Ivie Osobase

I am Ivié

[Ivie Osobase @IOsobase]

people don't resist change
they resist loss
loss of control over their workflow
loss of expertise built over years
loss of certainty about their role
loss of identity tied to how they've always worked
when you don't address these losses in your AI rollout
you get a grief response disguised as a technology problem

[Ivie Osobase @IOsobase]

resistance to AI is not irrational
it's logical
being asked to trust something you don't understand
give up control you've earned
change habits that have worked for years
resistance is information
it tells you exactly where your adoption design is failing
design for it. don't manage it.

[Ivie Osobase @IOsobase]

3 questions every AI transformation must answer before go-live:
who will resist and why?
what does success look like in behavioural terms?
how will we sustain change after launch?
most programmes cannot answer any of these at launch
then wonder why adoption stalls at 12%.

[Ivie Osobase @IOsobase]

traditional change management was designed for a different era
kotter. prosci. lewin.
built for transformations where the destination is clear and the path is linear
AI transformation is neither
the destination changes as the technology evolves
applying legacy frameworks to AI transformation is navigating with an old map

[Ivie Osobase @IOsobase]

most AI programmes have a project manager
very few have a change manager
that distinction is costing organisations millions in unrealised ROI
change management isn't a soft skill
it's the discipline that determines whether the technology delivers any return

[Ivie Osobase @IOsobase]

most AI programmes have a detailed implementation plan
very few have a sustained adoption plan
one covers what gets built and when
the other covers which behaviours need to change and how success will be measured in human terms
one gets you to launch.
the other gets you to transformation.

[Ivie Osobase @IOsobase]

the orgs winning at AI adoption aren't the ones with the best technology
they asked a different question before they started
not: "what AI should we deploy?"
but: "what behaviour change needs to happen for this to deliver value?"
that question changes everything

[Ivie Osobase @IOsobase]

most organisations measure AI adoption wrong
licences deployed
logins completed
features accessed
none of these tell you if AI is changing how decisions get made
six months after launch, leadership can't answer whether the transformation is working
because they never defined what working looked like

[Ivie Osobase @IOsobase]

previous tech transformations asked people to do the same work in a new system
AI asks people to delegate judgment to something they don't fully understand
to trust outputs they can't verify
to give up control over decisions they've owned for years
that's not a training problem
that's a psychological transformation

[Ivie Osobase @IOsobase]

three days of AI training before go-live
high completion rates
six months later everyone has reverted
training creates awareness
it does not create behaviour change
knowing how to use something and having a reason to use it are completely different problems

[Ivie Osobase @IOsobase]

i call it the pilot trap
successful pilot with 20 enthusiastic volunteers
leadership approves full rollout to 2,000 people
adoption flatlines at 12%
the pilot selected for people who wanted to change
the rollout hit everyone else
a successful pilot doesn't predict a successful rollout

[Ivie Osobase @IOsobase]

enterprise AI budget breakdown:
technology: 75%
training: 20%
behaviour change: 5%
that last number is why most transformations fail
you can't spend 95% on the system and 5% on the people
the return on AI is in the behaviour change
not the technology

[Ivie Osobase @IOsobase]

most AI transformations don't fail because the technology fails
they fail because nobody designed for the humans
the AI works
the people don't change how they work
and six months later leadership asks what went wrong

[Ivie Osobase @IOsobase]

3 types of curiosity that drive AI retention:
competence curiosity: "am i getting better?"
discovery curiosity: "what am i starting to see?"
social curiosity: "what are others finding?"
if your product answers all three too early
usage collapses
keep all three alive indefinitely

[Ivie Osobase @IOsobase]

netflix doesn't win by explaining everything
it wins by leaving things unfinished
"continue watching"
"new episodes"
"because you watched"
every interaction creates a curiosity gap
lesson for AI products:
don't just deliver value
design what's left undiscovered.

[Ivie Osobase @IOsobase]

AI products are uniquely exposed to the satisfaction trap
fast answers
clear outputs
immediate resolution
feels great once
without intentional design, curiosity collapses day one
AI products don't fail because they're confusing
they fail because they feel finished

[Ivie Osobase @IOsobase]

for a decade AI competed on capabilityfor a decade AI competed on capability
most accurate model
fastest inference
most features
that race is over
AI platforms are converging
97% vs 99% accuracy? users can't tell the difference
investors know this before most founders do
the new questions:
not "how smart is your AI?"
but "how curious does it make users?"

[Ivie Osobase @IOsobase]

question every AI leader should ask:
what if improving your AI isn't what increases your valuation?
a fintech closed Series B at 40% premium
without changing the model
day 30 retention: 18% to 61%
that single metric = $20M valuation uplift

[Ivie Osobase @IOsobase]

AI products are especially vulnerable
they solve problems too fast
without intentional design, curiosity collapses on day one
the question most teams ask: "is this easy to understand?"
the question that drives retention: "what reason does the user have to return?"
design for exploration, not completion

[Ivie Osobase @IOsobase]

humans aren't wired for instant mastery
we're wired for progressive mastery
curiosity appears when something is partially revealed
not when everything is explained upfront
tiktok never reveals its algorithm
netflix hints at what you might discover
wordle gives you one puzzle a day
none of these are accidents