Big man with flashlight helps businessman see artificial intelligence through binoculars
getty
Looking at the landscape of AI startups can sometimes provide a glimpse into how AI-powered business ideas work in the real world.
For me, there’s a sense that people who are watching this space closely have revised their outlook over time. For example, in 2021, many people had a vague idea that everything would be robotized, but now we have a more detailed outlook.
I’m thinking of a recent example in which a trio of Harvard Business School students, who are also part of the Y Combinator community, talked about applications to practical fields like HVAC. This was at the Imagination in Action event in Boston in April (which I was responsible for organizing), and we had a lively conversation about all things related to AI.
In this special segment, Ben Horowitz interviews Cortex’s Vidhan Bhaiya and Robbie’s Joe Schwartzman. Cortex has been proactive in acquiring various companies to bring AI to their operations, and Robbie has a particular focus on supporting HVAC technicians in the field.
AI trajectory
Initially, the group was discussing speed of implementation.
“What happened to software should also happen to hardware,” Bayer said. “With hardware innovation, it doesn’t take 12 to 18 months to design every little thing. I think what’s missing to make that happen is actually engineering judgment.”
This explains why Cortex buys companies.
Attend the call of the furnace
Meanwhile, Robbie is building an AI agent for home service technicians.
“We decided to go with HVAC technicians and see what we could do with AI to help them with their daily tasks,” Schwartzman said, reflecting on the company’s origins. “We pivoted a bit, but we were lucky enough to be part of the latest YC batch.”
glasses fail
Schwartzman then narrated an ultimately unsuccessful project to equip HVAC technicians with smart glasses.
“This was a great idea from HBS,” he said. “It sounded great on paper, but it just didn’t work in practice. When we put the glasses on the technicians, we realized they weren’t actually solving the most pressing problems. There was no clear link to the bottom line. Honestly, it sounded good. There’s a disconnect between something and a great business idea, so after spending more time on it, we decided to obviously change direction and just use software first.To be honest, the fundamental problem could have been solved with software.
stepping stones
Mr. Bhaiya then spoke about concrete ways to align development with incentives.
“It’s very difficult to align incentives,” he said, noting the importance of building trust. “For these tools to be valuable, they need to capture the judgment and reasoning of senior engineers.
That being said, he outlined three steps to progress.
“Step one is educating different people who are at very different stages of their AI journey,” he said. “I think step one to building trust is making sure everyone is on the same level, and that’s part of the transformation. We believe that we have been doing this for over 20 years. They are the best people in the world to tell us what the ideal state of this process would be. We believe that AI, we are the enablers of AI and taking it to that level. However, we do not intend to say, “This is the ideal.” Once we help bring AI to that level, we’ll ask them. ”
So what is step 3?
“Step three is, can you tell me how you think about the problem?” he explained.
Fear of redundancy: The problem of job loss in HVAC and elsewhere
In answering questions about unemployment, the pair kind of parried some of the despair and melancholy that might be heard around the water cooler.
“Engineers aren’t really motivated by money,” Bayer says. “Engineers actually want to do innovative work. They want to push the boundaries of innovation.”
He also distinguished between more general processes that could eliminate jobs and more surgical changes that focus on what AI is good at.
“We are pursuing white spaces that we would not have access to without AI,” he said. “And that’s consistent with what everyone wants.”
Schwartzman noted that many HVAC technicians are “smug.” Because they are less disruptive than engineers in other fields.
next frontier
Towards the end of the talk, Horowitz questioned the pair about the innovative system, asking, “Who is the audience?”
Mr. Baiya suggested that his audience was people interested in rapid progress.
“All the advances that have happened in software engineering are things that people can see, and we’ve seen how software engineers can now do things that they couldn’t do before,” he said.
Schwartzman said the focus in the lobby is on the technicians doing the work.
“There aren’t that many home services investors that are really on the front lines, and that’s kind of the cause of the disconnect,” he said.
Internal and external strategies
I think the next point made by Mr. Bhaiya was very interesting.
How do you create third-party influence?
Basically, Mr. Bhaiya started by saying that he just wanted to sell software rather than buy companies.
But Cortex’s ultimate strategy of taking the reins of its customers has created a lot of value, he said. he explained.
“To get that contextual data, you can’t do it as an external vendor,” he said. “I think it requires an alignment of incentives, which means we have to be part of the inside.”
He suggested that Cortex needs access to long-term data and the power to change business relationships to maximize the value of transformation. So they remained acquired.
Schwartzman then explained how the lobby works.
“The data is there,” he began. “We derive valuable insights from that data, for example, making recommendations based on the expected value of the booking. We translate that to the technician, give them their attention, and fully provide them with all the context they need to best serve that customer. And after the booking, our voice agent collects all the information from the visit and feeds it back into the original CRM system.”
It’s a life cycle.
“We have a better understanding of how to increase revenue from every visit,” he added.
Schwartzman said Roby’s approach also solves the problem for HVAC technicians, who must complete a significant amount of paperwork, often on iPads, when touring a project site for the first time.
“Even if they’re not our customers, we know they’re not paying us. At least as good as what they’re doing now,” he said of the engineers’ opinion. “In fact, it’s gotten even better since I started using it.”
What will drive adoption?
To conclude his talk, Horowitz asked two others for their outlook on what will move the AI race forward in general.
“This is the fastest-introduced consumer technology in history, but people forget about it as soon as they get into the workplace,” Bayer said.
Schwarzman suggested that people are overcomplicating startups for today’s world of business strategy. He touted Y Combinator’s influence in focusing on founders.
“YC was great at saying, ‘No, all that matters is the bottom line, and the only reason it matters is that people will only pay you if they find value in what you’re doing,'” he said.
They also agreed that simplification is at the heart of the new initiative. Basically, there was a consensus that common sense standards could drive startups, hiring, and engineering by clarifying this at the end of this segment.
I thought this was one of the better parts of the event at a time when startups and even more established companies are in dire need of advice to give them direction. Please look forward to future updates.
