AI's Promise Beyond Doubt: New Building Blocks for Work
Nvidia’s earnings defined the week in financial markets, as investors continue to fret over whether the hundreds of billions of dollars in spending will pay off soon enough. I took time with some leaders who are building on that infrastructure to see how it’s going, and what the roadmap to higher productivity looks like.
If you’d like to join me – and peers – for deeper conversations on innovation and leadership, get on this list for Fortt Knox Executive Communities, launching soon: mba.fortt.com.
Freshworks
Small and medium businesses use Freshworks to run their internal systems, so after the company reported earnings I spoke with CEO Dennis Woodside about the role customer AI expectations are playing in the sales process:
If we’re talking to a new customer, they want to understand, what is our AI strategy? Where are we going? What’s the vision? If we’re talking to an existing customer, they’re under pressure more so than they were six months ago, to put AI into practice, to make their teams more efficient and deliver better value for their own customers. So it’s critical to the buying decision. Now, it’s expected that platforms like us will have AI baked into their business and it really is a driver for us of our business. We’re moving increasingly upmarket, as you know, bigger and bigger accounts. Our typical customers, let’s say 5 to 20,000 employees, companies like New Balance or S&P Global, that are looking for a platform that is less complicated than the older incumbents, but offers the level of sophistication, enterprise grade and ultimately AI that they expect.
AI is becoming this essential part of the future-proofed roadmap. Your big customers want to know that you have an AI strategy that’s going to enhance the value of your product. Even if you have been in business for quite a while, before they’re going to agree to continue doing business with you and lock in for a number of years. That’s very different from what we were hearing three, four quarters ago where these companies were saying, Yeah, I’m preparing along the lines of AI, but really it isn’t a demand driver yet. Now you’re hearing it is a demand driver, but not necessarily just the product I have now, but the capability that I can show to have valid and valuable AI products in the future. Freshworks is more of an ERP, back-office play, driving efficiency throughout a business. Dennis is saying that customers are telling him, “We want to have faith that you guys have a plan to drive efficiency using AI.”
Scribe
How will the workforce use AI to get better? And how will those AI agents understand the steps to complete jobs humans have been doing? I spoke with Scribe CEO Jennifer Smith, whose company this month announced a $75 million Series C investment round at a $1.3 billion valuation. Scribe does documentation with the help of AI:
If you’re familiar with large language models, that’s what ChatGPT is. The way it works is it predicts the next word in a sentence. Our large workflow models predict the next step in a process. Very similar idea. We’re now just applying it to what you’re doing when you’re at work every day. We can use that to suggest ways that a process could be better. So we want to do something where we say “Hey, that was cool. Here are three different ways that you could actually improve this. You could do it faster or more efficiently, more accurately.” We can do that for you, Jon, the person. We can do that for your team. And we can start to look across your entire company and be like, “Hey, actually there are a lot of places where you’re doing duplicate work or you’re doing it like pretty inconsistently. Or you know what? If you built some software connectors, maybe people wouldn’t have to do so much manual data entry in things.”
I think Scribe is fascinating. I want to try it, because I liken it to performance capture for athletes. For movies, you’ve got these motion capture suits. Right now, a lot of elite athletes will use technology that watches them go through certain moves on the court or on the field, captures their motion down to the smallest detail and figures out, is that as efficient as it could be? Maybe it’s a golf swing. Exactly how does this look? How can this be improved compared to the best? Right now, part of what Scribe is doing is enabling that kind of performance capture with what people are doing on the browser, in mobile apps, etc. and breaking it down into pieces that I can then analyze and either train agents on how to do that same repetitive thing, or as Jennifer was describing there, just analyze that process and say, based on what the best are doing, here’s how you might be able to improve that process. This is a fundamental assessment and analysis piece that can lead to enhancement later, and strikes me as a very important step.
Notion
It’s not just workflows that need to get broken into modular pieces in the AI era; software itself can benefit from that, too. That’s part of the idea behind Notion, the AI-friendly productivity suite that recently passed $500 million in annualized revenue. CEO Ivan Zhao told me about Notion’s winding path to product-market fit, which might foreshadow the broader market’s journey in AI:
We tried to start as a developer tool. It didn’t get traction. Or we tried to build a web page builder and software builder. People didn’t know what to do with it. We tried a bunch of different ideas those four years, and we had to restart the company in Japan at one point. … So we decided to rebuild, lay off everybody, rebuild just with the two of us. And that’s why we went to Japan, actually, because when morale was so low, we thought “Let’s just go somewhere that we haven’t been to.” And turns out Japan’s cheaper than San Francisco for the cost of living. We leased our apartment, we actually made money living in Japan. Definitely eating a lot of noodles back then. But you can’t complain, the ramen in Japan is delicious, right?
Notion hadn’t gotten its footing yet. It’s the first four years and product market fit was elusive. And they just decided to go back to first principles. Now that’s the piece – going back to first principles – that I think a lot of companies in the AI era are going to have to do. What do we really believe in? That’s something that Notion and Scribe have in common is they’re trying to break down processes and software to their fundamentals – Scribe the processes, Notion software itself. Both of those fundamental building block approaches are very AI friendly, because AI needs to be able to put together different pieces in order to figure out where to go. It needs a language to work from. Ivan talks about being like Legos for software. In a way, Scribe is trying to optimize how best to put those Legos together, and Notion is providing the Legos themselves.
If you’d like to join me – and peers – for deeper conversations on innovation and leadership, get on this list for Fortt Knox Executive Communities, launching soon: mba.fortt.com.