Gleen's Innovative Chatbot for Discord and Slack Draws Solana Founder in Highly Oversubscribed Funding Round

There’s a growing number of chatbot services competing for attention on platforms like Discord and Slack. Among them is Gleen, a California-based startup founded by experienced professionals from Microsoft and LinkedIn, which aims to serve demanding technical communities—including those within blockchain infrastructure on Discord. Recently, Gleen secured new funding to enhance its enterprise-grade chatbot.

By focusing narrowly on this specific segment, Gleen is tackling a critical challenge in large language models (LLMs): hallucination. This phenomenon occurs when AI confidently provides inaccurate information, posing a significant risk in discussions surrounding intricate topics. A misrepresented fact could easily mislead participants in a technical setting.

“If someone claims that the price of Uniswap [the token from the decentralized crypto exchange] will soar to $200, it could lead to serious market manipulation,” said Ashu Dubey, co-founder and CEO of Gleen, in a recent interview. “We committed to addressing this hallucination issue before positioning ourselves as a serious contender in the market.”

Gleen's commitment has attracted a diverse range of investors in its $4.9 million oversubscribed funding round, including established software leaders and emerging players in the crypto space. Key institutional investors include Slow Ventures, 6th Man Ventures, South Park Commons, Spartan Group, and CoinShares. Its angel investors feature notable figures such as Anatoly Yakovenko, co-founder of the Solana blockchain, Mike Derezin, former COO of Chainlink, and Will Papper, co-founder of Syndicate.

On competition, Dubey pointed out that many chatbots are merely “wrappers” around models like ChatGPT, offering solutions that do not effectively address hallucinations. In contrast, Gleen has developed its proprietary machine learning framework, sourcing data from enterprise knowledge and cross-checking LLM outputs to minimize inaccuracies. LLMs constitute less than 20% of Gleen's technology stack; the bulk of its efforts focus on data storage and retrieval systems that ensure precise answers drawn from extensive domain knowledge.

Once Gleen's system identifies a potential response, it consults various LLMs, such as OpenAI, Anthropic, or a fine-tuned version of Llama, to generate the final answer. Overall, Gleen's model has been trained on 100,000 question-and-answer pairs.

“Search is our proprietary algorithm, which we consider our secret sauce,” Dubey explained. “The communities and businesses where technical precision matters significantly are where we find the most traction, as these users deeply value accuracy in responses.”

Every time Gleen onboards a new user, it learns the specific domain knowledge from its knowledge base and discussions on platforms such as forums and Slack. This capability to extract meaningful information, even from less-than-ideal documentation, is a core strength of the startup, according to Dubey.

Initially, Gleen launched its Discord chatbot for web3 clients, but it has since diversified, now generating notable revenue from non-crypto sectors. With a lean team of eight, Gleen is currently serving over 10 customers, who are billed based on their conversational volume with the bot.

“The customer support market is poised to exceed $10 billion,” Dubey noted. Gleen is now targeting medium-sized enterprises, with the long-term vision of revolutionizing customer service capabilities across various business sizes, from small local stores to large corporations. While its current user base is highly tech-focused, the goal is to become industry-agnostic, which will be a true test of Gleen’s AI adaptability.

With the recent funding, Gleen plans to invest in product development, sales, and marketing. A crucial part of its go-to-market strategy involves educating users on hallucination, security, and compliance challenges related to generative AI, Dubey stated.

“While we will have a dedicated sales team, inbound leads are expected to remain our strongest channel, providing a solid foundation for future growth,” he added.

Gleen's path is influenced not only by the rapid advancements in AI but also by the nature of its clients, who are navigating similarly evolving technologies. This unpredictability presents significant challenges for the startup.

“As a CEO, you're always investing in new technologies or bringing products to market. But what happens if the underlying technology shifts drastically within a year? You could find yourself back at square one,” Dubey explained.

“If we are midway or a quarter through our research, we stay vigilant about emerging technologies, ensuring that we are not tied to any particular solution. Instead, we remain dedicated to solving our customers' pain points, particularly in customer support for innovative companies.”

Most people like

Find AI tools in YBX

Related Articles
Refresh Articles