1 August 2026 · Strategy

Why We Build With Three Different AI Toolkits, Not One

Ask five AI vendors what they build with and most will name one tool. That's usually a sign of convenience for them, not fit for you.

We build with technology from Google, Anthropic, and OpenAI, depending on the job in front of us. Not because we can't pick a favorite. Each one is genuinely better suited to different kinds of work, and your business deserves whichever one actually fits, not whichever one we happen to know best.

Some jobs need the exact same steps, every single time

Invoicing, compliance approvals, daily reports. For work like this, Google's agent tools are built to lock down a fixed sequence, so the outcome doesn't drift no matter how many times it runs. We reach for this when a process has to be predictable, because predictable is the entire point.

Some jobs need an AI that can dig in like a technician

Some work happens behind the scenes: reading through files, fixing a broken workflow, investigating something that went wrong. Anthropic's tools give an AI direct access to files, logs, and systems the way an actual technician would work through a problem. We use this to build and maintain our own systems, not usually where a customer ever sees it.

Some jobs need an AI that talks to your customers

OpenAI's tools are built specifically for conversation: handling a live chat, answering a question, picking up naturally where the last message left off. When a customer messages you on WhatsApp or your website, this is usually the technology doing the listening.

Locked into one tool means locked into one company's roadmap

If we built everything on a single company's tool, your business would be exposed to every decision that company makes going forward: pricing changes, feature changes, service issues. Building across all three means your systems can shift what runs underneath them without you ever noticing a difference on your end. The output stays steady even when the engine changes.

What this actually means for you

You don't need to know the names of any of these tools. That's our job to track. What you should expect from any vendor is a straight answer to one question: why did you pick this particular technology for this particular part of my business. If they can't answer that clearly, they picked the tool they know, not the one that fits.

Not sure which parts of your business need which kind of AI?
We'll walk through your process and tell you honestly where automation fits and where it doesn't.
Talk to Us
← Back to all insights