Who Brings AI into the Working Day?

When I wrote The AI Came Later in September, the first step was to help the company see itself. Star Group’s reporting system is now built, and staff from security guards to property managers are beginning to use it. Management can see what is happening across different properties as it happens, allowing faster decisions and follow-up. Across other businesses, inventory and occupancy are becoming clearer, with records updated daily. Whether these changes improve customer and resident satisfaction over the long term will take time to establish. But as the system enters daily use, something else is becoming clear: tools can be built quickly, but making them useful requires someone who understands how the work happens and why people would choose to use them.

Recent conversations with specialists across the AI field have returned to a similar observation: every company has its own culture. The same features can produce very different results in different organisations. Technology companies have a role called the Forward Deployed Engineer, someone who works closely with a customer’s business and turns practical problems into working technical solutions. Companies need a similar capability internally: someone who knows which processes need improving, can break down unclear requests, use AI to build tools, and help colleagues test and refine them. Much of the work over the past few months has involved exactly this. Who submits a report, who receives it, and what requires follow-up may sound like small decisions. Each requires an understanding of the business and its people. AI can accelerate development, but someone still has to decide what is worth building.

When some internal tools can reach a testable version within days, the decision to subscribe to software begins to change. We are now testing tools built internally to see whether they can replace some existing subscriptions. The savings may be modest. The greater value is in workflows that fit the business, connected information, and the ability to keep adjusting the system as people use it. Established software still offers reliability, security and support, while building internally brings maintenance responsibilities. But companies now have another option, and evaluating it takes both business understanding and the ability to put technology to work. In the long run, every management team may need someone responsible for this: identifying problems worth solving, bringing AI and colleagues together to solve them, and taking responsibility for the results. The title may vary. Someone needs to do the work.