When AI underdelivers, look at the environment it landed in.
Most teams have more tools than they did two years ago, not fewer. When the AI sitting on top of them disappoints, the technology usually takes the blame.
The tools multiplied
Over the past two years, organizations have invested heavily in new tools, platforms and AI applications. Each one promises greater efficiency, better insights or faster decisions. Yet for many teams the day-to-day reality feels increasingly fragmented.
Work is spread across multiple systems, with information stored in different places and processes moving between applications that were never designed to work together. As the number of tools grows, so does the effort required to coordinate them.
Then the technology takes the blame
When AI initiatives fail to deliver the expected outcomes, the technology is often blamed. In my experience, the underlying challenge usually lies elsewhere.
AI performs best when it becomes part of the way a business operates. The value it creates is closely connected to the environment it is deployed into. Workflows, reporting, execution, knowledge and decision-making all need to function together before AI can become part of everyday operations.
The value AI creates is bounded by the environment it lands in.
What connected actually looks like
The most consistent theme in our conversations since launching BOS has been frustration with fragmentation. These are not organizations that under-invested. They have a CRM, a project tool, document repositories, communication platforms, AI writing tools and a growing list of applications supporting different parts of the business. The systems operate independently, and the gaps open up between strategy, operations, knowledge and execution.
Connected changes what each piece of work has available to it. When a proposal is created, information from previous client engagements can make it stronger. When someone in sales engages a customer, they can draw on proposals, marketing assets, organizational knowledge and established ways of working. HR teams reach onboarding materials, policies and capability frameworks in context. Finance works with operational and commercial information without searching across systems.
Information becomes available where the work is happening, and AI agents help every function work from the same source of truth.
Who keeps asking
What has been particularly interesting is the diversity of the organizations exploring this. Around half are professional services firms, which makes sense given how dependent they are on knowledge, expertise and collaboration. At the same time we see the same operational problem across industries that otherwise have very little in common.
If the fragmentation part of this is familiar, these pages carry the argument further.