A stitch in time saves nine... T-shirts. May be more!

HR Tech with AI can do more than clearing away HR’s calendar and bandwidth for more important tasks. Siva Kumar—CTO, Zaggle tells Pratima H how humans are still, and in fact, more relevant with the explosion of AI in HR functions

What made you foray into HR and payment areas with AI?

This question can be answered with an interesting anecdote—an employee of an MNC approached his HR Head with 35 T-shirts in his hand from events that took place over two odd years. He explained how those T-shirts were a waste for him because he wouldn’t wear them. That’s when the company realised, they needed digital rewards for their employees to enable them to benefit from rewards.

Zaggle was born to fulfil this growing need for digital rewards. We first came up with employee benefits, then with employee reimbursements and then forayed into expense management. We are a payments company working in the areas of Financial Tech, HR Tech and Bank Tech and the most amazing part of HR Tech is that it deals with People and mapping AI to HR. We found our sweet spot in work force performance management, the modelling we did using predictive analysis (AI) as to what could be the career growth of an individual after he joins the company. This modelling is being appreciated by a lot of our beta clients and we hope to release it soon.

Can AI soon turn into a hygiene factor for the industry? Any way in which it can also shape into a strong revenue area?

If we see how the world has evolved around data in the last 20 years, you would notice that, earlier, the problem was the quality and quantity of data. Anybody who made data- driven decisions would say I don’t have real-time data on hand. It was all based on historical data. Now that we have truckloads of real-time data, the skill required is (A) ‘What to look at vs. What not to look at’ and (B) How would you use that data. And that’s how companies with the exact same data set end up making completely diametrical decisions.

The game has changed around raw data to data intelligence and data analytics and what you do with that data. Payments—that is of course a very important aspect and the data we have will help us develop solutions for our users. If you are not thinking AI into your products, you would soon be out of business. So, it's already becoming hygiene. AI has already replaced the lower-end mundane jobs and is increasingly becoming a revenue generator or a cost saver. What we feel is that as we keep advancing in the product life cycle, AI will, possibly, become a major source of revenue rather than the cost cutter that it is majorly used as today.

It will replace the humans at the lower-end with the bots, but if you look at where the world has reached in terms of Sophia, the humanoid, I am pretty sure that we will be able to do most of the back-end analysis and analytics job with AI in the near future and sell the same as services to the other companies. It may, or may not, give direct revenue, but as we solve customer problems and simplify their lives by using AI, it would be adopted and loved by our customers and, hence, would indirectly give us growth.

Tell us something about how AI helps in expense management and reward functions? Do Propel and Save products use Deep Tech?

Yes, we do use deep-tech for both Zaggle Propel and Zaggle Save, which are integrated platforms for workforce reward and recognition, employee engagement, and state-of-the-art AI-powered performance management. AI is deep into our products and is present everywhere.

Can you explain with any examples?

Today, as part of the accounting process, you would have to submit the invoices to support the expenses that employees incur during their day-to-day operations. Imagine as an employee, I would need to read the bill, enter merchant name, date/time, amount spent including tax break ups etc which is pretty labour-intensive. But we have almost done away with that laborious process. The moment a bill is uploaded on Save, we convert the bills uploaded as images in to text by using various AI techniques like OCR, entity extraction, NLP etc. So, in real time all the information is converted to text.

On top of that, if the accuracy of the automation goes beyond a threshold figure configured by employers, we disable editing of the expense by employee. So, as a manager who is approving this expense need not figure out if the merchant name really reflects what employee is claiming, whether the date and time the expense happened and the amount claimed all are matching with the bills. It is a such a joy experienced by both employees and managers. For the rewards and recognitions management program, we offer rewards cards for employees and we have also tied up with more than 8500 merchants such as Abbot, Vodafone, Samsung and many more.

How crucial are moves like Open Banking, CMA guidelines and industry consortiums for the industry?

Open Banking is a beautiful concept and has been around for more than a decade around the world, but it hasn't taken off yet especially in India because it requires the necessary surrounding-infrastructure for it to function. But with the advent of UPI stack, it took off suddenly and now the number of use-cases being built around it is huge. With standardisations by means of guidelines and consortiums, it would become far easier for FinTech players to not worry about the bolts and nuts of how to build but they can rather focus on what to build.

Another important reason is to draft uniform regulations or codes of conduct that will govern individual companies throughout that industry, and to set standards for the licensing of professionals within an industry.

Is the impact of new APIs going to be strong for new revenue streams?

Application Program Interfaces (APIs) are basically interfaces which the third party can use to build a service of its own. We do assertively believe that the impact of these new APIs will be much stronger for the new revenue streams. The new APIs make it easier for different technology platforms and applications to interact with each other in a commonly-understood language. Furthermore, APIs offer companies the opportunity to increase the revenue, ease down the integration of backend data applications, adopt innovative solutions, reach a wider audience, and effectively support their sales and marketing activities.

AI's success depends on data—would players, including arch rivals, come together to expand the scale and depth of AI data?

I believe that AI is still at its nascent stage in India. As a thumb rule, I strongly believe in Andrew Ng's philosophy that we can only do those things with AI that a human can do in a second or less. To point an example; in HR, we can probably automate if an employee is happy, anxious, worried by taking a picture of an employee and feed into AI and it would pretty accurately be able to predict the emotional state. But, AI at its current state of development, is not good enough to go beyond that. For example, based on the emotional state, an HR manager can have deep conversations understanding the situation that employees go through which would take minutes or even hours to figure out the issue with the employee and then come up with empathetic suggestions. Here, AI is not even close though there is extensive research going on to match human capabilities.

Any challenges that AI faces when it comes to matching the quintessentially-human factors in HR functions?

One of the major challenges is that using AI while recruiting lacks the human touch in the entire hiring process. Another one is that reliability may still emerge as a question and prove to be a matter of concern. Since the data-learning process may still have numerous inconsistencies and flaws in it, reliability stays as a major issue.

What shifts, if any, are happening to embrace the impact of new concepts like Blockchain, AI, BaaS (Banking as a Service), Digital-only banking, FinTech-based financial inclusion?

We see a world where every company is embracing new FinTech products that offer all banking services and tons of additional features that banks cannot and should not imagine. With banking data and services opening up, FinTech players would come with so many innovative and disruptive ideas by combining AI/Blockchain along with massive banking data that we cannot imagine any business or even individuals going to physical banks anymore in future.

That new innovations are set to have such a major impact on retail banking in the next few years; and this shift suggests that regulators have come to a point where they can allow these technologies to flourish. With more governments, regulators and even businesses embracing these new concepts, it seems certain that it’s only a matter of time before AI-based business on a mainstream level becomes a norm.

Can AI help evade bias and abuse—especially in HR?

We humans make sense of the world by looking for patterns, filtering them through what we think we already know, and making decisions accordingly. When we talk about handing decisions off to AI, we expect it to do the same, only better. Harmful human bias, both intentional and unconscious, can be avoided with the help of AI, but only if we teach it to play fair, and to constantly question the results.

Humans are hindered by both their unconscious assumptions and their simple inability to process huge amounts of information. AI, on the other hand, can be taught to filter irrelevancies out of the decision-making process, pluck the most suitable candidates from a haystack of résumés, and guide us based on what it calculates to be objectively best rather than simply based on what we’ve done in the past. In other words, AI has the potential to help us avoid bias in hiring, operations, customer service, and the broader business and social communities—and doing so makes good business sense.

Is that where humans become more relevant?

Humans must tell AI what they consider suitable, teach it which information is relevant, and indicate that the outcomes they consider best—ethically, legally, and, of course, financially—are those that are free from bias, conscious or otherwise. That’s the only way AI can help us create systems that are fair, more productive, and ultimately better for both business and the broader society. Today, AI excels at making biased data obvious, but that isn’t the same as eliminating it. It is up to human beings to pay attention to the existence of bias and enlist AI to help avoid it. That goes beyond simply implementing AI to insisting that it meets benchmarks for positive impact.