How Software Toolbox cut support time to one minute with jonjon.ai
Andres Naranjo is CEO of Software Toolbox, a 30-year-old industrial connectivity firm that serves about 6,000 customers across 45 verticals and 75 countries. He has repositioned the company around industrial AI – combining decades of plant-floor connectivity know-how with agents like jonjon.ai that have reduced first-response support from under two hours to roughly a minute.
Start small and practical: connect the plant-floor data you already have, add the contextual documentation and naming conventions AI needs, then layer in models. You don’t need perfect standardization across every site before you begin – data, context, AI is the stack that delivers usable results quickly.
What’s the quick origin story of your business, and what makes what you do genuinely different from the other options your customers are looking at?
Software Toolbox started 30 years ago in Matthews, North Carolina, solving one unglamorous but mission-critical problem: getting industrial equipment to talk to software. Over three decades we’ve built connectivity into roughly 6,000 customers across 45 verticals and 75 countries, from OPC and Modbus drivers to data integration platforms like Cogent DataHub and OPC Router.
What makes us different now is we’re repositioning as a full-stack industrial AI company, with products like chatUNS.ai and jonjon.ai. But at the end of the day, AI in manufacturing lives or dies on whether you can actually get to the right plant-floor data and understand the context around it. That’s what we’ve been doing for 30 years. We know how those systems work, how to connect to them and, importantly, what it takes to earn the trust of a manufacturer running a real production environment. Now we’re bringing that experience together with AI to make that data more useful and actionable.
Where does most of your new business come from today, and where do you wish it came from?
Today, most of our business still comes from industrial connectivity and data operations. TOP Server is a big part of that, along with helping customers move data from point A to point B through platforms like Cogent DataHub and N3uron. That’s high-intent demand. Engineers know the problem they’re trying to solve and we’ve been doing that for decades.
Where I want more of our business to come from, and where we see that shift happen is AI. jonjon.ai is already being used by hundreds of people every week, and chatUNS.ai is helping customers actually work with the data we’ve been connecting for years.
That’s really the direction we’re headed. We don’t want to just be the company you call when you need to move data. We want to be the company that helps you understand that data and do something with it.
What’s one thing you’ve actually handed over to AI in the last year, and one thing you tried it on and went back to doing the human way?
The biggest thing we’ve handed over is the first line of our customer support. We built an agent called jonjon.ai that now handles the initial response whether a customer emails us, calls us or submits a web form. We used to be in the sub-two-hour range for responses, which was already pretty good for our industry. Now we’re at about a minute, and the accuracy has been really impressive. Our CSAT scores are still at 100, and customers can always escalate to a person if they need to. So we haven’t gone back on AI-enabled support.
We also haven’t really had a case where we said, "OK, AI isn’t going to work here let’s just go back to the old way." But that doesn’t mean everything worked the first time. This is an era of high experimentation. You have to tweak the prompts, look at the outputs, adjust the direction and keep improving it. For us, the lesson has been that when something isn’t working yet, you don’t necessarily abandon it. You figure out why and keep working at it. That’s been our approach with AI.
Have you noticed a change in how customers find you or what they already know before they reach out? Is anyone arriving through AI assistants like ChatGPT yet?
Yes, measurably. Classic organic search behavior is eroding. We’ve watched keyword positions and click-through patterns shift industry-wide, not just for us, as Google’s AI overviews and assistants intercept the informational queries that used to land on our educational content. Customers also arrive further along: they’ve already had an AI explain the basics to them, so the first conversation starts deeper.
And yes, we’re starting to see referral traffic from AI assistants in our analytics, and it’s small but growing. I think that’s going to become increasingly important for companies like ours. It’s not just about ranking on Google anymore. You also have to make sure the information you’re putting out is accurate, useful and authoritative enough that AI tools can find it, understand it and use it when they’re answering someone’s question.
If you were advising someone in your industry on all of this for the next 12 months, what would you tell them to actually do, and what would you tell them to ignore?
What to do: understand that getting from industrial data to insight can be simpler than people think. You really need three things.
First, data – which most plants already have a lot of it. Second, context — and this is the part people often overlook. The P&IDs, tag naming conventions, documentation and tribal knowledge need to be organized in a way that AI can actually understand and reason over. And third, an AI that can work with that contextualized industrial data. That’s what we’re doing with chatUNS.ai.
Data, context, AI. That’s the basic stack.
What I’d tell people to ignore is the idea that you have to standardize everything across the entire organization before you can get started. Don’t wait until every site is perfect. And don’t avoid AI altogether. Start with one site, one problem and prove that it works. Then build from there.
The companies that spend the next year trying to create the perfect data foundation before doing anything with AI are going to move more slowly than the companies that start small, learn and keep building.
Thank you to Andres Naranjo and the team at Software Toolbox for sharing what actually worked with Leaders Perception readers.
Want to share your playbook?
Leaders Perception publishes founder and operator interviews every week. A few questions, about five minutes, no calls. It is free and everyone we accept gets published.