There’s no scarcity of chatbot providers attempting to earn a spot in the myriad channels on Discord and Slack. California-based Gleen, based by Microsoft and LinkedIn veterans, is providing its enterprise-grade chatbot to essentially the most demanding phase of the market — technical communities, like a blockchain infrastructure channel on Discord, and it has raised some contemporary funding to work on the product.
The slim focus is an effective begin because it pushes Gleen to resolve essentially the most pressing difficulty in massive language fashions right this moment: hallucination. Hallucination, which occurs when synthetic intelligence makes up false data however solutions with confidence, is particularly high-stakes for discussions that contain esoteric matters, as a false however assertive reply may simply mislead individuals.
“If someone says the price of Uniswap [the token of the namesake decentralized crypto exchange] is going to go up to $200, then it can be a massive manipulation of the market,” Ashu Dubey, co-founder and CEO of Gleen, advised Ztoog in an interview. “We decided that we were going to fix this particular problem of hallucination before we could actually be a serious player, so we fixed it.”
Gleen’s imaginative and prescient attracted traders from each the established software program world and the nascent crypto sphere in its $4.9 million oversubscribed funding round. Institutional traders embody Slow Ventures, sixth Man Ventures, South Park Commons, Spartan Group and CoinShares. Among its listing of angels are Anatoly Yakovenko, co-founder of the favored blockchain community Solana; Mike Derezin, former COO of blockchain information supplier Chainlink; Will Papper, co-founder of decentralized investing protocol Syndicate; and ISM Angels.
When requested about competitors, Dubey argued that many chatbots available on the market are mere “wrappers” of ChatGPT and different massive language fashions, so the responses they provide are possible the identical as these from calling the OpenAI API.
That strategy doesn’t resolve hallucinations. Instead, Gleen created its personal proprietary machine studying layer that sources from enterprise data, which may then cross-check LLM responses to keep away from hallucination. LLM is lower than 20% of its tech stack; the remainder of the work goes to how Gleen shops the info and its proprietary system retrieves information to generate essentially the most correct reply primarily based on area data.
Once Gleen’s system is assured that it has a solution, it sends it to numerous LLMs, be it OpenAI, Anthropic or a fine-tuned Llama, to generate a response. All in all, Gleen’s mannequin has been skilled on 100,000 pairs of questions and solutions.
Gleen’s chatbot interacting with customers on Discord. Image: Gleen
“Search is our own proprietary algorithm, and that’s where our secret sauce is,” the founder stated. “The communities and companies, where the subject matter is highly technical or the quality of the answer matters a lot, is where we get the best traction because those companies or communities appreciate what a good response versus a bad response means.”
Every time it onboards a brand new consumer, Gleen must study its area data by gleaning information from its data base, boards, Slack or Discord discussions. The capacity to summary from that data is among the sta0rtup’s strengths, in response to Dubey, because it “[doesn’t] need very clean documentation.”
Gleen acquired its begin by offering its Discord chatbot to a web3 buyer, although now it generates extra revenues from non-crypto customers. Run by a staff of eight workers, Gleen is now serving greater than 10 clients who pay by the variety of conversations generated for the bot.
“Customer support is easily a $10 billion market,” stated Dubey. Gleen is shifting into the medium-sized enterprise market, and its distant objective is to “solve customer service for everyone, from mom-and-pop stores to very large corporations.” The early consumer base is extremely technical, but it surely desires to be “industry agnostic” in the longer term, which then can be an actual check of its AI system’s adaptability.
Going ahead, Gleen plans to spend its contemporary funding on constructing, driving gross sales and advertising. A giant a part of its go-to-market technique can be educating customers on points round hallucination, safety and compliance in the sphere of generative AI, in response to Dubey.
“Though we will have a sales force, inbound will continue to be our biggest channel. That’s also very defensible going forward,” he stated.
Gleen’s path is formed not simply by the fast-evolving AI applied sciences it leverages but additionally the kind of its clients, that’s, corporations coping with cutting-edge, altering applied sciences like itself. The unpredictability nature of those variables presents one of many startup’s greatest challenges.
“As a CEO, you’re investing in a particular technology or taking a product to the market. But what if the underlying technology completely changes one year from now? You have to restart from scratch,” Dubey stated.
“If we are halfway or 25% into research, we just keep an eye out on what’s best out there, so we are not married to the technology. We are married to the customer problem. We want to solve the customer support problem for these new companies that are coming up in the best possible way.”