Airbnb has been advancing cautiously on the user-facing AI front, but embracing the technology with enthusiasm on the engineering side. On its second-quarter earnings call, co-founder and CEO Brian Chesky said AI has reduced by 60% the time between conceiving an idea and shipping a feature. The company had already revealed, earlier this year, that AI writes 60% of its code.
"Today, we're building, testing, and iterating faster than we could just a year ago," Chesky said, summing up the internal cultural shift. For a platform Airbnb's size, this acceleration is not trivial: it means more experiments, faster fixes, and a much leaner product cycle. AI is being used in search, discovery, and support, and the executive said the company will finally begin testing AI-powered search.
The search test is the highlight for the end user. Airbnb is adding a toggle that lets users switch between traditional filter-based search and the new AI search, where users type natural language and get results in a visual format. The decision to make the feature optional is significant: the company does not want to force the change on millions of travelers used to the current flow. It is a gradual adoption strategy that reduces the risk of rejection — and also allows comparing engagement metrics between the two flows.
There is also an operational efficiency data point that shows AI's broad impact. According to reports from the call, automated AI support already resolves about 45% of requests and reduced support costs by 16%. Combined with the speed gain in development, the effect is twofold: the company delivers more with less, while improving the experience for those who need help. For a marketplace business, where trust between host and guest is everything, faster and cheaper support has enormous strategic value.
The open question is how much this internal advantage translates into competitive advantage in the market. Airbnb was not the first to bet on conversational search — travel competitors and commerce giants are already exploring the same territory — but the combination of internal acceleration and cautious product testing suggests a more sustainable approach. If AI really cuts iteration time in half, the question changes: it is no longer who launches first, but who learns and corrects fastest. And in that game, Airbnb's engineering speed may be the hardest advantage to copy.
There is a risk the company will need to manage carefully. Adopting AI at speed is no guarantee of getting it right — model-generated code can introduce bugs, inconsistencies, or technical debt if not reviewed rigorously. Airbnb's challenge, like that of any company accelerating this way, is to maintain quality while multiplying speed. If the strategy works, however, the business model shifts subtly: the differentiator stops being just the marketplace and becomes the ability to experiment and iterate in weeks what once took months. In that scenario, AI is not a product tool — it is the company's engine.
Whatever the outcome of the search test, the deeper signal from this earnings call is that Airbnb now treats AI as an operational multiplier rather than a consumer gimmick, and that distinction is likely to shape how the company competes for years to come.
Sources: TechCrunch, CNBC, Dataconomy
✓ Independent sources cross-checked and verified before publishing
