it, connect it across systems or act on it quickly enough to make a meaningful difference,” he adds. He sees enterprises moving away from BI as a narrow reporting tool and towards using data to improve their day-to-day business processes. This is also an area where AI tools can help.
“Although generative AI attracts much of the attention, traditional machine learning continues to deliver meaningful value across areas such as forecasting, matching, optimisation and classification because the outcomes are often measurable and well understood,” Chris Riche-Webber, VP of business intelligence and analytics at SmartRecruiters, an SAP company, told IT Pro.
“There is naturally a significant amount of overlap between analytics and AI, although I think the relationship is often misunderstood. AI is not a replacement for analytics. If anything, it raises the stakes because organisations are dealing with more complexity, more signals and more information than ever before,” he says.
If enterprises are to exploit AI’s potential, then they should start with both effective analytics and solid data.
