1Mby1M Virtual Accelerator AI Investor Forum: With Ashish Gupta, Partner at Clearvision Ventures (Part 2)

Sramana Mitra: I’ve recently talked to a number of people who are looking at the services aspect and the human-in-the-loop aspect. Have you invested in companies with human-in-the-loop?

Ashish Gupta: A few. For example, there is a company called Skit. They build voice bots. So, an agent gets replaced by a voice agent. However, let’s say somebody starts singing an Elton John song in the middle of a conversation about banking. Then, the bot would not know where to go. So you need to bring a human in the loop. This is a ridiculous example of more realistic boundary conditions that you can imagine. So, some percent of the conversation ends up happening with the bot and some percent gets handed over to a human.

Another one is a company called Inferred here in the Bay Area. They extract data for mortgage and insurance companies. There, you will end up with the odd data element, which does require a human being to look at things and reconcile them. Otherwise, 99% of the work will be done using AI models.

Or a company that serves the pharma industry. They do content management. The same piece of content for a pharma company can find its way into marketing literature as it can in an FDA disclosure. Content is a very important part for pharma companies. This company uses a combination of man plus machine to repurpose content to help them figure out how to create labels for different geographies.

They are all companies that are constantly readjusting the boundary of what machine learning can do and what human beings then have to do to finish off the job that was partially done by an AI-based solution.

Sramana Mitra: Ashish, what are your companies – these examples that you just shared – experiencing in terms of adoption by these humans in the loop? By the way, do these humans work for your startups or they work for the enterprise?

Ashish Gupta: Great question and I think that is one of the important open in-play dynamics, Sramana. We’ll see how it goes out. I really don’t even have a hypothesis. For these companies currently, all the people who are using the tools are internal. Part of it is because the tools keep changing so quickly that training a third party on using those tools is a constantly moving target. As a result, it is a hard one.

Sramana Mitra: So they’re selling product and services, technology and services, and the service element is performed by the startup.

Ashish Gupta: Yes, but the question that you raised is very much an elephant in the room. If, tomorrow, someone comes in with a solution that is usable by an internal person, there are internal HR questions. For instance, the person who owns a particular function is used to outsourcing the work.

Let’s take the Head of Marketing in a company. They’re used to working with agencies. Actually, one of the companies that I’m invested in does exactly that. It’s an AI-powered agency. The head of marketing is used to buying final content. But tomorrow, what if they only want to make minor changes, let’s say they want to change which to that in a website? Should they really be paying the agency for minor changes like that? Or for changes that don’t impact the SEO value of the content?

Those boundaries will soon start getting crossed and in much more meaningful ways. So your question actually is one of the big ones that all of these companies are thinking about, even though they’ve taken a certain choice right now.

Sramana Mitra: There is a very similar related question that is also coming up where the vertical AI company assumes that a human on the enterprise side is going to use their technology. The human who’s supposed to be using their technology is also supposed to be training that technology. That human is very nervous about using or training that technology with the fear that they’re gonna get eliminated from the loop.

Ashish Gupta: Very true. One of the reasons many of these companies have taken the approach of bundling both the technology and human expertise to deliver outcomes is because the HR aspect is complex. It’s the kind of question that’s beyond our pay grade, as there are larger forces at play. I completely agree with you.

This segment is part 2 in the series : 1Mby1M Virtual Accelerator AI Investor Forum: With Ashish Gupta, Partner at Clearvision Ventures
1 2 3 4