I run the technology organization at Aptitude 8, a go-to-market systems consultancy. We're the partner companies bring in to make their sales, marketing, and service tools actually work, from strategy through implementation. My side spans pre-sales, solution engineering, development, and internal product. Lately, most of my energy goes into putting AI directly into how we build.
A quick primer: the things I do well and genuinely enjoy, with a bit of where I've proven them out at Aptitude 8.
Designing how the pieces fit together: CRM platforms, custom integrations, portals, UI extensions, coded actions. The technical spine of client solutions.
Scoping aggressive, high-stakes migrations so they actually land. I built the framework my team uses on enterprise deals.
Running pre-sales, solution engineering, and development teams, and moving them to AI-assisted delivery. Same effort, sharper and more consistent output.
Not AI theater. Rolled Claude out org-wide and took daily use from ~50% to 90%+, then started embedding it into how we deliver. Now a company growth vector.
Stood up an internal product team to build the tooling, skills, and plugins that fill the gaps standard SaaS leaves behind.
On the exec leadership team, helping set company direction and shaping our AI service offerings. Sustained YoY growth.
There's a lot of noise about building your own CRM, or any core system, from scratch. It usually gets framed as saving on licensing, with AI as the thing that finally makes it cheap enough to try. In theory, sure. In practice, I've never once seen it pay off.
It’s the last thing you should touch. A free CRM already does the job, and your time and your Claude credits belong in your actual product. Saving a few hundred dollars a month sounds great until the "quick two-hour build" turns into days of real work to make something secure that a spreadsheet wasn’t already handling.
The money looks more real, mid five or six figures a year. But now security is the whole game. Your internal systems need to be SOC 2 compliant to satisfy your own customer contracts, and that license is a rounding error next to a breach and the lawsuit behind it. The math quietly flips back.
So the move is almost always the same: trust the core system, get everything you can out of it, and purpose-build only the piece that genuinely can't live inside it. That last part is where AI truly changes the math. The adjacent tool that used to take forever to build, cost a fortune, or mean a six-figure add-on is now something you can actually build.
As a trusted advisor, the real skill is knowing where that line is. First I help your team push the system to its limits, so you're getting every bit of ROI it can actually give. That's usually further than people expect, and it's the honest way to find the true edge of the platform. Then, and only then, I build what lives beyond it: sometimes an AI feature that couldn't exist before, sometimes software that finally got cheap enough to be worth it.
Built a startup's Salesforce managed package. The installable product their customers add to sync platform data in and out of Salesforce. Their actual product, not an internal tool, and they've kept building on it since.
Architecting a move off a legacy system into Salesforce. Consulting-heavy discovery, then building the process to match. The switch saves multiple six figures a year in license costs, many times over the one-time build.