A new a16z report looks at which AI companies startups are actually paying for

Thursday, Andresen Harovits released it The first AI expenses report Fintech Farm in part with Mercury. Using transaction data from Mercury, the report is analyzing the top 50 AI-Native Application Level companies that are spending startups, similar to previously published Top 100 Jenner AI Customer AppThe

A1 16 Z Partner Olivia Moore and Seema Ambble say that data shows that companies are still taking a range of different AI products for specific tasks – and new applications are growing and reading very fast.

“The tools are spread,” said the Ambble. “It didn’t just combine one or two in each category.”

The report also shows a lot of costs for “Human Agriculture” or “Kapilots” that can help increase productivity among the employees, suggesting that the proposed startups advise not to be fully transferred to agent workflow.

“As the use of the computer turns into further mode, there is more power to the end to the end agent flow, I think the shift will happen, where we will see the low coopilt and the last agent tools from the end,” said Ambble. “Especially since people are really interested in trying” to try ”

Surprisingly, the top of the list was dominated by the major labs, the Openai followed the top slot and ethnographic followed by the second number. Vibe-Coding Tools were also well represented, at number 3 and loving 18. The cursor landed on No. 48 and 48, which operates on more enterprise-oriented, which was more enterprise-oriented.

When the A1 16 Z made a similar list for consumer habits, Loveable alone gained much space than the authentic traffic transcript because a lot of people were using it to create a project. However, startups like the lack of enterprise features are as profitable as they are not spending as much money on money. However, different companies on the list seem that there are places for different companies at once.

“This is an open question that is moving forward in sibling coding,” said Moore. Does the space begin to unite, and a place code becomes the best platform for vib? Or is it the case where there will be four or five sibling coding companies that are really big business for different types of applications? We still don’t have the answer.”

Moore was surprised to see the startups by taking consumer-based equipment such as Moro Capkut and Mid Journey.

“We see that they are so many [consumer] Companies are becoming faster and faster in the enterprise because they make such pleasant consumer equipment that people accept and use as individuals and bring them to their team and workplace, “said Moore.

Horizontal applications have made at least 60% name on the list overall and consist of 40% vertical applications. The most popular vertical software companies have fallen into three buckets: sales, recruitment and customer service. However, the report has made progress in many sectors that fought to crack startups from the previous generation.

Moore said, “Whatever may be before, such as service companies or suggestions, is now the software company in the AI ​​era.”

Ambol gave the Ambol Crossby legal as an example, which can quickly review the legal agreement for a user, to replace the key at one point an internal general suggestion, spreading rumors through thoughts and research. He said now that most tools are used to help employees (like co-pilots), with automatic staff (ending agent) to make quick decisions instead of replacing the entire workplace and talent suits.

“As the technology gets better, we have actually been able to create full agents colleagues, you will see that mixture away from the end to the end agents and co-pilots,” he added that AI equipment could do much more work like outreach faster than human cans.

As was also taken in the list of lots of notes on the list Otter.aiRead AI, and Hapiscript – without any single alternative dominance. This is when an ambulatory means when he said that there is not a product that has dominated the market; Rather, startups are still picking “their own taste” to see which tools they like the best. It is also good for employees who, with many options, can choose what applications help them to work best by not using one-size-fit product “from top to bottom.”

The latest search of the report was the growing interconnection of consumer and enterprise business. People are bringing personal applications that they use for work at home and people starting companies are using their favorite private applications to help make their business. Earlier, there was a description between the two: a set stack for what to use when creating a startup.

Ambble and Moore canvas are mentioned as an example: This is a popular consumer application that has a huge enterprise audience. Even the canva has taken a few years to add an enterprise plan. However, if the distinction of the cases of the unique and enterprise use is even stronger, the companies are more interested in mixing the two.

“Your tam [total addressable market] One or the other is no more, but you can sell both, “he said. He said that the companies made by these products can quickly” professionalism “, which means to create enterprise teams, such as go-markets, sales and assistance, so that they can start selling and collecting enterprises without relying on separate customers.

Moore and Ambble are hoping that the list will change rapidly in the coming years. Old companies are now launching AI features to stay relevant and accessible, when new entrances bring new ideas.

“Inheritance player, inheritance almost means, ‘what was in the12 months ago,'” said the Ambble. “If we are pulling it back in 12 months, will the same note-taking applications be there? Or will the whole new set?”

This piece was updated to reflect the applications of notes.