Because the chief product officer for AI customer support startup TalkDesk, Charayna “CK” Kannan stated that enterprises typically say they wish to automate completely different workflows however that it’s actually onerous to implement AI. Enterprises are coping with clunky, legacy software program, that always doesn’t have APIs, creating a frightening job their IT departments weren’t prioritizing.
“Each firm that we talked to had anyplace from 50 to 1,000 automation requests from completely different groups of their backlog that they simply by no means acquired to,” Kannan (pictured above on the best) informed TechCrunch. “This simply doesn’t make sense. At the present time, you shouldn’t have a 1,000-line merchandise automation backlog. It’s best to be capable to do it actually quick.”
This realization grew to become the impetus behind Kannan’s new startup Pinkfish, which helps enterprise prospects construct AI brokers and different AI-driven workflows by means of pure language prompts. The software program has greater than 200 integrations, like Salesforce and Zendesk, and is targeted on deterministic execution, which implies the identical consumer immediate produces the identical end result every time.
Kannan stated that Pinkfish has tried a unique strategy than rivals when promoting to enterprises. As a substitute of pitching its platform as a golden ticket to automate each workflow, Pinkfish tells the businesses to attempt the software program simply to automate one or two completely different workflows at first. “In order that’s the place they begin, after which they go from two to 4, from 4 to 10, from 10 to twenty, and hopefully 1,000 [automations through Pinkfish],” she stated.
To this point, that technique has paid off. Pinkfish launched in stealth in January 2024 with Kannan as CEO and co-founder Ben Rigby as chief product and know-how officer (CPTO). The corporate focuses on just a few areas, together with retail and companies, and has landed a whole lot of customers and enterprise prospects together with Ipsy, Elevate, and TalkDesk, amongst others.
Kannan stated that whereas many workflow automation startups want to assist firms minimize out a few of the extra “further” features of a job like automating market analysis, or pulling potential gross sales leads, Pinkfish is targeted on mission-critical workflows.
She gave the instance of Ipsy, a make-up subscription service. One of many first workflows Ipsy used Pinkfish to automate was its worth request function, which was beforehand taken care of by a three-person staff. This staff must attend to every request manually no matter whether or not it got here in in a single day or on the weekend. Kannan stated now, that entire course of runs by means of Pinkfish.
“It’s so mission vital,” Kannan stated. “If Pinkfish screws up someplace, guess what, your costs usually are not in your web site. You permit cash on the desk. “
Now Pinkfish informed TechCrunch solely that it’s rising from stealth and has raised a $7.6 million pre-seed spherical led by Norwest Enterprise Companions with participation from Storm Ventures and angel buyers.
Scott Beechuk, a companion at Norwest who might be taking a board seat at Pinkfish, informed TechCrunch that he has identified Kannan since her time at TalkDesk and would faucet Kannan to be an adviser for numerous Norwest portfolio firms.
Beechuk informed TechCrunch that he was excited to again the corporate as a result of he thinks Kannan and Rigby have the best stability of understanding the underlying know-how and understanding the client base to face out in a crowded AI agent panorama.
“They’re launching with a bunch of great logos and paying prospects who’re discovering actual ROI, you again these seed-stage firms, they might take years to ship actual ROI,” Beechuk stated.
Kannan additionally thinks Pinkfish stands out from rivals as a result of it lets prospects use pure language to immediate the system whereas utilizing full code within the background to construct these AI workflows. She stated that whereas low code was widespread for years, and nonetheless is for a few of their rivals, she thinks in right now’s surroundings its develop into too limiting and is successfully “useless.”
She added that firms don’t wish to choose from a set of pre-coded constructing blocks, however reasonably would have an answer that provides them entry to a full-code again finish however with a simpler-to-use interface. Because the AI agent market will get rising crowded, she hopes that message resonates.
“How can we go carry tangible worth to the mission vital, complicated use circumstances? By grounding it with the agent and determinism, and bringing in a single platform with the best stage of guardrails for all of those connections,” Kannan stated. “I feel these are the 2 areas we’re pondering in another way.”
