Cognition, the company behind the Devin coding agent, has acquired Poke. The goal is to build Poke's conversational model and interactive style directly into the assistant. The move signals a paradigm shift: having the most powerful model is no longer enough. How the agent communicates with the developer matters too.
Conversational UX is emerging as a competitive differentiator in developer tools. Yet many Italian companies, from SMEs to mid-market firms, haven't built this perspective into their AI strategy. As a result, they risk adopting tools that are technically sound but feel cold or hard to use to their internal teams.
At SHM Studio, we follow these dynamics closely. And we turn them into concrete advice for those evaluating AI agents in their digital workflows. In short: an AI's personality is not an aesthetic detail. It's product architecture, and soon it will become a selection criterion for Italian businesses too.
The timeline: Cognition buys Poke, Devin gets a new look
On July 24, 2026, Cognition announced the acquisition of Poke. The news was reported by TechCrunch with a sharp analysis: Poke brings its own conversational model and an approach to user interaction that goes beyond simply executing commands. As a result, Devin, Cognition's coding agent, will get an upgrade that doesn't touch the underlying model but changes how it relates to the developer.
Poke had stood out among AI tools for a more natural, dialogue-driven style. Until now, though, few observers had seen this as a primary strategic lever. So the acquisition surprised part of the market, which is used to judging AI mainly on technical benchmarks.
Winners and losers: who gains from this move
Cognition gains time and positioning. Instead of building a credible conversational UX in-house, it acquires proven expertise. It also brings on a team with a specific vision for the personality of AI agents. The advantage, then, isn't just technological: it's cultural and product-driven too.
Devin's competitors, from GitHub Copilot to Cursor and the new vertical agents, are now under pressure on an unexpected front. Contrary to what you might think, the differentiator is no longer just the quality of the generated code. It's the quality of the experience of working with the agent. So those who don't invest in conversational UX risk losing ground even with technically superior models.
Developers, finally, are the real short-term winners. An agent that communicates more clearly, in context and in a more human way reduces cognitive friction. As a result, real adoption and trust in the tool go up.
What no one is saying: AI personality is architecture, not cosmetics
There's a tendency to treat an AI's conversational style as a superficial layer. In reality, it's a deep architectural choice. The way an agent asks questions, handles ambiguity and flags its own limits determines how much the user trusts it, and how much they really use it.
Recent research by Nielsen Norman Group shows that users abandon AI tools not only when they make mistakes, but when they communicate in ways users perceive as mechanical or unreliable. Personality therefore becomes a retention factor, not just a first-impression one.
Similarly, studies on developer behavior indicate that trust in a coding agent is strongly influenced by how clearly the agent expresses uncertainty. So an AI that knows how to say
Related articles
Discover more articles exploring similar topics, selected to offer you a more complete and stimulating perspective. Each piece of content is carefully chosen to enrich your experience.
