In March 2024, I wrote that without good data, generative AI systems will be as useless as a warehouse fire. As I teach my generative AI architect students, AI, and especially generative AI, is ultimately a data-oriented problem. If your data game is weak, so will your AI solutions.
To be effective, AI systems need gigabytes of clear, accurate data. They find patterns in the data and respond to user requests for what those patterns might mean and how to leverage the insights for strategic business purposes. Poor data hygiene, accuracy, or dysfunction can lead to AI systems making inaccurate inferences or giving answers that you don't realize are wrong until it's too late.
Well, at least AI is cheap to build and run… oh wait.
We're not yet ready for AI
A recent Enterprise Strategy Group report found that a survey of 800 IT decision makers revealed that more than three in five organizations have notable gaps in AI readiness, particularly in their infrastructure and data ecosystems.
While organizations are optimistic about AI's long-term potential, many still need to prepare for mass adoption. More than half of AI decision makers are concerned about their IT teams' ability to keep up with the rapid innovation that generative AI will drive. Businesses need to tackle a range of tasks, including strengthening data processes and hardening infrastructure.
Companies vary widely in how they measure the success of their generative AI projects. About two in five companies monitor progress based on qualitative impact analysis, accuracy of AI responses, or quantitative benefits to users and processes. Slightly fewer, about 38%, of organizations cite cost savings as their primary success metric. As AI initiatives demand a larger share of corporate budgets, the push for ROI is expected to intensify.
Overall, the report highlights that while businesses are eager to harness the potential of generative AI, significant infrastructure and data management groundwork is required to realise its benefits and ensure sustainable, long-term success.
The CIO's To-Do List from Hell
Long before AI had a major impact on the market, most companies knew they had a problem with data. In fact, most companies have avoided investing in AI and business intelligence due to a lack of trust in the data. No one in the company fully understands where the data is or what it means. Silo leaders own and manage the data, so there is no single source of truth for simple things like who a customer is or where customer data comes from. Redundancies are common in sales, production tracking, and other areas where data is not managed properly.
How did things get so bad? Most companies spent years focusing on new and exciting objects like ERP and CRM systems that contained critical data, but that data was locked away in proprietary data stores. After ERP and CRM came data warehouses, distributed systems, data integration, and now the cloud. Through it all, data has become more complex, distributed, heterogeneous, and lacking centralized management. Too many companies don't understand metadata and can't properly track data through business processes. Also, acquisitions have created data redundancy, and many companies are still running old systems that came with the businesses they acquired. Now we're facing AI, where the meaning, structure, and veracity of data are no longer optional.
For AI to deliver value, CIOs will need to fix the data, which may be too costly or risky for many companies. Hindering AI adoption due to data issues could doom some companies as competitors embrace AI as a key force for innovation. Ultimately, there will be haves and have-nots.
Bridging the gap
Take note, companies: if your data isn't ready, avoid AI. Many of the failed AI projects I've been consulted on stem from a poor data ecosystem and lack of willingness to fix it. Some companies think AI can fix their data, but that's never the case.
My advice is that getting your data in better shape has benefits beyond just AI readiness, and it's worth investing time and money into it. It seems like AI investments are being spent on fixing past mistakes, and many CIOs are happy to make it the next leadership's problem. Fixing years of neglect and lack of a data strategy without easily traceable direct business benefits requires asking leadership and the board for funding. Most CIOs don't have this conversation.
My advice would be to fix the data, whether you leverage AI or not. Maybe AI will be the incentive for companies to finally get their data into a more functional state. I'm okay with that.
