The direct answer
Before you automate anything, simplify it. AI is not the bottleneck. The workflow it runs is. Automate a process that wastes steps, and you have built a machine that wastes steps faster.
The fact: Knowledge workers spend 60% of their time on work about work: chasing updates, switching tools, attending unnecessary meetings. That is the process problem, and it exists with or without AI.
SOURCE: Asana, Anatomy of Work Index — Read the index
Find the real bottleneck
The test I use. Take one workflow and write down every step and every handoff. You will usually find the waste in a specific place: the same status update, asked for three different ways. The same file, rebuilt because nobody can find the original.
Fix that step first. Then automate what remains. This is also why I don't sell AI as the answer. I sell the system: simplified process, then a play that runs it.
The fact: McKinsey estimates generative AI could add $2.6 trillion to $4.4 trillion in value annually, concentrated in customer operations, marketing, software engineering, and R&D.
SOURCE: McKinsey, June 2023 — Read the analysis
The Play: Find the Bottleneck
State the problem. The system maps the workflow, spots the repeated waste, and shows you what to simplify before anything gets automated. You make the call. Then the play runs the cleaned-up version.
Where to start
Pick one process this week. Map it. Cut one step. Then run it. The win is smaller than a headline and easier to repeat, which is exactly what makes it real.