Talk to us
Menu

Where you're stuck

Can't we skip augmentation and go straight to AI-native?

You can design AI-native workflows today, and plenty of software will show you where AI fits. The hard part is getting people to give up how they've worked for years, and they only do that when they've redesigned the work themselves.

What you've built

You have a mandate to go AI-native, and tools that can map an AI-native version of almost any workflow in an afternoon. Some of those maps are good.

The thing most likely holding you back

Organizational inertia. People have done this work one way for years, and a finished workflow asks them to give that up with no say in it.

Same design, two pathsAn AI-native workflow design can take two paths. Handed to the team finished, it lands on people who don't know their role, they put the brakes on, and it stalls; when the models change, you start over. Redesigned by the team that does the work, the team owns the new version, and when the models change, the team redesigns it again, so the path loops and keeps going.An AI-nativeworkflow designHANDED TO THE TEAMREDESIGNED BY THE TEAMLands on the teamfinishedPeople don't knowtheir roleThey put thebrakes onStallsNext model update: start overThe team redesignsits own workflowThey own thenew versionThe modelschangeKeepsgoingThe team redesigns it againWhen the models change, only one of these paths keeps going.
Same design, two pathsAn AI-native workflow design can take two paths. Handed to the team finished, it lands on people who don't know their role, they put the brakes on, and it stalls; when the models change, you start over. Redesigned by the team that does the work, the team owns the new version, and when the models change, the team redesigns it again, so the path loops and keeps going.An AI-native workflow designHANDED TO THE TEAMLands on the team finishedPeople don't know their roleThey put the brakes onStallsNext model update: start overREDESIGNED BY THE TEAMThe team redesigns its ownworkflowThey own the new versionThe models changeThe team redesigns it againWhen the models change, only oneof these paths keeps going.

What happens if you skip the people

Usually the new workflow goes live, and people work around it. Gartner found that 41% of employees already work around formal processes, and 38% have had to create new processes because of technology. (That's before anyone hands them an AI-native redesign.) Source: Gartner, Oct 2025

Meanwhile the licenses keep billing and the value stays thin: 88% of HR leaders told Gartner their organizations haven't realized significant business value from AI tools. And every model update resets the clock. A team that never learned to redesign its own work waits for someone else to do it again. Source: Gartner, Oct 2025

Melissa · 38 sec

Don't just do your job, own the process

Short from YouTube, plays on click

It's fixable, and it starts small: one team, one workflow.

5x

Employees are five times as likely to be top AI users when AI solves their own work frictions. Source: Gartner, Oct 2025 (2,986 employees)

Why skipping ahead stalls

The Hyperadaptive™ Model lays out five stages of becoming AI-native. Stage 2 is augmentation (think AI helping people with the work they already do). Stage 3 is automation, where AI takes over steps or whole workflows. Most organizations try to jump straight to Stages 4 and 5, scaling AI across the business before their people have done either.

We've seen this leap before. Business process reengineering made the same promise in 1990, and we don't need to repeat what happened next.

We've seen this promise beforeTwo moments side by side. 1990: Michael Hammer's Harvard Business Review article 'Don't Automate, Obliterate' told companies to start from a clean sheet. Outsiders redesigned the work and left, and Hammer and Champy's own unscientific estimate was that 50 to 70 percent of efforts missed their intended results. The redesigned process then held still. 2026: the promise is to skip augmentation and go straight to AI-native. AI can design the workflow in an afternoon, and the models change every few months. The open question: who redesigns the work next time?REENGINEERING1990AI-NATIVE2026THE PROMISE“Don't Automate, Obliterate.” Startfrom a clean sheet. (Hammer, HBR, 1990)WHO REDESIGNEDOutside teams, who then left.HOW IT WENT50 to 70% missed their intended results,by Hammer and Champy's “unscientific estimate.”AFTER THATThe redesigned process held still.THE PROMISESkip augmentation. Go straightto AI-native.WHO REDESIGNSAI can draft the workflowin an afternoon.HOW IT'S GOINGThe people who run it oftenweren't in the room.AFTER THATThe models change every few months.So who redesigns the work next time?With AI, the work has to be redesigned every time the models change.
We've seen this promise beforeTwo moments side by side. 1990: Michael Hammer's Harvard Business Review article 'Don't Automate, Obliterate' told companies to start from a clean sheet. Outsiders redesigned the work and left, and Hammer and Champy's own unscientific estimate was that 50 to 70 percent of efforts missed their intended results. The redesigned process then held still. 2026: the promise is to skip augmentation and go straight to AI-native. AI can design the workflow in an afternoon, and the models change every few months. The open question: who redesigns the work next time?REENGINEERING1990THE PROMISE“Don't Automate, Obliterate.”Start from a clean sheet.(Hammer, HBR, 1990)WHO REDESIGNEDOutside teams, who then left.HOW IT WENT50 to 70% missed their intendedresults, by Hammer and Champy's“unscientific estimate.”AFTER THATThe redesigned process held still.AI-NATIVE2026THE PROMISESkip augmentation. Go straightto AI-native.WHO REDESIGNSAI can draft the workflowin an afternoon.HOW IT'S GOINGThe people who run it oftenweren't in the room.AFTER THATThe models change every few months.So who redesigns the work next time?With AI, the work has to beredesigned every timethe models change.

Sources: Hammer, "Reengineering Work: Don't Automate, Obliterate," Harvard Business Review, 1990 · Hammer and Champy, Reengineering the Corporation, 1993, via Enclaria · Hammer later said he had been "insufficiently appreciative of the human dimension" (Wall Street Journal, 2008, via Lean Enterprise Institute)

Start here

Teach people to redesign their own work

Right now

People redesign their work in an ad hoc manner, with isolated success.

What is needed

Teams learn how to identify high-value, automatable work, learn how to redesign it for AI, own the result, and have the skills to do it again as the models change.

Melissa · 49 sec

How one bank cut an eight-hour task to three minutes

Short from YouTube, plays on click

A community bank asked one or two people in each department to find their worst pain point and match one AI solution to it. Its biweekly asset-liability committee prep went from eight hours to three to five minutes. Fifty of its 200 employees joined as champions, and the bank pulled its full rollout forward six months.

From the field

8 hours

to 3 to 5 minutes, at a community bank

The bank's own team did this work, using ideas from the Hyperadaptive model.

The Applied AI Workshop

One team, one real workflow, six hours. The people who do the work redesign it with AI, which builds their buy-in and the durable skills to keep redesigning as the models change.

Designed by Melissa Reeve, delivered by a Hyperadaptive delivery partner.

Find a delivery partner

Not sure what stage you are in? Find your waypoint in five minutes.

Questions about going AI-native

What's the difference between AI augmentation and AI automation?
Augmentation means AI helps people with work they already do. Automation means AI takes over steps or whole workflows. Most organizations need both, and people who've practiced augmentation usually make better automation calls.
Can we go straight to an AI-native organization?
You can design AI-native workflows now. Getting people to adopt them is the slower part, and it moves faster when the people who do the work help redesign it.
How do we get people to redesign work that AI might take over?
Make a trust commitment first, then prove it with a decision people can see. When Toyota and GM reopened GM's Fremont plant as NUMMI, the plant adopted a no-layoff policy, and within a year the same workforce took it from GM's worst quality to its best. People tell you where the real opportunities are once telling you can't cost them their job.
How long does it take?
The Applied AI Workshop takes one team through one real workflow in six hours. The team then uses the same approach on the next workflow.

Keep reading