Instinct’s curated recommendations draw criticism over unsolicited prompts
AI startup Instinct has introduced Instinct Selections, a feature that offers personalised lists of dining, travel and shopping suggestions. The rollout began with product recommendations pushed to some users without an explicit request, prompting public criticism from several early recipients.
Founder Noah Shinn said the feature is intended to combine AI outputs with what he described as human taste. Instinct plans to work with local chefs, designers, architects, travel guides and other contributors, although it has not identified the partners involved. In examples shared by Shinn, the agent could use chef-picked restaurant lists, local trail guidance or independent designers' products that fit a user's style, space and budget.
Curated advice meets an unsolicited delivery model
Instinct's stated objective is to provide recommendations differentiated in quality from material found on the internet and from other chatbots. The initial objection was not necessarily to curation itself, but to recommendations arriving without a relevant request or an apparent need.
Shruti Gandhi, a general partner at Array VC, said she received suggestions for carry-on luggage, sunglasses and a brim hat that she did not need. Entrepreneur Andrew Yeung also said he received unrequested product recommendations. Chat Joglekar, founder of a business-buying marketplace, said suggestions based on email and upcoming trips created his first “ewwww” moment with an AI assistant.
Another user, Ash Thaker, argued that users should be asked whether they want recommendations and should be able to curate their own sources as well as the service's suggested sources. Those reactions underline the difference between an assistant responding to an expressed task and an agent initiating a shopping-oriented prompt from contextual information.
Commercial questions alongside rapid growth
Instinct has not said whether Selections is already generating revenue or whether it will do so in future. The format could lend itself to advertising, commerce or affiliate activity, but the company has not confirmed any of those models. Its launch follows Instinct’s $1 billion Series C funding, which placed the company at a $10 billion valuation after a $1 billion Series C round.
The company has also not disclosed its user count. In a podcast interview with investor Patrick O’Shaughnessy, Shinn said the agent is approaching $1 billion in annual transactions and that more than 50% of those transactions relate to travel. That context makes travel-oriented suggestions a notable use case, while also increasing the importance of matching prompts to user intent.
What businesses should take from the response
For businesses building recommendation features, the early feedback points to a practical requirement: make recommendations requested, naturally relevant to the current interaction and subject to user control. Commercial potential does not remove the need to establish relevance, especially when an assistant draws on sensitive context such as travel plans or email-derived signals.