What cricket quietly teaches us about the future of Artificial Intelligence

By Sanjay Kala, Director – Technology Delivery Practice, Asia Pacific, SAS

India changes during IPL season. Families that barely sit together for dinner gather around screens .

Every chai tapri becomes a mini commentary box. WhatsApp groups turn into expert panels. Somehow, every second person in the country becomes more qualified than the actual coach of the team.

It was one such night. The clock had crossed 11 PM. An IPL match had reached the final over. Twelve runs were needed from six balls. The bowler wiped sweat from his forehead. The captain adjusted the field for the third time in two minutes. Fans had stopped blinking.

The television screen showed analysts behind laptops, coaches debating possibilities, players reviewing field positions, statisticians studying matchups and the captain standing quietly, thinking, before making his next move.

At that moment, the game no longer looked like just cricket. It looked like Artificial Intelligence.
Strange comparison? Maybe not. Without realising it, IPL teams already operate very similarly to the way modern AI agent systems work. For years, most people thought of AI as a chatbot: ask something, get a reply, move on. But the new world of AI is changing rapidly. AI is no longer just responding. It is beginning to think, plan, review, collaborate, use tools, and make decisions step by step like human teams working under pressure.

This new era is called Agentic AI. And one of the clearest ways to understand it is not through coding books or technical diagrams but through cricket.

Why understanding Agentic AI system design patterns matters
Agentic AI systems do not work through a single, monolithic brain. They work through patterns repeatable structures that tell different parts of the system when to think, when to delegate, when to reach for tools, and when to collaborate. These are called AI Agent Design Patterns.
There are five that every business leader, student, and professional should understand.

01 The Reflection Pattern: Pause. Review. Then respond.
A bowler appeals loudly for LBW. The entire stadium screams OUT. The crowd is convinced. But the wicketkeeper pauses for a second thinks carefully and quietly signals for DRS.

That single moment captures one of the most important patterns in Agentic AI. Reflection means an AI does not deliver its first answer blindly. A second internal process reviews the response checking accuracy, tone, logic, and completeness before anything reaches the user.

Experienced cricketers do this naturally. Great captains know emotional decisions lose matches; calm review wins tournaments. Modern AI systems are learning the same habit. Without Reflection, AI can sound extremely confident while being completely wrong. With it, AI behaves like a senior professional who double-checks before speaking .

In practice, this looks like a Generator agent drafting a business proposal while a Reviewer agent silently audits it for gaps and errors, then returns it for improvement.

02 The Routing Pattern: Right expert. Right moment. Right result.
A dangerous left-hander walks to the crease. The captain immediately withdraws the pace bowler and calls for the spinner not because the spinner is the biggest star, but because he is the right specialist for this exact situation. Death overs call for yorker experts. Turning pitches demand wrist-spin.

Routing in AI works identically. When a business leader asks an AI system to analyse sales data, build a presentation, and draft a communication, a Routing-enabled system does not apply one generic brain to everything. It intelligently distributes tasks across specialised agents: one for data analytics, one for visual design, one for professional writing , much like a project manager routing legal questions to counsel, financial modelling to the CFO, and communications to PR, all at once.

03 The Tool Use Pattern: Intelligence is nothing without the right instruments.
As an innings progresses, graphics light up with wagon wheels, strike rates against spin, and predictive probabilities. Behind every modern IPL franchise sits an invisible army of analysts, software systems, and performance dashboards.

Tool Use in AI mirrors this exactly. An AI travel assistant does not operate in isolation. it connects to mapping services, weather APIs, airline booking systems, and payment gateways to complete a single request. The intelligence is not only in the AI’s reasoning; it is in its ability to reach for the right external capability at the right moment, the way a consultant might use legal databases, financial modelling software, and CRM tools together to prepare a client pitch.

04 The Planning Pattern: Championships are won in the strategy room, not on the field.
Long before the toss, coaching staff study pitch conditions, prepare batting sequences, map bowling rotations, and build contingency strategies for rain or dew. Matches are frequently won before the first ball is bowled.

Planning in AI means breaking a large, complex goal into a series of smaller, ordered sub-goals, then executing them as a disciplined workflow. If someone asks an advanced AI system to help launch a startup, a Planning-enabled AI does not throw random ideas at the problem. It thinks step by step: conduct market research, analyse competitors, define brand positioning, estimate budget, build a go-to-market plan, much like a CEO mapping a five-year expansion, entering markets in sequence and building supply chains before scaling sales.

Cricket fans say it well: talent wins matches, but planning wins trophies. Agentic AI is learning the same lesson.

05 The Multi-Agent Pattern: No superstar wins alone. Neither does a single AI.
Modern cricket is no longer about one extraordinary player carrying the team. Winning franchises depend on coordinated ecosystems: batsmen setting platforms, finishers adapting to situations, seamers targeting matchups, fielders operating as a unit, physiotherapists managing fatigue, analysts feeding intelligence into the dugout, and leadership synthesising it all under pressure.

Multi-Agent AI reflects this orchestrated complexity precisely. Instead of one giant AI doing everything, modern systems deploy an ensemble of specialist agents collaborating toward a single goal, one agent gathering research, another analysing numbers, another creating reports, another reviewing quality, another handling the final presentation. Individually, they are useful. Together, they are transformative.

A global bank deploying simultaneous AI agents for fraud detection, customer service, and compliance works on the same principle as an IPL squad.

In many ways, an AI ecosystem today resembles a dressing room more than a traditional computer program.

What this means for every industry
These five patterns are already spreading across banks, hospitals, and retailers building coordinated AI systems instead of single overloaded chatbots. The future workplace looks less like humans versus AI, and more like humans directing teams of digital specialists. Leaders don’t need to code they need to ask whether their AI reviews its own work, routes tasks to specialists, plans in sequence, and lets agents collaborate.

The ball is still in the air
AI is becoming more human, not less – thinking, planning, and collaborating under pressure, like a great IPL team chasing victory. The question isn’t whether Agentic AI will reshape your world. It’s whether you’ll be the captain who understood the game or the spectator who only watched it.