‘We back genuine AI startups, not wrappers’: Google accelerator programme lead | Technology News

Whereas the programme for the apps-only cohort of 20 startups runs from July to November, the AI-first cohort take part from September to December.

“The COVID-19 pandemic noticed a trough as a result of the variety of folks beginning up and the variety of corporations on the market was form of plateauing out, however for the reason that AI shift, we’ve got seen an enormous spike within the variety of startups who apply to Google’s accelerator programmes,” Paul Ravindranath G, senior programme supervisor, Google for Startups Accelerator India, mentioned. “The explanations are, individuals are in a position to utilise AI instruments and AI fashions to allow their options in a a lot faster and a quicker manner,” he added.

In an unique interview with The Indian Specific, Ravindranath additional mirrored on Google for Startups’ journey thus far, the state of India’s startup ecosystem, the challenges most founders when constructing AI options within the nation, and extra:

How has Google helped form India’s startup ecosystem, and what impression do you suppose it has had on startup founders throughout completely different sectors?

Ravindranath: I’ve had the privilege of being related to the startup programme from day one for Google in India. At first, it was only a one week-long mentorship programme. We might invite about 10 corporations collectively within the workplace and work together with them about their wants and a number of areas reminiscent of expertise, development, advertising, and so forth.

2015 was extra of the period of “Me Too” form of startups. I keep in mind Google was the foundational companion for the ten,000 Startups initiative by NASSCOM. What began as supporting founders to develop and innovate has snowballed during the last decade into so many significant areas of worth. Over time, we’ve got seen India’s startup ecosystem evolve not simply when it comes to quantity of concepts, however in fixing huge population-scale challenges on the again of platforms like Google Cloud and different AI instruments we’ve got on the market.

Immediately, we stand at over 237 corporations supported throughout 17+ cohorts. Over the past 4 to 5 years, we’ve seen startups constructing deeply on AI and leveraging expertise in unbelievable methods. These corporations have gone on to be actually profitable as nicely, elevating greater than $4.5 billion in funding, with a 96 per cent survival fee. It’s an actual testomony to the truth that these corporations are fixing actual issues and doing it in a manner that permits them to maintain and develop, with help from platforms like Google which are in a position to again them.

How does Google determine which startups make it into its accelerator programme? How has the choice standards developed through the years?

Ravindranath: Within the early days, we have been startups in what was actually the period of “Me Too” corporations. We had a lot of functions, a number of on the thought stage, wanting help from our programme. In the course of the COVID-19 pandemic, over these two to 3 years, we didn’t change off this programme. It remained very a lot lively, we pivoted it utterly on-line.

That was a difficult interval for the Indian startup ecosystem. The wants of corporations and the way in which we chosen them additionally modified. In our cohort on the time, we picked corporations targeted on healthcare or entry options, and we noticed how tech enabled these options.

The pandemic was additionally an attention-grabbing section the place startups in India noticed an entire shift of their consumer base. My dad, who’s over 70, had by no means used an app apart from YouTube on his cellphone, and by the tip of the pandemic, he was ordering groceries on-line, consuming on-line content material, and totally partaking with the e-commerce ecosystem.

Within the final 4 years, with the leaps we’ve got made in AI capabilities, we’ve seen an entire platform shift. The variety of startups utilizing AI to resolve population-scale issues has grown quickly. We’re seeing sturdy tendencies in healthcare and fintech, the place AI platforms are getting used to deal with these challenges.

This, nonetheless, presents one other problem for us. How can we select which corporations are really constructing with AI? We see many “mild wrappers,” as we name them, startups or aspiring founders constructing fast UIs on prime of present massive fashions.

That’s not one thing we wish to help, as a result of these are simple to spin up and never essentially sustainable options. We go a degree deeper to see how startups are tackling attention-grabbing drawback areas utilizing proprietary knowledge, making use of AI and ML to that knowledge to personalise and enhance accuracy. That exhibits promise and provides us an opportunity to work with them to make their options even higher.

How are you seeing Indian startups use AI instruments to construct options for India?

Ravindranath: I’d like to possibly recount two or three which have actually caught with me through the years as being really impactful. One in all them is a startup known as Niramai, which is Bengaluru-based. Gita Majunath is the founder. They’re within the early breast most cancers detection area. So think about with a thermal picture of a girl’s chest, and making use of their proprietary AI mannequin with a 98-plus p.c accuracy, they’re in a position to detect early-stage breast most cancers.

One other startup that I actually love working with is that this startup known as Wysa. Wysa is a startup that really operates within the psychological well being area. It started as a easy AI chatbot that you would speak with for those who’re feeling down and low. However the AI chatbot has developed, and clinically cleared to have the ability to get you out of a tricky spot.

Kidzovo is an app that we labored with in our cohort final 12 months. It’s a nice instance of a startup that provides a layer of interactivity over a video. So let’s say a younger baby is basically spending quite a lot of display time, consuming quite a lot of YouTube movies, cartoons, Peppa Pig, and the like. With AI, Kidzovo is ready to add a layer of interactivity. It can pause the video the kid is watching and ask them an interactive query: “Are you able to faucet on the nostril of the pig? Are you able to inform me the variety of sharks on this Child Shark video? Faucet on all of the sharks.”

One other one I can cite is an organization known as Toon Sutra. They’re utilizing Google’s Gemini fashions, Veo and different video-generation fashions in very attention-grabbing methods to create India-focused comics.

As AI turns into extra central to startups, what are the largest challenges you see for early-stage founders in India?

Ravindranath: One of many greatest challenges is compute. A number of AI founders working within the area search for help in subsidising the associated fee concerned in constructing AI options. Usually AI options contain quite a lot of knowledge processing, and as founders, they undoubtedly attempt to strike the steadiness between a fast AI resolution that could be very pricey versus one thing that’s cost-effective and delivers the appropriate consequence.

Doing this cost-benefit evaluation as founders and CTOs is basically essential. This is likely one of the areas the place, as an accelerator, we assist founders architect for scale. We’ve very particular modules in our programme that consider the package deal structure, consider the fashions and their accuracy, and decide how costly an answer might be, and make it extra environment friendly.

One other problem I’d like to focus on is undergo this present AI increase and do it responsibly. Knowledge is out there in all places, and folks can rapidly construct options round it. However how do you construct thoughtfully?

When you’re constructing within the healthcare area, in finance, or in tech, as founders it is advisable develop a mindset of accountable innovation. How do you deal with the info supply responsibly? How do you’re employed with that knowledge and knowledge? How do you handle the interplay between folks and AI? I feel that’s actually essential for founders to give attention to.

With the rise of vibe-coding, how do you suppose this adjustments the way in which Indian startups are constructing, experimenting, and scaling new merchandise?

Ravindranath: There’s now an enormous improve within the capability for entrepreneurial minds to prototype options. So anyone right this moment who has an ideal thought, prior to now, they must determine go about fixing for that: determine getting any person else, like a co-founder, to come back on board, suppose by way of it, mannequin the answer, do UX analysis, go on the bottom, speak to potential customers, create a minimal viable product, work on preliminary traction, get consumer suggestions, and undergo a number of cycles concerned in that innovation.

However right this moment, with platforms like Firebase Studio or different vibe-coding platforms, we’re in a position to get to that preliminary prototype or fascinating MVP in a really, very quick time frame.

So the innovation cycles are a lot shorter in right this moment’s day and age. Now, that’s good as a result of with quite a lot of the concepts, we’re in a position to fail quick and transfer ahead. However on the identical time, it creates an issue the place quite a lot of help programmes on the market try to design filters, which is a skinny wrapper, which isn’t an actual AI resolution. As an ecosystem, we’re all studying to search out these gems, however general, it’s a optimistic motion in direction of innovation.

On the identical time, I feel it must be approached with warning, as a result of as founders are constructing, one of many questions that I get is: “Hey, the place do I discover some actually nice AI and ML or massive tech engineers to assist me construct out this product?”

So whereas vibe-coding is nice for attending to that preliminary imaginative and prescient and prototype, to construct a really sustainable, production-ready resolution requires funding in foundational data, having a group with deep tech experience to essentially construct that concept out into one thing production-ready that may scale at inhabitants scale, particularly in case you are fixing for India.

Any AI startups supported by Google which are specializing in constructing foundational AI fashions or native AI apps?

Ravindranath: One firm, known as Deep Imaginative and prescient Tech, a Bangalore-based firm, is working within the dwell translation, signal language area. They’ve an avatar on their cellphone and, as I’ve a dialog with somebody who can not hear, the avatar really indicators to them in actual time translating and signing again to them in Indian Signal Language. And there are a number of signal languages. That is an incredible innovation within the accessibility area.

One other startup is known as Iyaso. This startup is within the speech remedy area. The founder himself has a stammering drawback, and he developed this app to really assist individuals who stammer overcome that incapacity, with an ideal diploma of success and medical validation as nicely. He technically has one of many largest databases of stammering knowledge and is approaching the issue solely with an AI mindset and methodology.