Most teams are stuck
My perspective on moving from individual AI use to unlocking AI potentialThe framework
Most organizations treat AI adoption as a single decision. It is not. It is five distinct stages, and the hardest part is not starting. It is crossing the chasm between basic prompting and workflow integration.
Some teams embed Copilot into PowerPoint and call it a day. Teams designing and building AI are seeing higher productivity and growth. Between those two groups sits a gap that most teams do not cross, because crossing it requires changes to workflow, measurement, and incentive, not just tools.
Figure out where your team actually is. Then decide if the next level is worth the work.
The five levels
Level 0
AI-absent
The team does not use AI. Workflows unchanged in years
Built by hand in Excel
Manual emails, no template
Level 1
AI-curious
Using AI inside existing tools, Copilot in Word, Excel, Gmail
Copilot drafts bullets inside Word
Copilot suggests replies in Outlook
Level 2
AI-active
Going directly to ChatGPT or Claude, describe it, AI builds it
Claude writes the full report from raw data
ChatGPT drafts the full email sequence
The chasm
Most teams stall here because they lack workflow, governance, and experience
Level 3
AI-integrated
AI is in workflows with measurable outcomes
AI generates, humans review, time down 75%
AI drafts, humans review and deliver to clients
Level 4
AI-native
Company delivers AI products and markets them to clients
Dashboards write themselves, AI flags what needs action
Clients supported from contract signed to first use
Level 0
AI-absent
The team does not use AI. Workflows unchanged in years
Built by hand in Excel
Manual emails, no template
Level 1
AI-curious
Using AI inside existing tools, Copilot in Word, Excel, Gmail
Copilot drafts bullets inside Word
Copilot suggests replies in Outlook
Level 2
AI-active
Going directly to ChatGPT or Claude, describe it, AI builds it
Claude writes the full report from raw data
ChatGPT drafts the full email sequence
The chasm
Most teams stall here because they lack workflow, governance, and experience
Level 3
AI-integrated
AI is in workflows with measurable outcomes
AI generates, humans review, time down 75%
AI drafts, humans review and deliver to clients
Level 4
AI-native
Company delivers AI products and markets them to clients
Dashboards write themselves, AI flags what needs action
Clients supported from contract signed to first use
The chasm
Between L2 and L3, most teams stall
L0 to L2 feels natural. Install an LLM, try it, get comfortable. Most teams become proficient at these stages pretty quickly.
L3 is a workflow problem. That requires changing workflows, defining what “good” means with AI in the loop, and measuring whether the change actually matters. This requires leadership commitment, operational redesign, and honest measurement.
Most teams stop at L2 because it feels like progress – the team is using AI! But the company has not actually taken advantage of AI. That is the chasm.
The hard part is not knowing where you are. It’s deciding to go to the next level.
If you want to talk through where your team is and what it would take to get to that next level, book a 60-minute working session. We’ll evaluate where you are today and where AI can take you in the future.
Book a working session
Louis Levine 