In the last three years we've experienced the rapid expansion of AI as coding partner, researcher, and commander of agentic armies that can be deployed in customer support and operations.  The speed of change and associated fears, is orders of magnitude greater than move to cloud just a few years ago.  It's exciting, but the discernment and judgment we’ve always applied to enterprise tools is still required. Do features align with business operations? Does it make sense for the scale of the organization?  What about supportability and cost of the solution? Security?  Given the quantity of AI start ups that have emerged and shuttered, the stability of the vendor and longevity of the technology underpinning the solution, has to be considered.

Experimentation is critical. Willingness to reimagine how processes are solved is the only way to prevent a crisis of technical debt and for businesses to manage costs and operations. But experiments that aren’t refined and operationalized create their own technical debt.

Navigating the current revolution requires imagination and curiosity. It also demands the same systematic approach to tool selection and implementation which enterprises have always applied. A new tool has to be better than the thing it replaces. It’s always impossible to automate a process which is not fully understood.