Team behind Aadhaar now creating digital infrastructure that aims to disrupt financial services, make service delivery cheaper and efficient and boost start-up ecosystem
On any given day, at any of the upmarket cafes that dot Koramangala in Bengaluru, you will find a bunch of youngsters huddled around mugs of coffee, trading ideas on how they will change the world and make tonnes of money.
There is one place in Koramangala that strikes an even more audacious note. The people there don’t talk about solving one problem in one sector. They talk about disrupting entire industries. They draw plans to change how the entire state machinery works. Their idea of scale is touching the lives of a billion people.
Only, they are not talking about the fanciful. They are talking about the mundane, about people getting access to money, better healthcare, better education. All this with the help of technology.
We are talking about the people behind Aadhaar, the world’s largest biometric identity project: Nandan Nilekani, who made his name at Infosys Ltd, and went on to lead the Aadhaar project; the band of men he put together to pull it off, within budget and ahead of time; people like Pramod Varma and Sanjay Jain, who quit their marquee jobs to build the technology behind Aadhaar; men like Sharad Sharma, former head of Yahoo Labs in India and a venture investor who spends a good part of his time bringing together volunteers who work on the project.
And Aadhaar, the project, is massive. In May, it crossed one billion registrations. It took just five-and-a-half years to achieve that—the fastest by any digital platform.
Every quarter, Aadhaar was registering the equivalent of New Zealand’s population. And it had to be accurate at this scale. Even 99.5% accuracy would have been like Singapore getting it wrong for its entire population.
But that’s not what this story is about. By itself, a unique number for every Indian is of no value. The value comes from people and organizations using the Aadhaar system to bring down the cost of service delivery dramatically (in the case of mutual funds, from ₹ 1,500 to a negligible ₹ 10 for authenticating a new user) and opening up choice of service providers for the poor, from education, to healthcare, financial services and ration shops—that’s the big story.
The use cases extend to more prosperous Indians too—for example, easing road toll collection to decongest traffic and reduce waiting time at the toll booth.
All of that is made possible by the system’s ability to authenticate that a person is in fact who he claims to be—all online, in seconds, using an inexpensive device.
The team is building a number of layers on top of the Aadhaar system. These layers are all digital, open (as in others can work on the application programming interface, or API, to create services) and connected to each other. Think of it as a collection of digital goods. The team calls it the India Stack.
It creates a digital infrastructure for presence-less (no need for physical authentication), paperless and cashless service delivery.
It does so by bringing together five elements:
1. e-KYC (know your customer)
2. Aadhaar Payments Bridge System, which essentially turns an Aadhaar number into the person’s financial address.
3. eSign, a digital signature.
4. Unified Payment Interface (UPI), which makes sending money as easy as sending an email or an SMS.
5. A consent layer to share personal data (mark sheets, health records, financial transactions) with a bank, insurer, employer or university for a limited time for a specific purpose.
India Stack can potentially make government services more efficient, and bring down leakages. It can give a boost to the start-up ecosystem by providing an open digital infrastructure, much like how building highways can boost the economy.
“India Stack is for Bharat," says Sharma.
“Aadhaar was designed to make life convenient for the average person," adds Sanjay Swamy, who was a volunteer at Aadhaar and is now a managing partner at venture capital firm Prime Venture Partners.
It is an ambitious project, and has its share of naysayers.
Sceptics are concerned as much about the security of all this personal data as about its accuracy.
Indeed, it will need continuous technological improvement and the growth of an ecosystem for disruptive applications for all of it to come together. More on that later. Let’s first understand the intent behind India Stack.
It evolved over the past few years because of demand from users, encouragement from regulators and the team’s belief that the possibilities are endless.
After identity and authentication, a need that arose almost immediately was for physical copies of Aadhaar, especially to fulfil bank and telecom companies’ KYC requirements. Thus came e-KYC, an authorized, digital document that contains the individual’s name, address, gender and date of birth. An Aadhaar holder can authenticate his information biometrically, consent to sharing it and digitally send it to anyone.
Regulators were quick to see Aadhaar’s scope. During his early days at the Unique Identification Authority of India (UIDAI), Nilekani got a call from Reserve Bank of India’s (RBI’s) deputy governor Usha Thorat, who spoke about how Aadhaar can be used to create a more inclusive financial system. The result of such conversations—and Nilekani’s own role as an adviser to the National Payments Corporation of India—was Aadhaar Payments Bridge System. So, when a government department sends direct benefit transfers to an Aadhaar number, it is credited to the bank linked to it. The beneficiary is free to opt for any bank.
With UPI you can send or receive money through a UPI-enabled app to your virtual address (yourname@bankname)—you don’t have to fill a mobile wallet first. It uses the 24x7 immediate payment service IMPS.
Yet another layer was added when Ram Sevak Sharma, who was the director general of UIDAI, asked why so few people used digital signatures. Jain and Varma’s team worked with the Controller of Certifying Authorities to launch eSign. So, you can draft your rental agreement on your mobile phone, eSign it, and it would serve as a legally binding document.
As identity, authentication, signatures and transactions move online, hundreds of documents will turn digital. For example, this year’s CBSE mark sheets are available in digital format. You can give a college permission to access it digitally, just like you would share a document from Dropbox or Google documents, and obviate the need to photocopy and attest it. There are hundreds of use cases. The team is working on a digital locker which will make all of this possible.
And finally the consent layer, which ties in with the digital locker. The team is working on a framework that will let you share your data with various entities for a specific period and for a specific purpose. For example, a bank might want your income and expense data to assess your creditworthiness, or you can use your health data to get better insurance.
Each of these might seem small, but together, their impact could be far reaching and there are reasons for optimism.
The Aadhaar card is based on just four data points—name, date of birth, gender and address—with an option to add email and phone number.
This minimalism helped Aadhaar achieve scale. The fewer details you want, the fewer documents you need to support them.
The most important reflection of this idea is its technical architecture—an hourglass architecture. It has a thin waist (the Aadhaar number), but it brings together several layers of applications and devices above and below it. Its strength lies in its simplicity and ability to work with a large number of systems.
It is a principle they have carried on to other elements of India Stack. UPI, for example, is just an interface that sits on any banking system, enabling easy transactions on the mobile.
By narrowing the scope of what they would do, the team could think broadly about the role others would play. For the biometric systems for example, UIDAI gave the specifications, built the APIs to make sure different systems spoke to each other and let the vendors compete.
In the broader market such competition typically leads to lower prices and better quality, and so it was with UIDAI.
The value of Aadhaar, however, is in people using it. For Uber or Ola, the number of app downloads is important, but not sufficient. The value is in the number of times customers use it. That, to a large extent, depends on the number of drivers and passengers on the platform. Once they have sufficiently large numbers, the network effect comes into play.
The same dynamics will work with India Stack.
Some of the biggest backers in the government immediately grasped this. When Nilekani and his team presented the idea to the then finance minister Pranab Mukherjee in the early days, he used the metaphor of a railway platform to describe its significance. By itself, a railway platform has no value. But with tracks and trains, its impact can be huge.
What sets Aadhaar and India Stack apart is not its technology per se. After all, Apple’s iPhones use biometric authentication to unlock the phone and its apps. Both Google and Apple have their own payment systems, Google Wallet and Apple Pay. But they are owned by Google and Apple. India Stack, on the other hand, is open. It is public goods.
With India Stack a good part of that platform is ready. The government has started testing some of its bigger trains—cash transfers for workers under the Mahatma Gandhi National Rural Employment Guarantee Scheme (MGNREGS) and for those who have opted for cooking gas subsidy, and in some parts digitization of the public distribution system (PDS). The private sector is showing interest too. In the next few years we will see digital products and apps built around it. Only then will we see the true impact of Aadhaar.
When people talk about an ecosystem, they talk about entrepreneurs, investors and catalysts (typically industry bodies such as MAIT and Nasscom), says Sharad Sharma, one of the founding members of think tank Indian Software Products Industry Round Table (iSPIRT). However, there are three other important elements:
1. The government especially playing the role of a market maker.
2. Regulators making sure that the rules don’t stifle innovation.
3. Availability of public goods.
In India, the different pieces are picking up speed. Not in full measure, but in the right direction. State agencies have used Aadhaar for subsidies and cash transfers, in effect acting as a anchor client, and seeding a large number of users. Regulators appear to be open. RBI has approved new payments banks and has said it will give licences more regularly. The Securities and Exchange Board of India (Sebi) is digitizing equity and mutual fund investments through e-KYC.
There was a missing piece—public digital infrastructure. This is where the volunteers, who are keen on building software products in India, are making a difference. Much of the institutional support comes from iSPIRT. It has brought together a set of people who realize that there is a need to create a digital public infrastructure to enable creation of businesses.
The initial push came when Nilekani took charge of UIDAI. He had left Infosys on a high note. He had just published Imagining India, which set forth his vision of the future. His interviews and speeches fired up engineers and managers. Some of them took a sabbatical to volunteer for the Aadhaar project. Some, like Jain and Varma joined full time.
It’s no different today, except that the volunteers do their day job and in their spare time conduct hackathons, seminars and workshops, evangelizing the technology. Though the ownership of India Stack rests with government agencies (UIDAI for Aadhaar, the Controller of Certifying Authorities for e-Sign, the National Payments Corporation of India for Aadhaar Payment Bridge System and UPI.)
The best among them don’t even take credit for all the hard work. “They are like Sherpas. They help the mountaineers climb up the peaks, and they remain in the background," says Sharma.
Last September, Nilekani took the stage at a TiE event in Bengaluru to persuade his audience that a number of technology trends are conspiring to create a WhatsApp moment for the financial sector.
His basic argument ran thus: Some years ago, people exchanged messages on mobile phones using SMS and on the Internet using chat services such as Yahoo Messenger. When phones became an Internet device as well, one would have thought that these companies would continue to dominate the messaging space. Yet that privilege went to WhatsApp, a start-up with just about 40 engineers. More messages are sent over WhatsApp than SMSs through all telecom companies put together.
The financial sector, Nilekani said, was facing a similar disruption. Because of the digital infrastructure for cashless (UPI), paperless (e-KYC, eSign and digital locker) and presence-less service enabled by India Stack. Together this will bring the cost down by a huge factor, increase reach and open up opportunities.
Take mutual funds for example. The cost of acquiring a customer is ₹ 1,200-1,500. A lot of it has to do with collecting documents, verifying them, and going back to customers if there are errors. Now, if you do all of it online, the cost dramatically drops to ₹ 10.
This could change the economics significantly. If you are spending ₹ 1,500 to get a customer, you would use your commission to cover the costs. If your commission is 50 basis points (0.5%; a basis point is one-hundredth of a percentage point), you would approach only those who can invest a couple of lakhs. By default, you would exclude those who can invest only a couple of thousands. Sebi’s new e-KYC norms allow this digital migration, and it’s only a question of time before that actually happens.
The India Stack team often uses the example of Rajni, a persona they created to explain how India Stack can transform the financial sector and livelihoods. Imagine Rajni to be a street vendor. Every day she borrows ₹ 90 from a moneylender, buys vegetables and sells them at a mark-up. She makes about ₹ 130-140, and pays back ₹ 100 to the lender. She pays a usurious interest rate because the cost of servicing small loans is high—more importantly, because she has no other option. A new moneylender is likely to trust her even less and the rates would be just as high.
Now consider how her life changes with India Stack. A lender will know who she is, thanks to biometric authentication, can look at her transactions and take a call on her creditworthiness, send her digital cash, and get paid back in digital cash. Each transaction Rajni does leaves a digital trail, which she can use to borrow on better terms. From being locked in with a single moneylender, she can now choose from many and get better rates. Rajni can now borrow more and increase her income. This is not merely a technologist’s dream. A pilot is going on in Delhi right now.
Banks are already facing the heat in the US, where a whole range of start-ups are unbundling the industry. A recent CB Insights presentation on the future of fintech had a slide with a screen grab of Wells Fargo’s landing page with each of its offerings (loans, wealth management, insurance, banking) tagged to the plethora of fintech firms offering the same service.
In India, Aadhaar and India Stack will act as a catalyst.
Sharma says there are over 40 start-ups that are experimenting with India Stack right now.
In 2014, soon after Narendra Modi came to power, Nilekani sought a meeting with him to discuss Aadhaar. In the months leading up to India’s parliamentary elections in April-May 2014, the fate of the project seemed uncertain. Some politicians from Modi’s Bharatiya Janata Party swore they will pull the plug when they come to power. They saw it as yet another wasteful project.
Modi was keen to know how Aadhaar would help the government fix leakages. Nilekani felt that Modi had already made up his mind to go for the project. Gujarat under Modi was among the most enthusiastic adopters of Aadhaar.
Modi also saw the big picture—of how the trinity of Aadhaar, mobile phone and access to banks—could bring about a real change. As a result, his government pushed banks to open more accounts, even as it stepped on the gas on Aadhaar registrations.
Its impact is already visible—the government says it has saved ₹ 1,000-1,500 crore on cooking gas subsidy, mostly by removing duplicates.
Another typical case is PDS. A ration shop customer is tied to one shop—and has to accept its services, good or bad.
In Krishna district in Andhra Pradesh, which has seen Aadhaar-based reforms, people have more choice. The ration subsidies are tied to their Aadhaar number. They can get their rice, sugar, etc., from any PDS outlet. If an outlet gives them a poor deal, they vote with their feet, putting pressure on the entire system to get better.
When it comes to giving away cash—as in MGNREGS—the problem is even more complicated because cash is more fungible. The Aadhaar Payment Bridge System solves this problem. It’s now possible to credit the payments directly to a beneficiary’s bank.
What about their access to banks? India Stack’s answer is micro ATMs. A bank customer has to just go to one of these micro ATMs, typically a device operated by a banking correspondent, authenticate himself, transfer the amount from his account to the banking correspondent and receive the cash.
The direct benefits transfer can be applied to any subsidy—fertilizer, healthcare, education. It also provides a way to test an idea that is popular among economists—that the government has no business running ration shops or for that matter schools. Leave them to free markets and directly transfer money to the needy. The market will compete to get his money, lowering the cost and improving the service.
It’s not just about the poor. The broader idea of giving unique numbers and building tech layers around it can be applied to other areas as well. Take for example national highways, which the government hopes to fund predominantly through toll collections. But toll collections impose more costs on vehicles—there is waiting time and delays. Recently, a team from IIM Calcutta surveyed major routes in the country and concluded that the total cost of delays caused by check posts, toll plazas and checking by enforcement agencies is over $21 billion. Part of this problem can be solved by electronic toll collection.
In 2010, Nilekani chaired a committee that looked at this issue and suggested a system based on unique identity for all vehicles. A tag on the vehicle, sensors at toll plazas, a charging system (prepaid/postpaid) built on the principles of interoperability and other enabling infrastructure should solve the issue. The same system can be used within cities to manage traffic by applying congestion charges.
Technology is disconcerting to many. In part, because our views are shaped by popular culture. Films—Blade Runner, Terminator, The Matrix, take your pick—tend to depict machines as too powerful and dangerous.
Aadhaar, by bringing cutting-edge technology (biometric scanners, digital authentication) to such a large number of people, has kindled those fears.
Some fear new technology because it might turn out to be a dud. (This reflects a plot point of John le Carré’s Russia House—that missile technology is dangerous not because it’s advanced and precise, but because it’s not and it could misfire.)
As Aadhaar rolled out, there were many articles in the press that echoed this sentiment, and the concerns exist to this day. Aadhaar could end up excluding a lot of deserving people because the system might generate too many false negatives. Government services would not reach intended beneficiaries, because technology didn’t recognize them.
The standard answer from the people we spoke with was that the system is not perfect, and the glitches would be sorted out. Technology is improving fast. Just a few weeks ago, Samsung launched a tablet with an iris scanner, which is more accurate than fingerprints as fingers can get bald because of manual labour. It wouldn’t be long before every phone has one.
It would be unwise to dismiss technology based on how the first versions work. Information technology tends to get better at exponential rates.
The second set of arguments is based on the fear that Aadhaar could turn out to be a powerful tool in the hands of a state—an Orwellian dystopia where Big Brother watches you all the time. The revelations made by Edward Snowden showed that the state can and does indulge in such surveillance.
The big question is, should we pull the plug on the entire system because of such fears, or should we think of ways to fix the issues because it promises to make the lives of a large number of people significantly better?
While reporting this piece, one issue that came up again and again was around protection of data. Discussions with those involved in the project suggested that they were indeed agonizing over it. For example, in the early days there was a debate on how to deal with the spelling errors that are spotted by an applicant after the details are stored in the computer. Allowing the person doing the data entry to pull it out for correction would compromise data security. So, Varma and Jain wrote code to append the corrections. It was more complex, but it protected the system’s integrity. Over time, Aadhaar has added more checks and balances. Every time your identity is authenticated, you get an alert. You can also block your biometrics. The consent layer will give users more control.
India also passed a law earlier this year proposing punishment including fine and jail term for breach of privacy.
It cannot, of course, be perfect. Hackers will never stop trying. Governments can always rationalize overreach citing national security. Implementation of law is never easy. But pressure from activists and users will likely keep the offenders in check.
There is increasingly more interest in the project. Sharma remembers organizing an event around India Stack about a year ago. The initial plan was to host around 150 people from start-ups, but there were over 320 who wanted to participate.
There is a lot of action but most of it is happening in the meeting rooms, in front of computers and between the ears of those who see a lot of potential. And that’s how disruptions happen, step by step.