Is AI overload leaving your organization drowning in insight?

AI is a boon for data analysis. Professionals can automate routine tasks, such as data processing and anomaly detection, while complicated mathematical equations can run in almost real-time. As a result, they have access to information faster than ever before, and the traditional decision-making cycle has been reduced from weeks to seconds with AI-driven insights, leading Gartner to predict that 50% of business decisions will be augmented or automated by AI agents by 2027.

Such access is great, but it also creates expectations about the pace of decision-making, and increases the risk of overlooking key things along the way. It’s something that’s crossed Satpal Chana’s mind.

“All this technology puts insight into the decision makers’ hands way sooner than they’ve ever had it before, which leads to an interesting discussion because now, they feel like they need to think about things,” says the deputy director for data analytics at Visit Britain, an organization that uses Databricks tech and AI to analyze sentiment, travel trends, and UK tourism. “There’s a lot of information, but is there too much of it?”

Jakob Rissmann, principal product owner of data services at transport specialist FlixBus, is another digital leader who recognizes the scale of the issue. He refers to a recent conversation with one of his firm’s data scientists, who talked about the challenges of trusting outputs, before posing a rhetorical question about the exploitation of AI.

“He asked if we should do it less and I thought about it, and replied, ‘I never thought about that concept before,’” he says. “Usually the direction is the other way around, and it’s all about how we can use AI more.”

Regardless of quantity of use, evidence so far certainly points to ever-growing investments in AI. Gartner also recently predicted that global gen AI spend is expected to increase over 76% this year to reach $644 billion, indicative of an imminent time when AI becomes even more integral to business operations and consumer products than it already is.

However, a decision to further invest in technology is no guarantee of success. While Carruthers and Jackson’s annual Data Maturity Index found that 68% of digital leaders believe the technology in their organizations either overwhelmingly or mostly supported data use, close to a third of CDOs, in contrast, said it hindered data use.

The consultancy’s CEO Caroline Carruthers says investing in AI is simply a starting point. “Most business leaders I talk to can see the connection between AI and data,” she says. “They’re excited by these developments, but they also recognize their business needs to feed these technologies the right stuff. There’s this dawning recognition that other issues such as data governance and management are important.”

Setting expectations for AI exploitation

One of the most important things to get right is education. While companies continue to invest in AI, many of the most popular tools, like ChatGPT and Gemini, can be tested for free by employees at home.

Amit Patel, CDO for wholesale banking at Truist, who spoke on a media panel at the recent Snowflake Summit in San Francisco, says digital leaders are likely to encounter users who’ve witnessed the rapid pace of change outside work and feel frustrated about the slower pace of enterprise transformation.

“The ease of being able to use these large language models in your personal life has influenced the perception of how quickly people think they should be able to deploy models in a business setting,” he said at the event. “I think there’s an education process that says you can’t just turn AI on and point it at a database or an application and generate answers tomorrow.”

Patel says he regularly encounters employees who expect AI tools to be accessible and available. However, IT departments exist for a reason: to ensure new technologies are adopted safely and securely. He adds that CIOs must work with business partners to establish expectations and ensure employees understand that any use of AI tools is compliant with business policies and industry rules and regulations.

FlixBus’ Rissmann also recognizes the challenges of implementing new technologies in an enterprise setting, and the importance of education. He suggests introducing AI into a business should involve a process that’s well-rehearsed with every innovation.

“First, it’s overwhelming, and then we, as a society, learn how to use it the right way, but also know where it can be dangerous,” he says. “That process takes place through education, by using these technologies and understanding them. People should make their own informed decisions about where they want to use AI and where they don’t.”

The message that emerges from discussions with digital leaders is that exploiting insight that’s generated through AI involves nuance. Employees are eager to use tools that can generate quick answers to intractable questions. But the role of CIOs in this balancing act is to help ensure AI-generated insights are applied in the most appropriate places to create maximum business value.

That’s an approach being taken by Antony Hausdoerfer, group CIO at UK’s auto breakdown specialist The AA. He’s driving a digital transformation program, using key partners such as Ericsson and the company’s car-health assistant Vixa, to deliver data-enabled services. Hausdoerfer’s experiences lead him to suggest CIOs must take a targeted approach.                                                                               

“I don’t think you should fear AI, because the technology is probably going to give you more insight than you’ve ever had before,” he says. “But success is related to how you take that information and turn it into a meaningful decision, so you’re coherent. That process is about how data contributes to that kind of decision-making, as opposed to just constantly going from one thing to the next.”

Taking a strategic attitude to insight

This targeted approach will be the key to success for CIOs who want to help their businesses turn insight into groundbreaking decisions. HPE global CIO Rom Kosla, like other digital leaders, understands the scale of that challenge. He refers to AI as an equalizer, as employees in all firms now have easy access to hugely powerful tools.

“Knowledge is going to be everywhere, and it’s going to be super-commoditized,” he says. “But decision rights still need to be maintained. I do think that even if we have agentic AI systems now, there still needs to be decision rights around what they can and can’t do. I don’t think it’s clear yet how those guardrails are defined.”

Some organizations are grappling with it all. But Mikko Jarva, head of portfolio and architecture at telecoms giant Nokia, is making strides by helping to develop Network APIs that bring mobile network capabilities to pioneering transport areas such as self-driving cars, robotaxis, and drone flights.

Jarva and his team have combined the Snowflake AI Data Cloud with Nokia’s Network as Code platform to create an optimal route path algorithm that predicts relevant network KPIs, ensuring autonomous vehicles have high-quality connectivity. He’s eager to set the guardrails to investigate how AI can help further improve operational processes.

“We’re exploring the use of generative tools connected to our data,” he says. “We could use that technology to help ask business-level questions, and then get the right insights to see what changes need to be optimized. So we see the AI tools as an interface to make data access simple, and more understandable and comprehensible.”

Making data work for you

One of the key things for CIOs to prioritize as they explore their options, suggests Vivek Bharadwaj, CIO at clothing manufacturer Happy Socks, is a process to find a needle in the haystack of information that bombards modern businesses. His guiding mantra is don’t do data for data’s sake. He advises other business leaders to be clear about their objectives.

“You need to think about how data leads to the insights you’re creating, and what types of insights lead to a particular action you want to take,” he says. “Then success is about moving from data to insights, and insights to action, where the action drives the kind of data you collect and analyze, and how you steer the strategy.”

Richard Masters, VP of data and AI at Virgin Atlantic, is another digital leader focused on building a data strategy. Virgin uses Databricks’ Unity Catalog to bring disparate sources together to create a centralized location for the information that powers the firm’s AI-enabled decision-making. He recognizes the importance of setting employee expectations.

“The biggest challenge we’re seeing is in some areas, you can get the information quickly, but you still need to model that data and put the semantics on top to establish the content. Some data might require a bit more work before it’s ready.”

Like other digital leaders, Masters says employees must be educated about the right balance between exploiting AI and respecting governance frameworks. At Virgin, people already boost operational processes and hone customer services with data insights. Now the priority is to help employees do even more with information.

“Data analysts have a wealth of things they can access now with the right governance and processes,” he says. “They can start to be curious while their managers look at what they’re going to do next. Effective AI-enabled decision-making creates this great cycle of curiosity across the teams.”