Fix the Process Before You Add AI

After leading a number of AI solutioning workshops, a pattern keeps emerging.

Teams come in focused on AI. More often than not, the real opportunity turns out to be process realignment and data cleanup.

Mapping what actually happens today, not what the documentation says, almost always surfaces the gaps: steps without clear ownership, data scattered across disconnected sources, handoffs that exist for reasons nobody can quite explain anymore. Addressing those typically gets to 60 to 80% of the efficiency originally targeted.

AI then adds the next 20 to 40%. Meaningful improvement, built on a foundation that can support it.

When organizations skip that step and lead with AI, the tools often go unused. Not because the technology is wrong, but because the underlying process and data weren't ready for it. The model ends up solving a problem that was never properly defined.

The engagements that produce lasting results share a common thread: willingness to do the foundational work first, and treating AI as something that amplifies a working system rather than fixes a broken one.