Pinkfish helps companies build AI agents through natural language processing

As the product director of the Customer Service Startup of the Talkdesk, Charayna “CK” Kannan said that companies often say they want to automate different workflows, but that it is really difficult to implement the AI. Companies are dealing with equipped and inherited software, which often has no API, creating a discouraging task that their IT departments did not prioritize.

“All the companies we talked about had between 50 and 1,000 requests for automation from different teams in their request for orders that never arrived,” Kannan (in the photo above right) said to TechCrunch. “This simply makes no sense. At present, you should not have an automation portfolio of 1,000 -line items. You should be able to do it very fast. “

This realization became the momentum behind the new Kannan startup Pink fishthat helps business clients to build AI and other workflows driven by AI through natural language indications. The software has more than 200 integrations, such as Salesforce and Zendesk, and focuses on deterministic execution, which means that the same user notice produces the same result every time.

Kannan said Pinkfish has tried a different approach from competitors when sold to companies. Instead of launching its platform as a golden ticket to automate each workflow, Pinkfish tells companies to test the software only to automate one or two workflows different at the beginning. “So that’s where they start, and then they go from two to four, four to 10, from 10 to 20, and luck 1,000 (automation through Pinkfish),” he said.

Until now, that strategy has been worth it. Pinkfish threw himself in stealth in January 2024 with Kannan as CEO and co -founder Ben Rigby as Director of Products and Technology (Cpto). The company focuses on some areas, including stores and services, and has led hundreds of users and business clients, including IPSY, Elevate and Talkdesk, among others.

Kannan said that, although many new working flow automation companies seek It focuses on the workflows of the critical mission.

She gave the example of IPSY, a makeup subscription service. One of the first workflows that IPSY used Pinkfish to automate was its price application function, which was previously attended by a three -people team. This team would have to meet each request manually, regardless of whether it arrived during the night or the weekend. Kannan now said that the whole process crosses the pink fish.

“It is very critical of the mission,” Kannan said. “If the pink fish is ruined somewhere, guess what, its prices are not on its website. You leave money on the table. “

Now Pinkfish told Techcrunch exclusively that he is emerging from stealth and has raised a round prior to the $ 7.6 million seeds directed by Norwest Venture Partners with the participation of Storm Ventures and Angel Investors.

Scott Beechuk, a partner of Norwest, who will take a seat of the Board in Pinkfish, told TechCrunch that he meets Kannan since his time in Talkdesk and will play Kannan to be an advisor to several Norwest portfolio companies.

Beechuk told TechCrunch that he was excited to support the company because he believes that Kannan and Rigby have the right balance of understanding the underlying technology and understanding the customer base to stand out in a landscape of artificial intelligence agent full of people.

“They are throwing themselves with a lot of significant logos and customers who pay that are finding a real ROI, you support these companies in seed stages, they could take years to deliver a real ROI,” said Beechuk.

Kannan also believes that Pinkfish stands out from competitors because it allows customers to use natural language to request the system while using the complete code in the background to build these AI work flows. She said that while the low code was popular for years, and remains for some of its competitors, she believes that in the current environment it becomes too limiting and is effectively “dead.”

He added that companies do not want to choose between a set of precodified construction blocks, but would have a solution that gives them access to a posterior part of complete code but with a simpler interface to use. As the AI ​​agent market is filled, it expects the message to resonate.

“How can we provide tangible value to the cases of use of critical and complex mission? By connecting it with the agent and determinism, and bringing a platform with the right level of railings for all these connections, ”said Kannan. “I think these are the two areas that we are thinking differently.”