Come Together: Unlocking the Power of Integrated Data with AI

I'm going to edit a sacred string of Beatles' lyrics for my newsletter opener: "Come Together, right now, all for me."  In this case, I am referring to what I constantly hear as a challenge for companies to pursue personalization -- the fragmentation of their data.  Data locked up in different, often legacy, databases, tied to marketing, sales, product use, billing, customer service, ecommerce, etc. It's all over the place, and getting it to "come together" seems daunting. 

Yet, from what I have seen in many AI tools now, one of the core capabilities they enable is writing the code and bringing together disparate data. Could the walls that divide up data about a customer be melting? Here's a few examples:

• Verusen combines several AI capabilities to enable supply chain and production managers to optimize their ongoing spend on MRO (maintenance, repair and operations) materials. One of the keys to doing so is combining all of the scattered data around the company on inventories and consumption of MRO materials across production lines, warehouses, and factories.  They also use AI to combine internal information with price and availability data they scrape from provider sites.
• Genesys provides new transparency of a customer's journey across all their different touchpoints (through their purchase of Pointillist). The foundation is connecting all the instances of how a customer touched the company, across marketing, apps, product use, etc., time-stamping it, and then turning that into data that can spot anomalous problems in people's journeys, or even just bring that data up to a call center rep so they know what problems someone was having when they called in
• Narrative's Rosetta uses AI to understand the different schemas of a cluster of different databases, even from more than one company, and writes the code to integrate the data sources, normalize the data in common formats, and highlight data that needs to be cleaned. Multiple companies, including The Trade Desk, are using this to create "clean rooms" where companies can share data in a managed, privacy, and security-compliant way

The more I dive into newer AI tools, especially those focused on unlocking the potential of specific use cases, usually tied to a sector or a functional process, most of them start from the premise that their value comes from bringing data together. Then, the larger pool of data becomes the base for new analytics, spotting triggers, trends, and targets to act upon that could never have been possible when the data was balkanized. 

There are large numbers of outsourcing firms, consultancies, and pools of data engineers all pointed at integrating data.  It often takes months and unusually large budgets. Is that always necessary now? The tools I cite are not for every application, but there are more capabilities now out there than most IT teams and executives realize. I've seen companies using these tools realize breakthroughs that spending vast sums on data integration could never unlock. 

The art of the possible is greatly expanding.  Yes, the promises of GenAI for creation, providing answers, and now, with agents, executing operations are real.  But let's also acknowledge that some of the more mundane areas, such as data integration, standardization, and hygiene, are now also transforming, providing large and small companies with fewer excuses based on the state of their data. Yes, basic processes and instrumentation to capture data in the first place is a huge gap for many companies, but bringing one's data together is now within reach.

AI can help companies unlock their own data to power AI. What could you unlock by bringing the data that you do collect together? What would change in your operations? Where is it worth trying specific tools to unlock the potential?  I am very curious to hear more from leaders experimenting with these capabilities and how they see the pros and cons involved.