ServiceTitan exec on AI trends in the trades

ServiceTitan in December released its first annual ‘State of AI in the Trades’ report, which showed that just 12 percent of the more than 1,000 contractors surveyed have embedded AI in their processes, while 34 percent are experimenting with it, and 41 percent have taken a wait-and-see approach.

What’s happening: Homepros caught up with Vincent Payen, a product executive at ServiceTitan, who shared additional insights into what the company is seeing regarding contractors’ use of AI — and ways it may evolve in the near term. 

This conversation has been lightly edited for length and clarity. 

Based on ServiceTitan’s data (aside from the report), are there any interesting insights that you have, broadly speaking, about how contractors are using AI right now?

We see AI as a trigger to fundamentally rethink business operations. If you need to teach a machine to dispatch for you, you actually have to think about your job types, your skills, all of the logic that goes behind making the right decision. And in many businesses, this has been tribal knowledge — sometimes documented, sometimes not. 

Effectively, if you don’t have good rules [about] how you want to treat an after-hours emergency, how you want to treat members versus non-members, [and other scenarios], the AI is not going to do what you expect.

So what we’re seeing is that the adoption of AI is triggering an upleveling of rigor in terms of, ‘How should the business function? What do we expect? How do we want to set things up?’ And I think that’s a very interesting byproduct of AI adoption, which I think is going to be incredibly positive for the trades.

What else are you seeing?

We see contractors experiment with lots of different AI solutions — and often disconnected [ones]: ‘I’m going to have a marketing AI solution, and then I’m going to have a CSR solution, and I’m going to have a dispatch solution.’ And more and more, we’re seeing the experiments run into trouble, because you end up having these point solutions that don’t talk to each other. 

And I think they’re realizing that greatness will come from connected AI solutions. It just creates weird experiences if you have a virtual CSR that answers something and then an automated marketing reminder that says something else. 

Where are you seeing the technology as more impactful today: Revenue generating or cost cutting?

I think the answer today is very much top-of-funnel efficiencies. If you think about virtual agents, they do two things. They allow you to have zero missed calls, which drives growth. (If you capture all the inbound volume.) That also means you may need fewer humans to answer the phones, because they can go beyond just answering the phones. 

Now, someone quits: ‘Do I replace them?’ We usually see that motion happen — from ‘great, let’s drive growth!’ to efficiency driving. But the interesting thing is that the reinvestment of that efficiency doesn’t necessarily mean fewer humans. 

It can mean, ‘Now we’re going to have humans who can be focused on more valuable and complex tasks that we didn’t have time for, like getting closer to our customers on happy calls, being more proactive on memberships, or estimates.’ So there’s a virtuous cycle playing out right now. 

We hear a lot about the dark side of AI, so how do you think about the risks?

The hallucination of AI, the incorrect information… the danger is real across the board. 

Virtual CSRs are less dangerous than installing a furnace wrong, yet how should they react to a customer who’s giving signs of distress? We had a good example recently: A 90-year-old person who just had heart surgery and [a new furnace installed], and there’s a weird smell. This is actually a life-or-death AI security that cannot go wrong. That’s why [starting] with AI-assisted before going fully AI-trusted is very important. 

You worked at eBay earlier in your career, an auction platform. In other applications, we see dynamic pricing, surge pricing — based on capacity. How do you imagine that evolves in home services?

I do not believe that AI will be a great equalizer, because the way you use it is incredibly personal. The way a virtual agent is going to behave for Contractor A could be very different than Contractor B, based on how they want it to behave. I think that’s an important foundation for dynamic pricing.

There are lots of things that were historically very difficult for contractors to do in real time — demand generation, pricing adjustments. The real-time tuning is something that AI will absolutely optimize in a way that I think is going to be very beneficial to contractors and transformative to margins. If you have a valuable job coming in, what’s the best reshuffling of your capacity to actually take that job, with less damage or risk of people being moved or canceling? 

All of these together are going to get much easier, as the AI understands — at the same time — your marketing campaign, your demand spend, your capacity, and where technicians are in the field, and can make decisions based on all of these data points.

I think it’s also going to bring much more transparency to the market. Google now effectively has an AI agent that’ll call five companies for you to get pricing and report back, and I think there are going to be very interesting things in terms of dynamic pricing. For example, [a contractor] would love to say, ‘I’m overbooked today, but if you want to pay $100 more, I’m gonna pay a technician overtime and send someone to your house.’  

What else do you anticipate is on tap for this year in particular?

This is the year of agents being aware of other agents in the business — the virtual CSR knows what marketing promotions exist — which creates a world of orchestrated action. 

For example, ‘We have a truck that was two doors down, and they finished a job already. We know we’re going to come to your house next week for your annual maintenance. Would you like us to come right now?’ If you say yes, the dispatch agent just changes things. 

We can only do that if all the agents are true ‘coworkers’ in the same system — so they can talk to each other — [but] these flows across multiple parts of the business are what we’re going to see come more and more to life this year.