1Mby1M AI Investor Forum: Anupam Rastogi, Managing Partner at Emergent Ventures (Part 1)

Anupam Rastogi, Managing Partner at Emergent Ventures, discusses his firm’s AI investment thesis.

Welcome to the program, and let’s talk about what you are up to with Emergent Ventures and AI. We started with this watershed moment of generative AI coming into the picture in 2023 and now almost two years hence, AI has moved at a breakneck pace.

So, my first question is, where is your head vis-a-vis AI? How are you parsing all the happenings of the world? How are you framing your investment thesis and so on?

Anupam Rastogi: First of all, thanks for having me here. It’s a pleasure. By way of introduction, I’m Anupam Rastogi with Emergent Ventures. We’re a Bay Area-based, seed stage-focused venture firm.

We’ve been focused on enterprise AI since we started in 2016. We have been focused on this space for a while, and our model is to lead or co-lead pre-seed and seed rounds – partner early with founders and help them get to the next milestone, typically a Series A round. We work pretty closely on product, go-to-market, downstream fundraising, and navigating a bunch of company building milestones.

Now coming to specifically ai, we have about 45 portfolio companies now that we’ve invested in over the last eight years. These are across the length and breadth of the landscape. All of these are B2B enterprise, are mid-market focused primarily, and in essentially both layers – Application layer and infrastructure layer. Application layer companies are the ones that are going deeper either into function across sales, marketing, HR, finance, operations, etc or vertical-focused companies, which are going deeper into specific verticals such as manufacturing, media and entertainment, logistics, etc.

I’d say that there’re different types of opportunity in AI. All of SaaS and cloud combined today is worth about $750 billion per year. That’s a big number, but all of corporate spend globally is in the order of $50 trillion. So all of SaaS and cloud today is just about 1%- 1.5% of the global corporate spend.

So, after all these years, SaaS and cloud make up just a tiny drop in the bucket. For corporates, SaaS is not a big spend today. The reason is that SaaS is primarily an enabler in its historical form.

What’s happening is that now you can deliver outcomes with AI. SaaS was basically just enabling someone to use a tool. A salesperson, for example, would use a tool for $50-$100 a month, but that salesperson may be making $10-$20K a month. That’s that same 1% type ratio, even less.

However, now SaaS can deliver outcomes with AI, and that number can go up 10x or maybe more over the next decade. That’s the opportunity we are actively investing in.

What we are looking for is, how do you get to these large opportunities that are opening up? Pretty much every workflow, every function, and every industry will get rewired. What’s happening today is just the tip of the iceberg in terms of AI adoption. When all these things fully get adopted into entire workflows, then everything or how work is done broadly is just going to look very different.

Gen AI started hitting mainstream around 2022 November with ChatGPT. Before that, some of the imaging models had come out, but a lot of that has been actually simmering in the background. Before that, we of course had the whole machine learning or deep learning wave, which is already today at a meaningful scale. It has a lot of proof points; it already works. When you combine the two, you get even better results in the enterprise setting.

Your machine learning or deep learning model could be creating a prediction for something. For example, is this customer likely to churn? Or what topics is this customer likely to be interested in? Machine learning is really good at a lot of those type of things that are based on your own data. Then Generative AI helps close the loop on let’s say framing a communication with that customer.

When the two are combined, that makes for some interesting businesses.

Sramana Mitra: Do you have a case study in your portfolio of what you have invested in that combines deep learning and machine learning as well as Generative AI?

Anupam Rastogi: Yes. Several companies do that. I can take examples from customer service or employee productivity. So let me pick a couple of quick examples.

There’s one called Prezent.ai, which helps corporate employees create presentations aligned with their data and brand. You can just give it inputs on who you are looking to present to – a senior member of the executive team, a CEO, a peer. You provide the persona and the underlying data and can create an entire presentation. It’s much deeper than what a ChatGPT or a Claude, or any other LLM today can come up with. Those slides wouldn’t be very customized to your data or to what should the story be.

This product from Prezent is trained on millions and millions of slide decks on what’s good, better, best, and not good on presentation decks. Then based on the persona, the storyline, the data, the machine learning model will come with the topics that should be covered and the topics that should not be covered. Then GenAI is used to generate the presentation.

Similarly, we have a couple of companies which are doing really well in the customer service space. We have been their investors for a few years, and those are now sizable companies.

Two of those companies were one of the first adopters of LLMs back in 2018 when it was still in a very formative stage. They run the machine learning on the customer data, the sentiment, the tonality, and a lot of the underlying data when they’re trying to understand the conversation. GenAI is used to generate the output – what should you say to this customer in an email or a virtual AI agent chatting with the customer. So again, it’s a combination.

So, if you just use GenAI, you’ll be immediately able to tell that it is GenAI generated. It’s not as customized to the problem, but when you combine the machine learning element with it, that is much better at surfacing the problem. This is an angry customer; this is a happy customer. This is the issue, and whether they’re a churn risk.

A lot of those things are identified by the machine learning piece.

This segment is part 1 in the series : 1Mby1M AI Investor Forum: Anupam Rastogi, Managing Partner at Emergent Ventures
1 2