Serendipitously, Chandrasekaran chanced to fulfill a pair of good technologists with many years of name heart expertise, Tod Well-known and Slava Zhakov, who shared the identical imaginative and prescient. Even higher, they’d already written a working prototype. They didn’t suppose their very own trade would get up earlier than it was too late. Crescendo was born.
AI that works with people
Crescendo’s novel method of utilizing AI with people nonetheless within the loop additionally addresses two elephants within the trade’s room. The primary is deflection, which is a well mannered method of describing how onerous corporations make it for purchasers to even learn how to contact an organization earlier than getting shunted into the FAQ and telephone tree wilderness. Then there’s churn. Unsurprisingly, lots of the trade’s jobs are fairly boring, resulting in stratospheric worker churn charges of as much as 50% a yr.
Crescendo’s method to AI appears good. As an alternative of throwing a whole lot of tens of millions of large clusters of GPUs and competing with OpenAI, Google, Meta, and many others., for the most recent PhD graduates in AI from MIT, they’ve put people in cost. For Crescendo, AI doesn’t displace people, it enhances them. On the similar time, Crescendo’s proprietary AI software can nonetheless profit from tapping into the world’s largest LLMs and even the personal information bases of their clients to reply the overwhelming majority of shopper questions. Better of all, clients declare they’ll go into manufacturing with Crescendo in simply two to 4 weeks, and after the primary month, AI is dealing with greater than 90% of the queries mechanically and precisely. So far they’ve but to expertise a single hallucination and there was zero buyer downtime.