Ben Solari - CRO @ OpenHands - on AI Agent Engineeting

Your company, All Hands. You're an open source AI coding agent platform for software engineers and you recently raised in 2024 $5 million seed round and you joined recently as the chief revenue officer. If your company Open Hands succeeds, how do you think the day to day life of a software engineer is going to change?

A great question. I mean, and we're seeing, you know, a shift right now over the last few years from companies that are just figuring out how to use these coding assistants. I would say what OpenHands is kind of on the frontier of is these cloud coding agents. If, if OpenHands succeeds, which I very hopeful we, we will, I think what you're going to see is the rise of a new Persona within software engineering orgs around kind of agent engineering. And this is someone who is looking for opportunities within software engineering that are typically repeatable use cases where software engineers that are using AI in a more of an ad hoc basis are now able to kind of put some automation and code around those use cases where an agent can kind of be managing that at scale, which is a little bit different than the kind of call and response that you're seeing with software developers today using AI tools. So I'd say if we succeed, I think one major barometer to know that we succeed is what I consider kind of the rise of the agent engineer or the agent developer.

You're leading the sales team at a fast growing startup called Open Hands. You raised a $5 million seed round from Menlo Ventures at the end of 2024 product as an open source AI coding agent platform. A lot of business and technical leaders are listening to this show. What are the first tasks you think every engineering team should automate first using AI agents?

Great question. I would say where we see the and who knows what, 12 or 24 months down the line looks like this is probably the fastest, most disruptive time for software development that at least in my lifetime. I think what you're going to see in the near term for use cases for cloud coding agents is going to be around large code bases and it's probably going to be the unsexy work of maintaining that code. Only a fraction of engineers actually get to work on the fun stuff. Estimates my last company, jellyfish, only 20 to 40% of engineering effort really goes into roadmap and innovation work. So that leaves a pretty big chunk for agents to kind of help automate the redundant tasks of fixing bugs. So CVE remediation, looking for upgrading code bases, maintaining that code, migrating that code from one language to Another, so I'd say code migration, maintenance, bug fixing, support. Those are all very ripe use cases for cloud coding agents like OpenHance.

In the near term, you're at a fast growing startup leading the revenue team. Your startup is working in the open source AI coding landscape. Everyone talks about AI co pilots, but your company, Open Hands is building AI teammates. What's the real difference?

Yeah, so I'd say I'd take it even one step further. I'd say, you know, the premise of the company was kind of around this AI developer that was the origin of the company, company called Devin. A year ago, you know, put out a video that basically made it seem like the software engineering was going to change with the kind of rise of an asynchronous agent that could go off and do work, you know, without you needing to be by your computer. And I'd say the AI assistant was kind of the 1.0 generation of these tools, the Microsoft Copilots and the Cursor, you know, early products of the like where you're sitting in front of the computer and you're basically without changing your workflow, you're accelerating your typical job by utilizing one of these assistants. The AI developer or you know, AI teammate, kind of takes that one step further. I'd say that's like Gen two. That's an asynchronous workflow where you're able to kind of task an agent to go off and do a complex, usually kind of long running or long form task that requires many different steps. That's kind of at the frontier today. And then even one step further is kind of these cloud agent orchestrators or cloud agent architects who are building kind of end to end workflows that don't even require a human in the loop other than maybe to kind of check in and ensure that the agent is doing good work or they approve that agent's work. And so that's kind of the evolution that we've seen in terms of kind of first generation is like AI assistants. Second generation is maybe kind of the AI teammate or AI developer. And then really, really frontier stuff that excites me with open hands and some of our more sophisticated customers is this concept of kind of agent orchestrators or agent engineering.

You have had a long career in revenue leadership. You were a VP of global mid market sales at Datarobot. You were there for over four years. Now you're the Chief revenue Officer at Open Hands, a fast growing seed stage AI engineering startup. What have you learned about AI that truly delivers ROI versus AI, that's just hype.

Yeah, that's a good question. I would say, you know, across the last 10 years, I've spent at 2 now 3 companies that were kind of on the edge of changing people's behaviors. You know, with Datarobot it was kind of automated machine learning. With Jellyfish it was engineering intelligence. And now with Open Hands, it's kind of agent engineering and orchestration and software. These are kind of radical things that it's very hard to change people's behaviors at scale. And so in order to deliver ROI and actually find that effective step between where an organization is and kind of the status quo and where they want to be with these new technologies, I would say that the successful case studies are really around picking a smaller group within your organization. And it's always kind of relative to the size. If it's 100 people, pick 10. If it's a thousand people, pick 50 or a hundred. But finding some success of getting some small fraction of your team to buy in and be really fully bought in, they find some quick wins and then use that kind of momentum to evangelize in the organization is really kind of underlying pattern that every company that I've been at that has sold to or helped these companies transform, it's really finding a small cohort within these companies that help build that success versus expecting that your entire organization is going to buy into this new technology all at once without having some underlying success. That usually starts with a smaller team. So yeah, I would say smaller team go deeper with them versus wider and shallower with the broader team.

You've sold deeply to technical buyers your entire career. You were the VP of global mid market sales at Datarobot, you were there for over four years. You were the SVP of revenue at Jellyfish, you were there for over five years and now you're the Chief Revenue Officer at Open Hands, a fast growing seed stage AI engineering startup. What's the number one mistake that sales teams make when they're selling to engineers?

First thing that comes to mind, do not BS them. They are highly skeptical individuals. It's what makes especially engineering leaders great, but makes great engineers great, is that they are by nature skeptical people. They also by nature don't trust salespeople and so typically leading with vulnerability, being quick to admit that you don't know what the answer is because even if they're still new to your product, they, they can smell kind of the BS factor from very far away. So I think the most successful teams and sellers that I've, I've led have been extremely authentic. They've also been technical so that they can actually do some sharing and add value to the conversation. But they're very quick to admit that they don't know something. And, you know, all of that kind of underpins like building trust with these technical buyers. And you know, that takes time as well. So I would say the other thing is it doesn't come immediately. You know, you have to kind of do that over the, the course of the sales cycle and the customer relationship.

Awesome. Love that. Okay, you've gone from an individual contributor to a sales leader to a chief revenue officer. What leadership lesson hit hardest in your journey?

I would say that, you know, one is, uh, the people that you are leading tend to respond best to people that are in the trenches with them. Um, so I've always taken the mentality that, you know, the best leaders are often leading from the front, which means, you know, in the early days of Datarobot or Jellyfish and now back to it at Open Hands, I'm doing cold emailing, I'm doing cold LinkedIn outreach, I'm making calls on call blitz days with the rest of the team. Um, and I think the other kind of factor to that is like authenticity, where it's nice to work for and work with people that are kind of authentic and you know how they, they show up. And so, you know, I think that's hard when, for example, you know, they're high pressure situations where, you know, it's want to sometimes shield the team from that and still show up every day. But I think if you can kind of work hard and do some of the small things with the team along the way and then also kind of be authentic in the way that you're, you're showing up, you know, that tends to kind of work well for folks, at least it's worked well for me.