By Ken Rona, Chief AI Officer, JWX
Priority two of JWX's AI mandate was rapid prototyping: getting new products in front of clients fast enough to get real feedback before committing to a full build. Here's the story of the prototype that gave the team the confidence to build what's now QuantumPath by JWX.
If you're evaluating an “agentic AI" pitch from any ad tech vendor, this is what a real prototype-to-product path actually looks like, three days of build time, then months of hardening before it touches a real client.
The prototype that gave us the confidence to build QuantumPath by JWX
The second pillar of our transformation strategy was to create a rapid prototyping capability. In addition to the CoE developing our internal tooling, they also took responsibility for creating proofs of concept and MVPs of products we wanted to show clients for feedback. This was the one area where I contributed directly. John Nardone and I would discuss and identify projects that didn't have a natural home in the organization, or where the existing team was too busy to tackle them.
In these situations, I'd act as the developer of first resort and put together a front end and some product requirements documentation. Once the front end was in place, I'd bring in the rest of the team to get the product closer to an MVP. Our focus on rapid prototyping created urgency and forced us all to explore new techniques and practices to use AI more effectively. That was a critical decision: because our focus was on development speed, we experimented with a variety of techniques, spec-built development, workflows, Slack for project coordination, and more.
As an example, nine months ago I built a fully functional integration with a well-known DSP in three days. A user could traffic an entire campaign through a chatbot, uploading a campaign brief and letting the system handle setup from there. None of this was ready for production or for clients, though it looked pretty good, but it was more than good enough to show clients, get their feedback, and give us the confidence that we could fully build out QuantumPath: a platform that sits on top of DSPs and SSPs and lets a user and an AI agent work from a single, shared control pane.
That three-day prototype is the reason QuantumPath by JWX exists today as a real product agencies and brands can use. If you want to see where that idea landed, have a look at QuantumPath by JWX.
Key Takeaway: If you're weighing whether a vendor's rapid-prototyping claims hold up, ask them the same question this post answers: what did the three-day version actually do, and how far was it from what shipped?
Next up: The AI Residency
Prior blogs in the CAIO series:
Why AdTech Needed a Chief AI Officer
The Bot Family
The Trojan Horse