Customer.io Uses AI as Translation Layer
Colin Nederkoorn, CEO and founder of Customer.io, on AI's effect on marketing automation
This marks a shift from simplifying the product itself to using AI as a translation layer on top of a powerful product. Customer.io has spent years adding more moving parts, from journeys to data pipelines to email building, because its best customers wanted more control, not less. The bottleneck became understanding that complexity, not building it. LLMs let the company keep the advanced workflow engine and make it usable by marketers who cannot write logic or HTML.
-
Customer.io started as a tool for technical marketers. Teams pipe in product events, build segments off behavior like last login or feature use, then send branched email, SMS, and push campaigns. That flexibility created switching costs and strong retention, but it also made the product harder for non technical teams to own as companies grew.
-
The recent product direction is to keep adding context and capability, then let AI explain and operate it. The assistant can answer what a screen means, create segments, and use docs plus company specific context like brand voice and business type. That is closer to a copilot for setup than a prettier interface.
-
This is the same pattern seen in adjacent software. Intercom did not win with a nicer chat widget alone, it won by using LLMs on top of deep data and workflow scaffolding. In both cases, the advantage comes from owning the system, the context, and the action layer, not from shipping isolated AI tricks.
The next step is marketing software that feels conversational on the surface and highly structured underneath. If Customer.io executes, it can widen beyond technical growth teams without giving up the sophistication that made it valuable, and AI becomes the bridge that turns product depth into broader adoption and higher spend per customer.