For the past two years, most of the conversation around AI has focused on literacy. What tools exist? Which model is best? How do you write a good prompt?
But as we move deeper into 2026, something more interesting is happening.
The most effective founders, consultants, and small teams are no longer asking what AI can do. They are asking how intelligence should be allocated.
A recent masterclass by Professor Rem Koning of Harvard Business School, co-lead of the AI Founder Sprint, sheds light on how AI-native organizations are actually building profitable companies. His research reveals a simple but powerful shift:
The winners are not using AI as a tool. They are using it as a workforce.
The New Competitive Edge: Allocating Intelligence
For decades, business success came down to how well leaders allocated scarce resources.
- Warren Buffett built an empire by allocating capital.
- McKinsey built influence by allocating talent.
But today we are entering a new operating model.
The strategic resource is intelligence itself.
The modern entrepreneur’s job is increasingly to decide:
- Which work should be done by AI models
- Which tasks require human judgment
- Which processes should be automated entirely through agents
In practice, this means orchestrating multiple forms of intelligence.
A founder might use:
- one model to analyse research
- another to generate marketing assets
- an agent to update the CRM or website
- and a human expert to make the final strategic decision
The role of leadership shifts from doing work to directing intelligence.
For small teams and independent consultants, this is transformative.
It means a team of three people can now operate like a team of thirty.
The Judgment Gap: Why AI Isn’t an Equalizer
One of the most surprising insights from Koning’s research comes from a field experiment involving more than 600 entrepreneurs in Kenya.
Participants were given access to AI tools to assist with business decisions.
The results were unexpected:
- High-performing entrepreneurs increased profits by about 20%.
- Lower-performing entrepreneurs saw profits decline by around 10%.
The difference wasn’t the technology.
It was judgment.
The weaker performers followed AI recommendations without questioning them. The stronger entrepreneurs used their own experience to filter the advice.
They treated AI as an amplifier, not a replacement for thinking.
For the Imbila community, this insight matters.
AI will not fix a weak business model.
But in the hands of someone who understands their customer and market deeply, it can dramatically accelerate growth.
From Chatbots to Virtual Employees
Most organizations are still using AI as a conversation tool.
They open ChatGPT, ask a question, copy the output, and move on.
But the most advanced teams have already moved beyond this stage.
Instead of asking a chatbot how to do something, they deploy AI agents that execute the work.
Think of the difference:
Traditional workflow:
- Ask AI for advice
- Copy the response
- Implement manually
AI-native workflow:
- Define the objective
- Deploy an agent
- Let the system execute the task
Examples include agents that:
- update websites
- manage CRM records
- generate marketing campaigns
- analyse sales pipelines
- produce reports automatically
This shift changes how companies scale.
Traditional organizations scale through headcount.
AI-native organizations scale through compute.
Small teams can suddenly operate with the capacity of much larger companies.
The Risk of “Vibe Coding”
The rise of natural-language software development tools—often called “vibe coding”—means that building products is becoming easier than ever.
Founders can now create functional software simply by describing what they want.
But Koning warns about a new trap.
Because building is now fast and fun, it is easy to build too much.
Teams can fall into a feature loop, continuously adding functionality without solving a real problem.
The winners in the AI era will not necessarily be the most technical founders.
They will be the people with deep insight into a specific customer problem.
Technology is becoming abundant.
Understanding what actually matters to a customer remains scarce.
The Imbila Perspective: Build for the Real World
At Imbila, we often emphasise a simple shift:
Move from AI literacy to AI building.
Professor Koning’s research reinforces that direction.
Three principles stand out.
- Develop Judgment
AI will produce endless answers.
Your advantage is the ability to recognize which answers are useful.
This comes from experience, domain knowledge, and understanding customers.
- Build Systems, Not Prompts
Prompting tools is useful.
But the real leverage comes from systems that execute work automatically.
That means thinking in terms of workflows, agents, and automation rather than individual prompts.
- Solve Real Problems
The biggest opportunities are not necessarily in building the next general AI tool.
They lie in solving very specific problems.
The world doesn’t need another generic AI platform.
But it may need the best AI-enabled CRM for a niche industry, or the best automated workflow for a specific profession.
📘 Looking for practical ways to apply AI in your consulting or business workflow? Explore our guide: AI Quick Wins for Independent Consultants https://www.imbila.ai/ai-quick-wins/
Final Thought
The cost of intelligence is rapidly approaching zero.
When intelligence becomes abundant, the scarce resource is no longer knowledge.
It is judgment, taste, and human agency.
The entrepreneurs who succeed in the AI era will not simply use AI tools.
They will learn how to allocate intelligence itself.
And that may become the defining skill of the next decade.