1Mby1M Virtual Accelerator AI Investor Forum: With Jishnu Bhattacharjee, Managing Partner at Nexus Venture Partners (Part 3)

Sramana Mitra: Since the hype cycle has been so rapid, enterprise and business buyers have made AI a priority now, which was not the case, right? When you started doing H2O, it was not that case. It was much more niche, much more selective, and much more difficult to get the buyer attention, and that has completely changed. AI has become one of the top budget items in the enterprise IT expenditure.

Jishnu Bhattacharjee: Absolutely. So a few things are happening, Sramana, in the enterprise world. The initial wave was driven by consumer adoption. All the consumer companies we know started adopting AI heavily and they are rolling it out either as part of their products or as separate product lines for consumers.

As for enterprises, there are some use cases for which enterprises are seeing ROI directly at this point such as automation in software creation and automated content creation. People are getting some value, but by and large, there are a lot of questions. Enterprises are spending money, but at least a majority of that comes from the innovation and experimental budget. People are still figuring out the sustained applications and use cases that will give them value. I think we are still 12-18 months out to say that these are sustainable budget line items and these will be the applications and use cases that people will get the most value from.

In fact, along with the hype, there is already a trend of de-hyping of AI in the enterprise, because enterprise executives are almost getting overwhelmed. Thanks to the marketing horsepower of the corporate world, people feel that if they’re not investing in AI today, their competitive edge will go down, but they don’t quite know what to invest in.

The initial wave of investments are going into the infrastructure side of AI. At this point, it’s mostly GPUs, but they haven’t yet figured out if these investments are proportionately good. They can tell themselves that they’re doing something here, but what are the use cases that will emerge? There are signs that it is going in the right direction. The timing is not clear.

The investment world gets very excited. People are imagining and investing upfront because they feel that they can’t wait for two years when things get really clear. They’ve to preempt it and try things out, right? So that is driving a huge amount of enthusiasm and in some cases, very hyped up, irrational valuations, kind of what we saw in 2021, at least in select AI companies.

Sramana Mitra: So, there’s one other category I would add in the enterprise use cases that has some clarity, which is customer service, right? Their ROI is very obvious, very easy to understand.

Jishnu Bhattacharjee: We have invested in a couple of companies in that area.

Sramana Mitra: Generative AI is a natural for it. Of course there is the hallucination problem, but I think there are ways to mitigate that by training the system with domain-specific constraints. So that’s another category where things are more clear.

On that topic, we’d a very good conversation last week in this same program. There’s FOMO, but there’s also actual fear of people losing their jobs on the ground inside of enterprises. People are afraid. There is a tension that is going on between the AI companies who are building applications for enterprises who are being used by large numbers of people inside the enterprise, but normally when a software company goes into an enterprise with a product, there’s a lot of give and take lots of input that comes to refine the product, to make it better and so forth. There’s a tension right now that these people who are using these products, if they train the product too much, they may lose their jobs. So that loop has friction in there.

I don’t know if you have something to add to that point.

Jishnu Bhattacharjee: Yes, there is friction. It’s happening in a very TenX scaled up way, but this occurs whenever there’s a any technology wave. We see that certain kinds of jobs become less important, but many more jobs get created.

Today, most of the AI systems getting adopted are all very good at exception handling and human-in-the-loop intervention, not only for efficacy, but also to make the models better. For example, a job category called prompt engineers, which was not there before is coming up. Thirty years ago, database administrators (DBAs) or database specialists started coming up. So you’ll probably see lot of different kinds of jobs come up. Several services companies are taking up jobs such as AI ethics monitoring. You’ll be surprised by the things people can do with AI.

In certain categories, we will see a shift of jobs. Like you mentioned about customer service, if the company can increase the efficiency to the level that they don’t need that many agents before, they will obviously make those decisions. But those people, who get freed up from doing regular basic question answering, will get into judgment level jobs. There’re some things where human judgment is still needed and can’t be done by machines. So you will see those things, but in the short term, there will be this shift.

This segment is part 3 in the series : 1Mby1M Virtual Accelerator AI Investor Forum: With Jishnu Bhattacharjee, Managing Partner at Nexus Venture Partners
1 2 3 4 5 6