Saying no, automating triage: how Spearhead kept its projects running
Joe Phillips is CTO of Spearhead Tecnologies, an electronic-invoicing software firm born from two decades of custom-software work and a health-tech exit. He runs a disciplined practice that often rejects work that lacks a clear decision to improve, and this pragmatism extends to using AI for specific tasks like inbound media and partnership triage while keeping final judgment human.
Write down the decision you want software to make before choosing tools, then use AI to automate repetitive steps and speed approvals. Focus on process documentation and human oversight rather than tool comparisons or promised headcount cuts – AI delivers consistency and speed, not original judgment.
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?
I started Trifecta Software in 2011 building custom software for U.S. companies from Costa Rica, and later Spearhead Technologies for electronic invoicing. In between, I co-founded a health-tech company that Huli Health acquired in 2018. So I’ve built, sold, and kept operating software businesses for 24 years.
What makes us different is unglamorous: we say no a lot. Most shops will build whatever you ask for. We push back and ask what decision the software is supposed to improve. If the client can’t answer that, we don’t take the project. That has cost us contracts, and it’s also why the ones we take tend to still be running years later.
Where does most of your new business come from today, and where do you wish it came from?
Most of it still comes from referrals and people who worked with us years ago and moved companies. That’s flattering and it’s fragile: it doesn’t scale, and it means our pipeline depends on other people remembering us.
Where I want it from: people who read my new book or saw a talk and already understand how I think before the first call. Those conversations start at a completely different place. No convincing, no education phase. That’s why I’ve spent this year writing and speaking instead of chasing leads.
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?
Handed over: triage of incoming media and partnership requests. An AI reads each one, scores how relevant it is to what I actually know, and drafts a reply. It used to eat two hours of my day. Now it’s ten minutes of approvals. Nothing goes out without a human clicking send.
Went back to human: writing the actual arguments in my book. I tried using AI to draft sections and the output was fine, which was exactly the problem. It read like something anyone could have written. My readers are executives who’ve heard the generic version a hundred times. The whole value is that I’ve made these mistakes myself, and a model can’t fake having made them.
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, and it caught me off guard. People arrive already knowing my position on things, sometimes quoting arguments from articles I wrote, without ever having visited my site. They found it summarized somewhere else.
Through AI assistants specifically: it’s starting, in small numbers. What I notice more is a shift in the first conversation. Prospects used to arrive with questions. Now they arrive with a hypothesis they got from somewhere and want checked. It changes the job. I’m no longer explaining what AI can do; I’m correcting what somebody told them it could do.
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?
Do: write down how you make one decision that you currently make by instinct. One page, the way you’d hand it to a new employee. Most people can’t, and that’s the finding. Everything useful I’ve automated started as a page I’d written months earlier for a human.
Ignore: the tool comparisons. People spend weeks choosing between assistants and zero hours defining the process the assistant is supposed to run. The tool will change in six months anyway. The process document will still be right.
Also ignore anyone promising headcount savings as the first result. In my experience, the honest wins are speed and consistency. Fewer things fall through the cracks. That’s less exciting to sell and it’s what actually happens.
Thank you to Joe Phillips and the team at Spearhead Tecnologies 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.