Kindo: The Story Behind This Rapidly Growing AI Security Company

By Amit Chowdhry • May 13, 2024

Kindo is a company that provides a way to securely adopt and centrally manage AI across your entire workforce. Pay only once for AI access from any model, use it across your organization, all in one easy-to-use interface. Supporting commercial, open source, and private AI models as well as 200+ SaaS integrations, Kindo empowers AI workforce productivity while keeping everything secure, compliant, and centrally managed for IT and Security teams. Pulse 2.0 interviewed Kindo co-founders Bryan Vann (CTO) and Ron Williams (CEO) to learn more about the company.

Background Of The Founders

(Left: CEO Ron Williams / Right: CTO Bryan Vann)

Pulse 2.0 (Amit): “What is the background of the founders?”

Kindo (Bryan Vann): “My academic background is in distributed systems. I got my BS/MS from UCLA in computer science. I worked at a few startups between undergrad and grad school but joined Google right after graduating with my masters. At Google, I worked on Google Photos, Technical Infrastructure, and Google Drive. I spent the majority of my time on Google Drive and was one of the original architects of the system, before later moving into organizational management. By the time I left Google I was a Senior Director managing the majority of the Drive engineering team. During my time on Drive we first spent time building out the consumer product, and then later pivoted as an org to focus on enterprise. I led the team through this transition including building out enterprise security controls, and later led efforts to support various industry and global compliance and security regimes.”

“My ‘20% project’ while at Google was leading a tech outreach program in LA where we partnered with a number of accelerators to host pitch events, mentor startups, and help in any way we could to build a thriving LA startup scene. I’ve always been passionate about startups and had just been looking for the right time to jump in.”

“That moment came in the fall of 2022 with the start of the genAI revolution. I had already decided to leave Google by then, but once genAI took off I decided I had to be part of it and narrowed my search exclusively on early stage Gen AI startups. Ron had started Kindo (formerly Usable Machines) a few months before I joined, but I really liked the pitch was exactly what I was looking for. So I joined in April 2023 as co-founder and CTO.”

Kindo (Ron Williams): “I started my career as a programmer in the Air Force and later moved into IT and security leadership. I’ve been the IT and Security executive for 3 tech unicorns over the last decade – from early days of the companies to billions of dollars in revenue and market valuations so I understand the pain that CIOs and CSOs deal with when new technologies waves happen.”

Formation Of Kindo

Pulse 2.0 (Amit): “How did the idea for the company come together?”

Kindo (Bryan Vann): “Ron conceived of the idea after stable diffusion took off. Given his experience as a CIO and CSO, he knew that open source was going to take off in this space, and also that large enterprises would be looking to adopt these tools, but would be concerned about these AIs interact with their private data. So the original pitch was we provide a way to securely control the orchestration of AIs in an organization. Since the industry is so nascent, product strategy evolved to ship with an end-user product for interacting with the AI models providing a single place for employees to take advantage of the productivity gains from AI tools, while providing the IT and Security admins a single pane of glass to control and manage the flow of data into AI models, and have complete visibility into how users are interacting with the AIs and what the AIs are doing.”

Kindo (Ron Williams): “In August of 2022 Stability AI launched the open source image AI called Stable Diffusion. The open source developer community exploded around it and it was clear that AI had reached the crucial moment where a tech is accessible, affordable and useful enough for broader adoption. This was the big moment for AI not the launch of ChatGPT later which only launched in response to the tech community being so excited about Stable Diffusion. Seeing moments in the past where new tech waves started in my career I quickly came to the conclusion that the AI market was getting ready to have a Cambrian explosion of startups that CIOs and CSOs would soon have to figure out how to wrangle as it flooded into their organizations. So we started Kindo to help.”

Favorite Memory

Pulse 2.0 (Amit): “What has been your favorite memory working for the company so far?”

Kindo (Bryan Vann): “What I’m loving most is the blistering pace of innovation. We’re riding the bleeding edge of technology and working to incorporate those advancements into the product as quickly as possible. For example when Google’s new Gemini Model came out, we hooked it into our product on the day it was released, and then followed it up the next day with their vision support adding the ability for our product to understand and interact with images. It’s super fun to be able to move that fast.”

Kindo (Ron Williams): “There have been some great moments but I think signing our first customer has been my favorite so far. It’s always a big deal to get your first believer and learn how they are transforming their business with AI.”

Core Products

Pulse 2.0 (Amit): “What are the company’s core products and features?”

Kindo (Bryan Vann): “Kindo is the first AI orchestration platform specifically built for Enterprises. With compliance, security, and management woven into our DNA, we allow business knowledge workers to directly benefit from AI’s incredible offerings in a fully secured and governed no code environment.”

“For the end-user, there are 5 core aspects to our product. 1) Users can interact with any model they like from the same chat interface since we support all the major commerical models and the popular open source ones today (you can add any model), 2) Users can upload connect to their private data (e.g. Google Drive) and we’ll index it allowing users to ask questions of their data or chat with individual documents, 3) users can build assistants which are like building their own chat bots by specifying a system prompt and attaching a specific knowledge store, 4) users can build agents which are like LLM driven workflows. Using agents, customers can build a customized solution for repeatable tasks, or those that would involve processing a large amount of data that would be infeasible for a human to do manually. 5) We support and manage plug-ins like developer IDE coding models so software developers can use any coding model they want inside their tools without having to expose company code to third party AIs.”

“On the security control side, we follow all of the industry best practices for security, authorization, and authentication. We have a dashboard which allows the admins to control which AI models are allowed in their organization, which data sources can be connected, and what individual users are allowed to do in the product. We also have audit logging which allows admins to gain complete visibility into everything happening within our product such as which models are  being used, how users are querying them, what data is flowing where, are their security or compliance issues happening, and what agents are being used.”

“Combined this gives customers a one-stop shop for their business AI security needs. Due to the generic nature of the capabilities our platform offers, users can build verticalized solutions to their specific problems in minutes. This allows admins to adopt a single platform that can meet all of their AI needs with a single control and visibility point.”

Kindo (Ron Williams): “We are a full stack AI orchestration platform for IT and Security team to control all the AI capabilities of the company by allowing fully controlled and audited accsss to any AI model (commerical or open source), any data source inside or outside the company and all of your employees individually customized for who is allowed to use which model with which data source. We offer a standardized chat interface that works with any LLM, a no code agent builder that allows any employee to apply AI to repetitive tasks, privacy filtering of data for PII, HIPAA, PCI-DDS, etc.. All of these features are turn key, no developer or data science people needed, and can run in the cloud or on-premise to ensure your data is never training another company’s AI on learning how to run your business.”

Challenges Faced

Pulse 2.0 (Amit): “Have you faced any challenges in your sector of work recently (and how did you overcome those challenges)?”

Kindo (Bryan Vann): “Things are moving so fast that there aren’t industry best practices established yet for many aspects of running production AI systems, especially those tuned to meet the specific needs of customers. We’ve had to do a lot of research and experimentation to find the best solutions as well as connect with others in the community to compare notes.”

Kindo (Ron Williams): “Keeping up with the rapid changes in the AI world is a lot of work. I have never seen any tech improve and expand so fast as AI has over the last 18 months. It’s currently improving at double the rate of Moore’s Law which means we are seeing 2x the quality and 4-10X cost reduction every 6-9 months.  Since the launch of ChatGPT about 17 months ago we now have over 20 open source models that outperform it across most if not all of the benchmarks used to rate model quality while open source models are now offering 10-40X lower cost to use then ChatGPT. We are also seeing businesses adopt many different AI models to tackle different use cases, cost, compliance and security issues.”

Evolution Of Kindo’s Technology

Pulse 2.0 (Amit): “How has the company’s technology evolved since launching?”

Kindo (Bryan Vann): “We had an early prototype with a completely different UI when I joined. It had some of the security controls, and had a rudimentary workflow builder. Over the past 14 months, we’ve launched Chat, the ability to ask questions over data you’ve connected to our system, we’ve completely overhauled the UI design, built a new more fully featured workflow builder (now called agents), and have launched a chat assistant builder. We’ve also built out many more security and user management controls, and have added more than a dozen additional powerful models into the product to help everyone get started. We’ve also done some backend rearchitecture so that we can now be deployed on-premise for customers that want complete control over the system which is important especially for government contracts.”

Kindo (Ron Williams): “Our experienced team of engineers and AI researchers have been writing code for about 20 months now so we have built a lot of capability and have totally overhauled the user interface.”

Significant Milestones

Pulse 2.0 (Amit): “What have been some of the company’s most significant milestones?”

Kindo (Bryan Vann): “Mostly answered above with all the launches. I’d also add to the list hooking up Gemini Pro Vision so we have image recognition, and we’ve started using LlamaIndex for better information retrieval from customers private knowledge store.”

Kindo (Ron Williams): “Riot Ventures (not associated with Riot Games where I was an executive) believing in us and giving us our preseed check in October of 2022. Our first free user in December of 2022, our Seed round led by Riot Ventures and Eniac in June of 2023, our first paying business in July of 2023, and our totally new product launch in August of 2023 have all been great moments to be a part of. As we sign each new customer it is still a big milestone for us.”

Customer Success Stories

Pulse 2.0 (Amit): “Can you share any specific customer success stories?”

Kindo (Ron Williams): “We don’t today talk about specific customers but we have been supporting well known brands across a range of markets since July of 2022 and have users from over 1,000 companies and other organizations which provide us with helpful insight into how knowledge workers are using AI to improve their work in a secure enterprise ready platform.”

Funding Information

Pulse 2.0 (Amit): “Are you able to discuss funding and/or revenue metrics?”

Kindo (Bryan Vann): “We announced that we closed $7 million seed round in September. We have revenue but can’t disclose specifics, but we sell to enterprise customers  and our customer base is growing rapidly.”

Total Addressable Market

Pulse 2.0 (Amit): “What total addressable market (TAM) size is the company pursuing?”

Kindo (Bryan Vann): “Given that the overall GenAI market is projected to be a few trillion dollars in the next few years, we’re estimating that our TAM is above $50 billion.”

Kindo (Ron Williams): “We think the impact of AI is going to be massive across every company with the entire software stack being obsoleted, rebuilt or disrupted. Kindo is positioning to be the go to platform from easy and secure AI rollout across companies with significant security and compliance concerns. So we believe the TAM for Kindo is over $50 billion.”

Differentiation From The Competition

Pulse 2.0 (Amit): “What differentiates Kindo from its competition?”

Kindo (Bryan Vann): “A few things differentiate us from most other players in the space:

  • We are AI model agnostic including all the commercial models and any open source ones so our customers don’t have to be locked into any one model provider
  • We allow customers to bring their own custom models and plug them into our product. This allows customers who have built custom models but that don’t have a great UI or all the enterprise IT controls to deploy and manage across their employee base a way to do so.
  • Our platform was designed from the ground up with enterprise security and IT control in mind. We’re really the only way to securely manage and control data flowing across many models, agents and sources of data. We can also run fully on-premise which is a key differentiation from most competitors.
  • We are a full stack turn key deployment focused on helping our customers solve their end-to-end problems. There’s a lot of tools and services out there that are helpful, but it’s generally not straightforward to solve a problem end to end such as automating a whole business process and almost all of the other solutions in the market require your own software developers to implement. We’re helping our customers not only automate repeatable or mundane tasks, but we allow them to accomplish things at a scale and cost with AI that they never could before.

Kindo (Ron Williams): “We are 100% focused on solving the biggest problem of large companies adopting AI: centralized IT management of employee access, centralized security controls to protect company data, and centralized compliance capabilities to deal with existing regulations and the new regulatory burdens targeting the use of AI.”

Future Company Goals

Pulse 2.0 (Amit): “What are some of the company’s future company goals?”

Kindo (Bryan Vann): “We’re doubling down on agents this year which will allow a significantly greater level of automation than is currently possible. We are also investing in information retrieval quality improvements so that more complex user queries and information in bespoke formats or data sources can be seamlessly utilized. We’re also going to be working on a number of AI driven security controls to increase intelligence inside our platform to ensure employees have the safest and easiest to use AI experience.”

Kindo (Ron Williams): “Kindo is on a mission to be the most trusted platform to securely manage and deploy AI across the enterprise.”

Additional Thoughts

Pulse 2.0 (Amit): Any other topics you would like to discuss?

Kindo (Bryan Vann): “Just that we are super excited and passionate about being part of this AI revolution and blazing the trail ahead. We use our product countless times a day and we’re passionate about how we can help guide large enterprises through this transition in a way that allows them to revolutionize how they work and drastically increase the productivity of their workforce while doing so with the peace of mind that all their data is in a fully secured and controlled environment.”

Kindo (Ron Williams): “As a three time top security leader of tech unicorns I believe that business and IT leaders should be much more careful about which AIs employees are allowed to use and what data and use patterns are allowed to leave the control of the company into those AIs. Using someone else’s AI is not like storing your data on someone else’s cloud. You are teaching that AI provider your entire business, how you run your business and how each employee does their job. You are betting the future of your business when you are trusting someone else’s AI model – so this is not just about a data breach risk like when you store your data in the cloud, this is about equipping someone else’s superintelligence on how to compete with you and hoping their promise to not train on your data (whatever they mean by that) is real (it’s really hard to sue someone when you are out of business). Security aside it’s clear AI is a huge transformation wave happening quickly, businesses that invest in controlling their own AIs will emerge much stronger then those that don’t. Kindo is built to make it as easy as possible to control your company’s AI future.”