We were in a routine discussion when my boss asked a seemingly straightforward question:
How do we encourage cross-functional teams to adopt AI in their workflows; reduce bottlenecks; experiment faster; and improve productivity?
On the surface, this sounds like a tooling problem. Introduce AI tools, train teams, and expect outcomes. But the question took me somewhere else.
The Story I Shared
In a village, a new barber arrived. He was skilled, efficient, and consistent. Word spread quickly. His shop became busy; customers kept increasing day by day.
In the same village, three struggling boys noticed this. They did a rough calculation; number of customers multiplied by price per haircut. To them, it looked like easy money.
They approached the barber and asked:
“What tools do you use?”
He showed them a comb, scissors, and a machine. That was enough for them.
They pooled money, bought the same tools, and opened their own shop.
For a brief moment, things looked promising. Curious customers walked in. There was attention, even excitement. But the boys didn’t have the skill, the discipline, or the understanding of the craft. Haircuts were poorly done. Experiences were ruined. Within days, the village knew. No one returned.
The tools were right. The outcome was not.
The Parallel With AI Adoption
This is exactly what is happening with AI today. Teams see others using tools like ChatGPT, Claude or many others and assume:
“This is the formula.” So they invest on the tools:
- Content teams generate copy
- Developers use AI for code
- Product teams use it for documentation
But without clarity and structure, the results are inconsistent or even damaging.
Just like the boys with the barber tools.
The First Batch Illusion
The boys did get customers initially because new things attract curiosity. Every product, feature, or workflow change gets a first batch:
- Early adopters
- Curious users
- Internal champions
This phase often creates a false sense of success.
Read more about the six stages of product lifecycle
The real signal comes later:
- Do users return?
- Does quality improve?
- Does trust increase?
This is where most teams fail.
Where Teams Actually Fail
From my experience across product and agile environments, failures around AI adoption don’t come from lack of tools. They come from:
- No Skill Development: Teams use AI outputs without understanding context, accuracy, or limitations.
- No Workflow Integration: AI is added as a layer, not embedded into decision-making or delivery systems.
- No Validation Loop: Outputs are not tested with real users or real scenarios.
- No Ownership: No one is responsible for quality when AI is involved.
The result is not acceleration; it is amplified inconsistency.
What the Barber Got Right
The barber’s success was not because of tools. It was because of:
- Repeated practice, skill built over time
- Understanding of customer expectations
- Consistency in delivery
- Clear ownership of outcomes
Tools were just enablers.
A Better Way to Introduce AI in Teams
Whether it is a barber shop in a village or a modern AI-powered product, the principle remains the same:
Initial attention comes from curiosity. Sustainable growth comes from capability.
If the goal is to reduce bottlenecks and increase productivity, the approach needs to shift:
1. Start with Problems, Not Tools
Identify where teams are actually stuck:
- Slow documentation
- Repetitive tasks
- Decision delays
Then map AI use cases.
2. Build Skill, Not Just Access
Train teams on:
- Prompting
- Validation
- Context awareness
3. Create Feedback Loops
Every AI-assisted output should be reviewed, tested, and improved.
4. Define Ownership
Someone must own the outcome, even if AI assisted in producing it.
Closing Thought
The gap between tool adoption and actual impact is now visible in real numbers.
- According to McKinsey’s State of AI 2025 report, AI adoption has broadened significantly, with 88% of organizations reporting that they use AI in at least one business function. Yet only a small fraction report meaningful impact on the bottom line.
- Through 2025/2026, roughly 80% of AI projects are expected to fail to deliver on their projected value. Gartner, RAND Corporation.
- The 70% Rule: In successful AI implementations, only 10% of the value comes from the algorithm, 20% from technology/data, and 70% from redesigning how work gets done. Boston Consulting Group (BCG)
What this means in practical terms:
- You may see a 20–40% speed improvement in isolated tasks using AI
- But you may also introduce quality drops, rework, and decision noise if systems are weak
- Over time, this can reduce user trust and retention, which directly impacts growth
Teams that pair AI with structured thinking, validation loops, and accountability can see significant gains in speed, cost efficiency, and experimentation.
Teams that don’t will simply move faster in the wrong direction. The tools are not the differentiator. The system behind them is.
Tahir Shahzad
Tahir Shahzad is a Product Manager, Product Owner, and technology consultant with over a decade of experience helping startups and organizations build products people actually use. If you are working on a product problem worth solving, reach him at tahirshahzad.com/contact/
