Two years after Vishal Sikka stepped down as the CEO and executive vice chairman of Infosys, an Indian IT services company, he has launched a new venture in artificial intelligence. Vianai, a Palo Alto, Calif.-based startup, last week announced its arrival with $50 million in seed financing.

Sikka believes AI has the potential not just to transform business but also to “amplify humanity,” as he puts it. He sees AI as a force multiplier that can tackle issues ranging from climate change to self-improvement. “I would love to see tens of millions of people to be able to build intelligent systems, and billions to be able to bring basic intelligence into anything that they do,” he says.

Sikka, who holds a Ph.D. in AI from Stanford University, demonstrated the Vianai platform at a keynote address at the Oracle OpenWorld conference on September 17. Before his tenure at Infosys, he spent 12 years at German software company SAP, where he was last a member of its executive board.  Sikka also serves on the supervisory board of the BMW Group and as an advisor at the Stanford Institute for Human-Centered Intelligence (HAI).

Vianai’s board of advisors includes, among others, John Etchemendy, former provost of Stanford University; Henning Kagermann, former chairman and CEO of SAP, who is now chairman of Acatech, Germany’s national body vested with technology innovation; Alan Kay, Turing Award winner and computer science pioneer; and Indra Nooyi, former PepsiCo CEO who is a board member at Amazon and Schlumberger.

An edited transcript of the conversation follows.

Knowledge at Wharton: When we last spoke in February 2018, you said that you were passionate about building technology that amplifies human potential. You also mentioned that you were working on a venture that would explore ways in which AI can help to elevate humanity. Why did you choose to go down this path?

Vishal Sikka: I always have been passionate about the idea of technology being a human amplifier, something that improves our ability and makes us see more, do more, and be more. Both [my wife] Vandana and I are deeply committed to this idea. At a time when you hear so much about technology’s negative effects and unprecedented ability to scale, and to propagate things in a negative way, why can’t we build technology that improves us?

In the course of last year, Vandana and I thought and talked a lot about this and concluded that the solution lies in startups. She is working on her own startup, in the consumer area. I thought that after working in large companies for 16 years, and leading them in many ways, it was time [for me] to go back to the drawing board.

Knowledge at Wharton: These days much conversation is focused on AI as a destructive force that will eliminate millions of jobs in many industries. But you seem to see it in a different way, as a constructive force, as a force for good. Could you explain that perspective?

Sikka: Most of the negative effects that we see of technology have almost nothing to do with AI. They have to do with the scale and the reach of digital technology in our lives. A lot of the negative press [on AI] stems from the unique nature of the technologies we have today. A lot of it also stems just from misinformation and a lack of understanding. That itself is an attribute of technology that must be addressed.

Knowledge at Wharton: You’ve led large organizations such as Infosys – a company with more than 200,000 employees – and before that at SAP, also a very large organization. What have been some of the lessons you have learned, in the time that you’ve taken to get started with the new venture?

“There is just a sheer amount of legacy that one has to deal with in a large company, which means that your ability to do things from scratch and going to the drawing board is much tougher.”

Sikka: I was fortunate to participate in two transformational journeys in both those companies. Large companies have massive scale, and they have many benefits because of that scale. Every successful company does things in big ways, reaching large numbers of customers and being able to affect the work of massive numbers of people.

[However], they also face significant challenges in making transformation happen. The burdens of legacy are real. There are existing markets, existing businesses and existing processes that have a huge amount of inertia and entrenchment that must be dealt with. There is a sheer amount of legacy that one has to deal with in a large company, which means your ability to do things from scratch and going to the drawing board is much tougher.

In the fall of 2017, after I left Infosys – I had just turned 50 – I was thinking about the next 25 years. I imagined myself as 75 years old and looked back. I realized that 25 years ago from that point, which would have been 1992 or 1993, Google wasn’t around, Facebook wasn’t around, [and neither were] Uber, Tesla and Airbnb. Steve Jobs wasn’t back at Apple yet. Amazon wasn’t around or had barely just started. All that happened in the past 25 years. And I thought, the new things that will define the next 25 years aren’t around right now.

David Patterson and John Hennessy, who won the 2017 Turing Award, in their Turing lecture last year talked about how it’s a fundamentally different age now of semiconductors. The traditional Moore’s Law is being replaced by domain-specific architecture for chips, especially for AI.

What would you do if there was no burden [of legacy]? That means that you have to abandon the benefits of scale and the benefits of working with super large companies. You have to do something literally from scratch. I decided to do that.

I have a Ph.D. in AI and have had the opportunity to work in large companies in enterprise software and services. I understand transformation in a way that few people do because I have lived through two large-scale, successful transformations. I thought it was time to take advantage of the unique gifts I have been given. That’s how I ended up here.

Knowledge at Wharton: What did you imagine your new venture to be?

Sikka: I always thought that teaching and doing are great ways of learning, so I went back to school. I taught two classes last year on AI – one in California and one in China. I also worked on a few prototypes over the course of the year and realized that there was both a tremendous weakness in the current state of AI, and also a tremendous opportunity to do something better.

Last week, there was the example of an eighth grade test that a system from the Allen Institute for Artificial Intelligence passed. If you look at the SQuAD leaderboard (Stanford Question Answering Dataset), there are several systems that perform better than humans. You hear all these wonderful examples of things that AI techniques can do, but on the other hand, there are also remarkable weaknesses. The systems are incredibly powerful perceptual engines, but they have no idea about what is happening in the world, what is the meaning of various concepts or entities, the semantics, of how things work. The techniques have significant weaknesses.

To further exacerbate that inherent limitation, you have this massively asymmetric situation where a very small number of people understands AI techniques. There are roughly 35 million or so programmers in the world, but the number of machine-learning engineers is in the few hundreds of thousands. Tencent did a study which showed that there are about 300,000 AI engineers in the world. These are people who can build machine-learning applications. The number of people who can explain to you how particular algorithms work is only about 20,000.

We can do better. We can bring AI to everybody, we can build AI systems that are transparent, easy to explore, easy to understand, and explainable in the sense that the developer understands. The technique itself is inherently opaque in these deep neural networks but you can make it easy for the developer, and for the designer of the system to understand what is happening in their system by making the system exploratory and transparent. We can dramatically simplify the footprint that these AI systems carry, and we can bring problem-finding into the mix, by bringing business people, the domain experts and IT people into the construction and evolution of intelligent systems, not only the data scientists and the AI engineers.

I decided I wanted to build a company that on the one hand helps enterprises build AI systems that can help them become better both at growth as well as operations. On the other hand, I wanted to build that using a new platform that is built from scratch, that is empowering, accessible to tens of millions of developers, system designers, analysts and so forth, and on which enterprises can reliably and transparently build dozens of applications. We spent the last two years getting to that point.

Knowledge at Wharton: Where does the name Vianai come from, and how will your company and its platform work?

Sikka: In Bali, the firstborn child is called Vian. That’s where Vian comes from. It’s also a beautiful word that means “full of life.” Vianai is the name of the company. It is aimed at bringing AI at a massive scale to enterprises in a purposeful way. I believe that over the next 25 years, we are going to have systems of intelligence, just as starting from the 1980s there were systems of record and then starting from the late 1990s we have had the development of the systems of engagement like CRM and marketing systems. The next wave will be systems of intelligence where enterprises will have dozens of intelligent applications running. When you are certain that over the next 20 years enterprises will have dozens of intelligent applications, then what is the platform for building and delivering those in a way that takes into account the emerging heterogeneity? That is the platform we are building. In parallel, we are also working with some of the biggest banks and manufacturing companies to build solutions and systems for them while simultaneously creating our platform.

Knowledge at Wharton: It sounds like an ambitious vision. Many AI startups are being funded right now because of all the hype around it. Who else is active in this space, and how do you differentiate what you are trying to do from others that are trying to bring AI to the enterprise?

Sikka: There are several thousand AI startups…. I’m sure they’re all unique and doing fantastic things. We have assembled a team that has a unique background in that we have a deep understanding of enterprises, and of transformation in enterprises. We have a unique understanding of not only the current state of AI but also where AI is coming from. We understand the strengths and weaknesses of previous AI approaches and [also those of] the current AI approaches. We have the perspective to transcend individual trends and build things that can be brought to enterprises at a very large scale.

“I would love to see tens of millions of people be able to build intelligent systems, and billions be able to bring basic intelligence into anything that they do.”

We tend not to define ourselves in terms of whatever else is going on, but rather in terms of the problems and the opportunities that we have identified and going after those. In that sense, what we are doing is significantly different from what anybody else is doing, whether at a startup, or even at a large company. It is about dramatic improvement in simplification and explorability. We can simplify the expression and the execution of these systems massively – by hundreds of times, if not more.

Knowledge at Wharton: You have raised $50 million in funding and you have an impressive group of advisors. How did that come about?

Sikka: Over the Christmas holidays last year, we decided that this is the path to go forward on, [that we had done] enough experimentation and [should] invest in our roots for real. So, we built a some models and realized that $50 million should be enough to get us to profitability – so I raised $50 million.

No plan will ever survive the test of reality. So, you need advisors who can guide you, help keep things in balance, and provide perspective. We assembled several people, [including] my friends and teachers and mentors over the decades who bring a combination of technology, business, operations and finance.

(Editor’s note: Vianai’s advisors include Henning Kagermann, former chairman and CEO of SAP and chairman of Acatech; Alan Kay, Turing Award winner and computer science pioneer; Divesh Makan, founder of Iconiq Capital; Indra Nooyi, former PepsiCo CEO and member of the boards at Amazon and Schlumberger; Sebastian Thrun, CEO of the Kitty Hawk Corporation, and co-founder and chairman of Udacity; and John Etchemendy, former Stanford University provost and co-director of its Human-Centered Artificial Intelligence Initiative.)

Knowledge at Wharton: You have been a proponent of problem-finding rather than problem-solving. How will you find the problems on which you want Vianai to focus?

Sikka: That is in many ways at the heart of the work that we do. AI itself will help us solve problems. We want to enable problem-finding with our platform as a core part of building an AI system. This sounds antithetical because AI is about problem-solving. We want to make it extremely easy and exploratory for developers and for designers to build AI systems.

But we also want to make sure that they are building it [to solve] problems that they clearly understand, where they can collaborate with business owners in a seamless way. Usually, what happens is the systems you build are so obtuse that no businessperson will ever look at them. We want to bring that awareness of problem-finding in our platform as the root of the activity of building a system.

For example, we’ve been working with a large bank on identifying failures of certain transactions that go through their systems, and these failures are worth tens of billions of dollars a day. If you are a manufacturing company, even if you make-to-order, the issue of working capital is fundamental to your operational efficiency and financial efficiency. In companies with large numbers of physical assets, this directly impacts their bottom line and their stock price. So, working capital management using AI is one of these areas we are working on, along with others.

Knowledge at Wharton: What challenges did you face in getting Vianai from an idea to where it is now? How did you address them?

Sikka: Ultimately, organizations are no more and no less than the people that are in there. And so, getting the right group of people has been one of the most important things. Already there is a mad scramble for talent in AI. And, on top of it are additional constraints [related to finding people with] a deeper understanding of the history of AI, understanding of enterprises and so forth.

“We want to bring that awareness of problem-finding in our platform as the root of the activity of building a system.”

So we abandoned the idea of bringing everybody together in one location from the start. Even though we are quite small, our team is already very distributed. Our headquarters are in Palo Alto, but we have employees in Southern California, Utah, Idaho, New York, Boston and Seattle, and also in Israel. Every once in a while, the team comes together and we make heavy use of workspaces that break down these barriers of space and time. We are a team of more than 30 employees, consultants, contractors and advisors now and growing rapidly.

Knowledge at Wharton: What are Vianai’s most immediate priorities? What goals have you set for the next 12 to 18 months?

Sikka: Some of the priorities are getting the platform built and delivering customer success. [Also important are] intellectual property and protection, in the environment that we live in. We have this dual desire to build things openly and yet, we are fully aware that there are powerful companies [in our space] and we want to make sure that we have the intellectual property protection.

I took financial uncertainty out of the equation by raising a large seed round. [Startups] typically don’t do $50 million seed rounds, but with this we don’t have to worry about funding. Of course, we run a financially prudent ship, and Divesh [Makan] being there helps a lot as also Henning [Kagermann] and Indra [Nooyi]. We have revenue coming from big companies, but still, we don’t want to worry about money.

The burden of going back to the drawing board is that you have a tremendous need for diligence and detail. In large companies, things have worked for decades, there are processes in place, and you can count on them. In a small company that is not the case. The fact that you chose to go back to the drawing board means that nothing is there; everything, including the mundane things, must be built up. That means that you need attention and to make sure things don’t fall through the cracks.

Knowledge at Wharton: What are the main risks that Vianai faces? What have you done to try to mitigate those?

Sikka: We have clearly understood that what we are doing is a very promising, very big opportunity. The challenges are in execution. And that is where the difference between knowing the path and walking the path [matters].

Knowledge at Wharton: Given your passion for amplifying human potential, how will you measure Vianai’s success? Financial milestones clearly are important for any enterprise, especially a startup. In addition to those, what metrics would you like Vianai to be measured by? How does social impact factor into your plans for judging your success?

Sikka: Beyond financial success, we want to make sure that we measure the impact that we have, both in terms of the number of people whose lives we impact by the work that we do and by working on important problems of our time. Climate change, [for example], is a far more severe and problem than we realize and a far more immediate problem than we realize. Similarly, there are problems that move the needle on societal improvement. Personally, I am very interested in making a difference on this front in India, the country of my birth, where AI can either be a vast disruptor or an equally vast opportunity, over the next few decades.

There is a mystique associated with AI. The reality is that there are techniques, there are techniques that we can learn and use to improve things, to improve ourselves. I would love to see tens of millions of people be able to build intelligent systems, and billions be able to bring basic intelligence into anything that they do. That is not hard. We can do that. We can enable that. In fact, that is one of the reasons why I started this company.

Education is one of the key parts of our work at Vianai that we haven’t started yet, because we are still building the platform. I have started to do a master classes on AI techniques with some of our clients. With our platform, we want to make it easy for people to learn these things.

Knowledge at Wharton: What is your most ambitious dream or hope for Vianai for the long term? You talked about the changes that have taken place over the past 25 years, and how it’s almost impossible to imagine what will be around over the next 25 years. Where would you like Vianai to be over that timeframe?

Sikka: I would love to see people start to bring intelligence capabilities into everything. We can amplify our perceptual abilities, our processing abilities, with the techniques that we already know and by demystifying them. To see such a pervasive delivery of intelligence capabilities in the world, in all walks of life, and to have that be built on our platforms and our tools — that would be worthwhile.