7 AI Startups Featured in YC’s Summer Class of 22

At that time of year again. This morning, Y Combinator (YC) hosted a demo day for its summer cohort 2022 – the 35th demo day in the incubator’s history. Featuring founders from 30 countries and startups in various sectors including developer tools, fintech and healthcare, today there has been no shortage of compelling offerings.

Competition has been fiercer than usual, due to YC’s decision in early August to cut batch size by 40% to about 250 companies under adverse economic conditions. But a certain class of startups has emerged: those who apply artificial intelligence and machine learning to solve problems, especially for business-to-business clients.

This year it only had 14 startups compared to 20 last year, which makes sense because the overall pool is also smaller. But the groups share a common theme: sales. Their products largely target sales and marketing hurdles at a time when companies are facing recessionary pressures.

Regardless of the economic challenges, the large routable market makes sales an attractive problem for startups to deal with. Grand View Research tied Sales force automation software market alone at $7.29 billion in 2019.

AI pilot

AI pilot It develops a tool for salespeople that automatically translates call recordings into structured data that then updates your CRM system directly. The idea is to save delegates time, and reassure their managers that the pipeline data is up to date.

It should be noted that other platforms such as fireflies and Microsoft viva sales Do this too. But Max Low, founder of Pilot AI, a former software engineer at Salesforce, says his product is more comprehensive than most, and can generate a summary of each call as well as data points that link to CRM fields and questions delegates ask, as well as key parts of a recipient’s answer.

AI pilot

Image credits: AI pilot


style It’s also in the sales space, but it focuses on text prediction across web applications via a browser extension and server-side API. Typewise was initially developed as a smartphone app, which claims to have Fortune 500 customers in the e-commerce and logistics industries — it can auto-complete sentences, insert smart snippets, automatically reply to messages, and check for style and grammatical consistency.

It looks a bit like TextExpander And the magic. But founder David Eberle says Typewise is compatible with any CRM system and can be customized to company data, with an analytics component that suggests which words and phrases to use.

YC Summer 2022 AI startups that didn’t fall into the sales and marketing technology category tend to focus on development tools, which is another profitable avenue for growth. Considering that 55% of developers struggle to find time to build in-house apps in the first place, according to a recent release exploratory studyventure capital certainly sees an opportunity: hmm invest $37 billion last year for startups that create development tools.

Amnesty International Monterey

Amnesty International Monterey It addresses a completely different part of the product life cycle: development. Founder Chun Jiang presents it as a “leading product development assistant” that replaces documents with workflows that automatically generate product specifications, including feature ideas, metrics, designs, and launch plans.

With Monterey, customers choose a product model based on their use case (eg, “Software as a Service”), input configuration, and check dependencies to resolve conflicts. Jiang says the platform can detect conflicts between teams and dependencies while providing an overview of the portfolio to align features.

Amnesty International Monterey

Image credits: Amnesty International Monterey

AI development tools

AI development tools Perhaps it can be used in conjunction with the Monterey AI.

Dev Tools AI offers a library designed to make it easier to write tests for web applications in current development environments by simply drawing a box over a screenshot. Applying computer vision, it finds elements in web pages such as search boxes and buttons, and can also see controls within web games. It can also test for crawling errors on pages, including broken links, 404 errors, and console errors.

As founder Chris Navrides points out, writing web tests from start to finish is a time-consuming process, requiring a person to search the page’s code multiple times as the tested application evolves. Assuming AI development tools work as intended, they could be a valuable addition to the arsenals of QA testing teams.

Maya Labs

Maya Labs You are creating a platform for translating natural language into code. Similar to GitHub’s copilot, Maya builds programs incrementally and displays results in response to steps written in English.

Sibesh Kar, co-founder of Maya, says the service builds apps using a combination of conditional logic, AI-powered search and classification, tuning language models, and template generation. Currently, Maya can query and plot data from an external source such as Google Sheets, Notion, or Airtable, and perform actions on that data, such as sending an email, uploading a file, or updating a database entry.

The long-term goal is to expand Maya to include tasks such as navigating the web, linking APIs, and automating workflows, which – given the current state of AI text-to-text systems – seem to be within the realm of possibility.


For those who prefer a hands-on approach to programming, Welcome It claims to use artificial intelligence to “instantly” answer technical questions for developers with explanations and relevant code snippets from the web. The platform is powered by large language models (such as GPT-3) that point to multiple sources to find the most likely answers, according to co-founder Michael Roizen.

When Hello users send a query, the service pulls the raw location data from Bing, rearranges it, and then extracts the understanding using the above models. A different set of models translate the results into human-readable answers.


Image credits: Welcome


Another startup with linguistic models at its core is nomind, which provides data scientists, data analysts, and software engineers a tool to create custom models for natural language processing. Leveraging large language models similar to GPT-3, NuMind can be used, for example, to find job offers that best match a particular resume on a recruitment platform.

NuMind founders Etienne Bernard (former head of machine learning at Wolfram Research) and Make.org co-founder Samuel Bernard claim interest in the company is very high, with a paid client base growing to nine within a month.

Leave a Comment