For anybody seeing GenAI in action for the first time, it is nothing short of magic! While we have talked about the challenges, esp. for India and startups in the earlier articles, nobody doubts that it is showing us the future. As they say, the best way to predict the future is to create it! As a fund, it is exciting to have the opportunity to participate in our small way in this, by supporting and promoting the entrepreneurs, who are the ones really shaping the future.
Just a recap of our thesis at Pentathlon. With our background in B2B Tech entrepreneurship, we focus on early revenue stage B2B Tech Indian startups and help them go global. We don’t invest in the Tech itself but in how the Tech disrupts use cases across industries.
The previous 3 articles have articulated our thinking about this fast-evolving AI and GenAI landscape, which is THE Tech today How we navigate it is what I cover in this final article.
What will not change?
The advent of GenAI proves once again that Tech is the biggest disruptor. And AI and GenAI are probably the biggest of them all. So our focus on how AI disrupts use cases across various industry verticals remains the same. We will not invest in companies building the AI or GenAI Tech infra.
We will continue to focus on revenue and the path to profitability and not get carried away by the hype and the valuation game, just because they are using AI or GenAI.
And of course, the focus on founders will remain. Apart from their background, we always look for their unique insights, in using GenAI most effectively and their expertise in the domain where they are applying it.
While GenAI seems like the preserve of the large Tech giants, as shown in the previous articles, there is a great opportunity for B2B companies to succeed in narrow niches across industries.
An excellent example of this is one of our first investments in Fund 2 – Vodex.ai. Their product is GenAI-based bots that autonomously call and qualify sales leads. The bots are trained on the very enterprise and product-specific data which may not be available with Tech giants. They can hold 2-way conversations and even handle interruptions! The years of experience the founders spent in this space were crucial in building Vodex.
What will change?
In the software world, code drives everything, and companies own their code. In the AI/GenAI world, data drives everything, and who owns the data becomes most critical.
Enterprises may want to use this data they own (which even the Tech giants don’t have) to create competitive advantage for their products in the market-place, or internal solutions to capture their domain expertise. The huge opportunity that B2B start-ups have with data owned by the enterprises (which even the Tech giants don’t have), is building these GenAI solutions for enterprises., But this seems a services rather than product play for the startups!
So, as a fund, counterintuitively, we must figure out our models to fund service companies. For fundsAs a fund, services were always a no-no, as services businesses cannot scale as fast as product businesses. However, if there is one country where service companies have scaled fast, it is India.
At the moment, all enterprises – large and small – are investing a lot to build Proof-of-concept (PoC) GenAI apps using their data. So, this is a large but fiercely competitive space.
To stand out in the competition, some start-ups claim to build their own foundation models (models trained from scratch), instead of building on top of available LLMs. But do they have the data needed for that with them?
In India, there is great availability of data from our world-leading Digital Public Infrastructure and Goods (DPI/DPG), such as Aadhar, UPI, Fastag, ONDC etc.. But it is owned by the Government and potentially available to everybody.
We, as a fund, must figure out who can win this competition.
Who will be the winners?
Although GenAI does make it easier to do so, most enterprises are unlikely to try building the solutions themselves, as they will not immediately have the talent or the focus for this “non-core” activity.
Because of their urgency to build their GenAI applications, Enterprises will choose those GenAI specialists who can deliver fast. To do so, these GenAI specialists will build their solutions on top of their ready platforms of agents doing various GenAI tasks for prompt engineering, and fine-tuning of Small Language Models, RAG or GraphRAG.
But GenAI cannot do everything and need not be used everywhere. Dto do this well, deep knowledge of the domain of the enterprise is required, so that business rules can be applied in these techniques for effective solutions.. Further, GenAI need not be used everywhere. If well-established technologies, and AI techniques are available they must be merged judiciously with GenAI. Only with this domain expertise can they understand the business problems and come up with the most effective and cost-efficient GenAI solutions, by merging , using business rules (algorithms) and other formidable AI techniques judiciously with GenAIand leveraging their business knowledge. This may be crucial when GenAI is so data, compute and hence resource-intensive, although new innovations may improve that situation in the future..
Deep knowledge of the domain of the enterprise is required, so that business rules can be applied in these techniques for effective solutions. Further, GenAI need not be used everywhere. Well-established technologies, and available AI techniques must be merged judiciously with GenAI for cost-efficient solutions. This may be crucial when GenAI is so data, compute and hence resource-intensive, although new innovations may improve that situation in the future.
So, the winners would be those who enterprises will choose their GenAI solution providers based on their blend other technologies with GenAI expertise, have ready-made platforms that accelerate building the solutions along withnd their deep domain expertise.
And that brings me to another counter-intuitive conclusion – we are going to see many founders in their 40s instead of their 30s! As the technology building itself becomes commoditised, founders who have spent 15-20 years honing their domain expertise will be the winners.
The younger founders can find their opportunities by building expertise in upcoming domains such as Climate and Renewable Energy. Or with their exceptional tech chops, they can even target horizontal functions in enterprises (such as HR or Admin), building cross-platform solutions to beat the solutions from the Tech giants that work only within their platforms. Or there is an opportunity for them to help product companies keen to incorporate GenAI in their products, either with services or tools.
Since a lot of jobs are going to be lost and new jobs created, and this is esp. true for India, the leapfrogging ability of GenAI for training and skilling at speed and scale, is a big area of opportunity. Again it will be very competitive. The winners wcould be those who focus on specific domains, using GenAI to teach domain knowledge to Techies, or tech-enabling field workers in the domain (such as nurses and teachers who provide the human touch required for last-mile solutions). In general, all new technology users (e.g. non-English speaking users) have to be trained in GenAI to reap the benefits of the new technology.
In conclusion
We see great opportunities for our thesis of catching B2B Tech players in India early and helping them go global. They must have domain expertise in their niche, access to specific data and the expertise to build GenAI solutions faster, with product-led services. We will look for such startups across verticals.
We hope new technologies will emerge that will address the challenge with GenAI of huge data, computing and energy requirements. But in the meantime, startups that can blend other technologies and domain knowledge with GenAI to address this challenge would be the winners.
Startups building products for faster skilling in various areas of technology creation as well as end-user solutions will be great opportunities for us, as their need will only keep going up.
With this, we come to the end of our 4-article series on AI/GenAI. We would love to know what you thought of them. Our interaction will help all of us navigate the GenAI revolution better.
