Zylon launches to make generative AI adoption easy for SMBs

The second full year of the generative AI era began in November 2023, but surveys show most enterprises are still just getting a handle on how they want to deploy the technology — and which models and applications to use.

For those who are still in that exploratory phase — especially those businesses that are outside of Silicon Valley and the tech industry — a new startup has emerged to offer a path forward to harnessing generative AI, even without any technical know-how.

The new startup, Zylon, remote but founded in Madrid, Spain, is announcing exclusively via VentureBeat a $3.2 million pre-seed funding round and the public availability of its enterprise-grade, easy-to-use, flexible generative AI web workspace.

The round was led by Felicis Ventures, with notable participation from LifeX Ventures, Zypsy, and a cadre of strategic angels, including Travis McPeak, Kirill Tashilov, and Hans-Christian Zappel.

Founding principles

Zylon was co-founded in 2023 by Iván Martínez Toro and Daniel Gallego Vico, previously the co-creators of PrivateGPT, an open-source project to allow people to use generative AI without any tracking, that has amassed a following of more than 46,000 enthusiasts on GitHub.

Now, with Zylon, the duo aims to offer the capability to build a fully private gen AI workspace for every business that could benefit from it, especially small-to-medium-sized businesses (SMBs) that rely on software, but don’t build it themselves.

“We are focusing on a target audience that is usually forgotten by other AI products, which is non-tech professionals — small or medium-sized companies that do not usually have technical expertise,” said Gallego Vico in a video call with VentureBeat.

The product is built around a modular, customizable but cleanly designed web interface, powered on the back-end by fine-tuned open source models including Meta’s Llama 2 and Mistral’s Mixtral.

This framework allows Zylon’s SMB customers and their end-users, the non-technical employees, to simply upload the documents and files they want gen AI’s help with — say, invoices that need processing or transcripts that need editing and summarizing — and then click on buttons that describe the actions they want to take.

This virtual workspace ensures that users do not require technical knowledge or prompt engineering skills can still leverage the capabilities of gen AI to get work done faster, more efficiently and more affordably, while still maintaining privacy, high security, and accuracy.

Right now, the company is already working with several customers across sectors including healthcare, finance and legal.

“You find there are common workflows and use cases across different industries,” said Martínez Toro. “Especially in SMBs, they don’t need a super specialized tool, they just need some help with tasks that weren’t covered by any other tool before, such as structured report generation that takes ages of manual work to put together, or data extraction, or some help for decision making. These are things across industries, and we are finding those customers super happy and they’re finding a lot of value.”

A foundation in open source AI

The foundation of Zylon is deeply rooted in the creators’ prior endeavors with PrivateGPT. This open-source project laid the groundwork for a platform that not only respects user privacy but also addresses the nuanced demands of non-technical end-users in the gen AI market.

“When we were working on PrivateGPT and getting a lot of interaction with the [business] community, we saw that privacy was one of the issues, but that was not the only one,” explained Martínez Toro. “There was a huge blocker for adoption: the user experience itself. End users were not able to interact with AI, or they had a really tough interaction with AI through chat.”

Similarly, Zylon’s co-founding duo and team of about seven full-time employees are not big fans of the “magic” buttons that many new AI products and features have rolled out lately, denoted by the sparkle emoji.

As Martínez Toro wrote in a blog post earlier this year on the Zylon site: “The magic button interaction is a one-shot interaction: click, cross fingers and get a result. There is no refinement or informed process, generating a big misalignment between the user intention and the end result.”

Instead, they believe their framework of a canvas with different sections that clearly state which actions the user can take is a superior experience for most non-technical SMBs and enterprises.

Simplifying the complexities of Gen AI adoption by the enterprise and SMBs

Martínez Toro, serving as CEO of Zylon, brings a wealth of experience from his tenure as CTO of Ontruck, where he honed his skills in applying AI and machine learning (ML) to enhance software-as-a-service (SaaS) process optimization.

Co-founder and CTO Gallego Vico complements this with his extensive background in software development and his contributions at Amazon.

Together, they have created a tool that encapsulates the essence of user-centric design, aiming to simplify the complexities associated with gen AI technologies.

Zylon is set to bridge this gap by offering a solution that prioritizes user experience and privacy. The platform allows companies to engage with gen AI through a straightforward interface, facilitating tasks such as data querying, summarization and report generation with minimal input.

Secure and supported

The company’s vision has already garnered the support of industry veterans and investors. Viviana Faga, General Partner at Felicis, expressed confidence in the founders’ vision and technical prowess, foreseeing Zylon’s role as a catalyst in the AI ecosystem’s growth.

Moreover, Zylon champions collaborative efforts by enabling shared projects that foster team creativity and productivity, all within a secure SOC2-certified environment.

As Zylon embarks on this journey, its closed beta presents an exclusive opportunity for early adopters to experience and shape the future of gen AI in the workplace.

Zylon’s commitment is to evolving alongside its users, ensuring that the gen AI revolution is not only accessible but also aligned with the real-world needs of businesses across sectors.