it's not when to use AI...
it's how.
I help companies find the right application for AI, so teams actually use it
AI is powerful, but what is more powerful is how we choose to use it.
I think of AI like a car engine: no one wants to see it; it can power everything, but we are still driving it.
most AI projects don't fail on the tech; they fail on the people.
We're not building faster horses; we're designing all new roads.
For many organizations, adding AI to the process makes sense, but after months, the AI fails to deliver the promised ROI, and/or adoption still lags.
It's often not a technology problem. It's habits, trust, and the comfort of the familiar, human-rooted psychology.
I work on both sides: helping companies explore what to build and ensuring people actually use it.
where I help.
assess which workflows to optimize.
Before building any tool, I use a workflow framework to determine:
-
how work actually gets done,
-
who does it,
-
where the hours go, and the bottlenecks,
-
which tasks are repetitive enough for AI
-
which tasks require human-in-the-loop.
-
how we revise, evaluate, and iterate
-
reaching goals and prolonged success
-
map how work moves through your team
-
find the high-volume, low-joy tasks
-
decide what AI should and shouldn't touch
-
rank opportunities by effort and impact
work with developers.
As your liaison, I am between your business and the builders. The developers get a clear brief and guidance from someone working on your behalf.
The result: a solution that fits how your team works, on-brand, compliant, and scalable.
-
requirements written in plain language
-
user flows and interface design
-
brand voice, prompts and guardrails
-
testing with the people who'll use it
encourage adoption.
This is where most rollouts stall. Throughout development, we recognize resistance cues and help teams move beyond them.
-
a rollout plan built on behavioural principles
-
internal messaging and training
-
champions inside the team
-
usage tracking at 30, 60 and 90 days
adoption is a behavioral
problem.
People don't resist AI because they're stubborn. They resist because they are human.
status quo bias.
People stick with the default. So we make the AI-assisted way the default, not the extra step.
friction.
Every extra click is a reason to quit. We cut steps until using it is easier than not.
social proof.
People follow people. Finding early influencers or users can encourage adoption.
loss aversion.
"Save time" is abstract. "The four hours you lose every week" is not. Framing changes how we see it, that changes behavior.
habit stacking.
New tools stick when they attach to routines people already have and make them easily repeatable.
the IKEA effect.
People value what they helped build. Involve the team early, and they are more likely to adopt it than resist it.
accountability.
It's hard to trust a tool when your reputation is on the line. Realignment of responsibility and incentives is not the only tool to drive adoption.
identity.
Many of us tie our identity to our work or outcomes. AI can take away how we see ourselves.
adopting AI:
going beyond buying the licenses to understanding where it belongs
One size and one solution do not fit all.
There are multiple reasons why AI adoption declines or reverts, and most can be addressed by understanding the root causes of resistance. I've worked with teams to identify and overcome blocks.
ai is a tool, not a shortcut
Half the value is knowing where not to use it.
Here's how I determine the workload in my own practice.
research.
AI gathers data, reviews the landscape, taps into industry insights, and recognize audience signals.
HUMAN: understanding why, further interpretation, and strategic framing
planning.
AI can help seek and organize opportunities, map timelines, and consider resources
HUMAN: the final judgment, execution and collaboration.
creative.
AI: testing frameworks, insights, copy considerations
HUMAN: creative direction, design and final approval.
monitoring.
AI: performance tracking, competitor alerts, trend signals.
HUMAN: what it means, external influences, and what to do about it