Mark Hay and Ethan Ding wish to make each enterprise resolution data-driven. Formidable? Actually. However the two engineers, who met a number of years in the past through the pandemic, are downright optimistic.
Hay and Ding are the co-founders of TextQL, a platform that connects an organization’s current knowledge stack to massive language fashions alongside the traces of OpenAI’s ChatGPT and GPT-4. The concept, they are saying, is to present enterprise groups the flexibility to ask questions on their knowledge on-demand, utilizing instruments that – in Hay’s phrases – “the ‘nouns’ and semantics of their groups to know.”
“Information leaders have been fed a false promise for 15 years… Half of the Fortune 500’s Chief Information Officers are allergic to the phrase ‘self-service’ proper now,” Hay, TextQL’s CTO, informed TechCrunch in an e-mail interview. “Their 400,000 knowledge scientists spend 40% or extra of their time accumulating one-time knowledge requests, and their enterprise groups use phrases that seem in a different way of their databases, leading to months of misplaced productiveness when discussing numbers.”
Hay, previously an engineer on Fb’s machine studying group, and Ding, a former member of Bessemer Enterprise Companions’ knowledge group (and a fan of gardening metaphors), thought they might give you a greater answer.
In 2022, they launched their effort in TextQL, which makes use of a knowledge mannequin to map an organization’s database to the ‘nouns’ that symbolize a buyer’s firm of their language – for instance phrases like ‘order’, ‘merchandise’, ‘vendor’, ‘SKU’, “stock” and so forth.
TextQL connects to enterprise intelligence instruments and directs customers to current dashboards when a query has already been requested. It could possibly consult with documentation from enterprise knowledge catalogs like Alation, Hay says, in addition to notes on platforms like Confluence or Google Drive.
Particularly, this enables TextQL customers to ask a chatbot questions reminiscent of “Are you able to present me an inventory of orders that have been very late?” and “Calculate the distribution facilities with the very best focus?” Along with answering questions, TextQL can – by an automation element – take sure actions, for instance sending an e-mail to managers about specified knowledge.
“In an financial atmosphere the place everyone seems to be attempting to do extra with much less, we can provide entrepreneurs superpowers on one platform,” mentioned Hay.
Hay – the place TextQL competes with distributors reminiscent of Palantir and C3.ai – says TextQL at the moment has a half-dozen clients in healthcare, life sciences, monetary providers, manufacturing and media. Annual recurring income is “six figures,” he claims, giving TextQL “a number of years” of runway.
“The slowdown hasn’t affected us a lot; corporations are enthusiastic about our software program as a result of it will possibly assist them do extra with their diminished workforce,” Hay mentioned. “Our whole group is made up of skilled founders who have been beforehand enterprise capital backed, and that is expertise that is fairly troublesome to amass outdoors of this atmosphere.”
By way of enterprise capital backing, TextQL, which has a group of ~10 folks, has raised $4.1 million by pre-seed and seed rounds led by Neo and DCM with participation from Unshackled Ventures, Worklife Ventures, PageOne Ventures and FirstHand Ventures and Indicator Fund.