The highly anticipated AI startup Instinct has just revealed a new potential source of revenue: product recommendations. Unfortunately, since the implementation of this feature, some users have found it uncomfortable.

During the overnight period, this AI proxy began to push product recommendations to users as part of the new feature Instinct Selections. This feature is designed to help users create personalized lists in areas such as dining, travel, and shopping.

Founder Noah Shinn announced the launch of this feature later on Tuesday and explained that the company hopes to collaborate with "local chefs, designers, architects, travel guides," and others to bring "human taste" to the product.

In other words, Instinct no longer relies solely on the suggestions provided by AI based on training data collected from the entire network; instead, it collaborates with real people to provide manually curated recommendations. (It is not specified who these human partners are in Instinct.)

Shinn wrote on X: "Next time you want to find a restaurant, Instinct can provide recommendations from a list hand-picked by chefs who know about the hidden gems in your area."

He continued to write, "If you're looking for a new hiking route, Instinct can help you find several options that suit your preferences with the recommendations of local guides. If you want to add some color to your home, Instinct will find works that match your style, space, and budget from independent designers."

"Our goal is to provide world-class recommendations, which are significantly different in quality from what you might find on the internet and from the content generated by other chatbots," he said.

But these seemingly unsolicited product pushes did not receive entirely positive reactions at first.

AI Investment Institution Array VC General Partner Shruti Gandhi wrote on X: “I don’t need to buy a computer case, sunglasses, or a wide-brimmed hat. This is really disappointing. At least give me a personalized shopping list that I actually need!!”

Another entrepreneur, Andrew Yeung, also confirmed on X that he received product recommendations that he had not actively requested, and this experience felt rather uncomfortable.

The third user, the founder of the business acquisition market platform Chat Joglekar, wrote: "When Instinct started suggesting to me what I should buy based on the content of my emails and the upcoming trip, that's when I felt a sense of 'uh' for the first time regarding AI's assistant."

For planned recommendations to be truly useful, the key is to ensure that they are exactly what the user is looking for, or at least to integrate naturally into the current conversation. In this way, the assistant will appear to be helping, rather than just advertising to you.

If product recommendations, like Instinct, appear at random times like they do now, then they must be highly personalized and impressive enough; otherwise, users will feel as if they are being bombarded with spam.

At a time when it is competing with competitors from larger companies—such as Meta and the current OpenAI—for users, this approach is unlikely to make users prefer this AI proxy any more, whereas those companies do not adopt such tactics.

Instinct has not revealed whether they have made any money from these recommendations so far, or whether they plan to do so in the future, but this feature seems to be very suitable for certain advertising, commercialization, or affiliate marketing models.

TechCrunch has sought further comments from the company on this matter.

Not long before the launch of this product recommendation feature, Instinct just confirmed its latest round of financing: a $1 billion Series C funding round that raised the company's valuation to $10 billion. The company has not yet disclosed the actual number of users of its product, which could help explain this valuation; however, in a recent podcast, Shinn mentioned that the agent often handles transactions for users.

In the same podcast interview with investors Patrick O and Shaughnessy, Shinn also mentioned that the platform's annual transaction volume through agents is approaching $1 billion, of which over 50% is related to travel.