Our Journey of Transforming Utility Meter Reading with AI

There are some journeys in business that begin with a business plan. And then there are journeys that begin with a problem you simply cannot stop thinking about.

For us, AI-powered meter reading was one of those journeys. Our belief was simple: meter reading should not just be digital — it should be intelligent, accurate, transparent, and capable of protecting the revenue of power utilities.

The idea began during COVID-19, at a time when the world was learning to work differently. Field operations were becoming difficult, physical interactions were being restricted, and the traditional ways of collecting meter readings were under increasing pressure. We started asking ourselves a simple question: can technology make meter reading more accurate, more reliable and less dependent on manual intervention?

That question stayed with us. And somewhere during those uncertain days, the idea of using AI and OCR to read electricity meters began to take shape. We did not know then how far that idea would travel. We only knew that it could solve a real problem.

September 2021: Our First Step Towards Bihar

In September 2021, we approached Bihar with this idea. I still remember that period vividly because it coincided with Ganesh Chaturthi — a festival associated with new beginnings, hope and the removal of obstacles. For us, September 2021 became exactly that: a new beginning.

At that point, Bihar's meter-reading process was largely photo-based. On the surface it sounded simple: a meter reader takes a photograph of the electricity meter, and that photograph becomes the basis for reading and billing. But there was a fundamental question — how do you know whether the photograph is genuine? A photograph could be real. But it could also be a photograph of an old photograph, a manipulated image, a non-meter image, or something else entirely. The system could capture an image. But it needed to understand the image. That distinction became the heart of our work.

From a Photograph to a Trusted Meter Reading

We began working on a system that could look at the meter image the way an intelligent reviewer would. Not just “what number do I see?” but “is this actually the right meter? Is this a genuine image? Is the reading correct? Does it make sense?”

By November 2021, we introduced our Bharat Smart Services AI-powered OCR meter reading solution in Bihar for around 40,000 consumers, on the auspicious occasion of Ganesh Chaturthi. That was a small number compared with what came later. But to us, those 40,000 consumers meant something much bigger — they were our proof that the idea we had carried from the uncertainty of COVID into the field could actually work at scale.

Our system was designed to do much more than simply read digits. It could check photo accuracy and authenticity, identify whether the meter was real or potentially spoofed, and validate different meter types, including digital/electronic and mechanical meters. It could also flag unclear images, invalid meter readings, invalid parameters and suspicious or spoofed images. The technology was gradually turning a simple photograph into a validated data point for billing.

This was more than a technology achievement. It demonstrated that better meter-reading accuracy could directly contribute to better billing and, ultimately, better revenue realization for the utility. The solution we have today builds on that same philosophy: validate the meter, validate the image, validate the reading, validate consumption, and allow only validated information to move towards billing.

And Somewhere Along the Way, We Saw the Light

There are moments in an entrepreneur's journey when you know something has changed. For us, this was one of them. Until then, AI-powered OCR meter reading was an idea we believed in. Bihar gave us the opportunity to see that belief become reality. The first deployment taught us what happens when technology meets a real problem at a real scale.

We learned from the field. We improved the technology. We understood the challenges better. And most importantly, we began to see that this was not just about reading a meter — it was about creating trust in the entire chain between meter reading and billing.

A correct reading leads to better billing. Better billing improves revenue realization. And stronger validation helps protect against errors and revenue leakage. That fundamental chain remains at the centre of our solution even today.

November 2022: From 40,000 to 1 Crore

A year can change everything. By November 2022, our service had expanded to around 1 crore consumers in Bihar, at that stage operating through an agency model. The Bihar power utility recognized the potential of the solution, and we moved from being a technology service provider executing a limited deployment to becoming a trusted technology partner supporting the utility's meter-reading ecosystem.

Think about that journey for a moment: from an idea born during COVID, to approaching Bihar in September 2021, to starting with around 40,000 consumers, and then, within roughly a year, reaching 1 crore consumers. For our team, this was not simply a number — it was validation that a technology solution built around a real field problem could grow from a pilot-scale deployment into something much larger.

May 2024: The Journey Came Full Circle

The next chapter came in May 2024, when the Government of Bihar released a tender for the service. This was an important moment for us — we had started by approaching Bihar with an idea, we had built the technology, we had deployed it, we had scaled it. And now, we were entering a formal tender process for the very problem we had started working on years earlier.

We participated. And we succeeded. For me personally, this was one of the most meaningful moments of the journey. Because winning a tender is one thing. But seeing a solution that was once only an idea become something that an entire power ecosystem could trust — that is a different kind of achievement.

Uttar Pradesh: From Pilot Scale to Nearly 2 Crore Readings Every Month

Our next major journey was with Uttar Pradesh Power Corporation Limited (UPPCL). In the initial phase, around November 2024, we started with approximately 1 lakh meter readings, working alongside Quess Corp. Once again, the objective was simple — prove the technology through actual field operations.

As deployment increased, stakeholders could see the difference between conventional meter-reading processes and an AI-powered approach. OCR was no longer simply reading numbers from a meter display — it was becoming an intelligent validation layer between the field and the billing system, helping improve reading accuracy, reduce incorrect manual intervention, identify field-level irregularities, validate meter photographs, and bring greater accountability to the entire meter-reading process.

Today, in 2026, we are working as an official technology partner in the UPPCL ecosystem, supporting approximately 2 crore meter readings every month through AI-powered OCR technology. From 1 lakh readings to nearly 2 crore readings per month — when technology delivers measurable operational value, adoption follows.

Odisha: Scaling Across the Tata Power Distribution Network

Our journey then expanded to Odisha. We started our association with Tata Power Central Odisha Distribution Limited (TPCODL) and gradually expanded the solution across the wider Tata Power distribution ecosystem in the state:

• • Tata Power Central Odisha Distribution Limited (TPCODL)
• • Tata Power Northern Odisha Distribution Limited (TPNODL)
• • Tata Power Southern Odisha Distribution Limited (TPSODL)
• • Tata Power Western Odisha Distribution Limited (TPWODL)

This expansion was another important milestone. It proved that the solution was not limited to one geography, one utility, or one operational environment. The same AI-driven philosophy could be adapted and scaled across different utilities while maintaining the fundamental principles of accuracy, transparency, field accountability, and revenue protection.

Expanding Beyond States

The journey did not stop with Bihar, Uttar Pradesh, and Odisha. We continued taking the same technology and operational experience into newer utility environments, including Delhi, Chandigarh, Punjab, Jharkhand, and other regions — and we have also extended our technology capabilities within the Tata Power ecosystem in Delhi, while continuing to work with utilities and partners across multiple states.

Each new implementation has reinforced one fundamental reality: meter reading is not simply a field activity. It is the first and one of the most critical points in the revenue cycle of a power distribution company. If the reading is incorrect, everything that follows can be affected. But if the meter, image, reading, location, consumption pattern, and field activity are intelligently validated at the source, the entire billing cycle becomes stronger.

From OCR to Intelligent Revenue Protection

Our vision has therefore evolved beyond basic Optical Character Recognition. Today, AI can help validate the meter type, image authenticity, extracted reading and consumption pattern before billing. It can identify issues such as non-meter images, photo-of-photo/spoof attempts, abnormal readings and manual modifications — creating a stronger and more auditable connection between field activity and billing. We now look at AI-powered meter reading as an intelligent ecosystem capable of validating:

• The Meter — Is the correct meter being captured?
• The Image — Is it an actual meter image, clear and genuine?
• The Reading — Has the reading been correctly identified?
• The Consumption — Is the consumption pattern logical or abnormal?
• The Meter Reader Activity — Is the field activity genuine and compliant?
• The Location — Was the reading captured from the correct geographical location?
• The Billing Input — Is the information reliable enough before it reaches the billing system?

Our Journey in Numbers

Bihar Was Not Just Our First Market. It Was Our First Belief.

Today, our AI-powered OCR meter reading service has gone far beyond where we started. What began as an attempt to solve one problem — reliably reading a meter from a photograph — has evolved into a much broader intelligence layer around meter reading and billing.

But when I look back, I don't first think about the technology. I think about September 2021. I think about Bihar. I think about Ganesh Chaturthi. And I think about a small team carrying a big idea into a market and hoping that the world would see what we saw.

Every Big Journey Begins Somewhere

Ganesh Chaturthi has always represented new beginnings for millions of people. For us, that September in Bihar became our own new beginning. We did not know that the 40,000 consumers we started with would eventually become the foundation for a journey touching crores of consumers across Bihar, Uttar Pradesh, Odisha, Delhi, Chandigarh, Punjab and Jharkhand. We did not know how many times we would have to improve, adapt, persevere and believe before reaching where we are today.

But perhaps that is the nature of building something new. You don't always see the destination when you take the first step. You only need to believe that the step is worth taking.

Our first step was Bihar. Our first opportunity was a photograph. Our first challenge was to make that photograph trustworthy. And somewhere along that journey, we saw the light.

We believe the future of utility operations will not depend only on digitization. It will depend on Intelligent Digitization — where Artificial Intelligence does not simply collect information but continuously validates it, questions it, identifies anomalies, and helps utilities make better decisions. When AI ensures the right meter, the right image, the right reading, and the right consumption, the utility gets the right bill — and ultimately the right revenue.

We began with meter reading. We progressed towards billing intelligence. And now, we are moving towards something much larger.

Our journey continues.