What would you do differently, if you could speak to any website?
There is no product in the market that allows companies to easily integrate voice-based UIs into their web-apps. Yet 58% of U.S. adults use voice technology daily and voice is the most natural way for humans to express their needs.
Verbose AI will transform the way users interact with websites, making voice control an integral part of the web experience. In simple terms, it’s a "Siri for your website".
The global voice user interface market is projected to grow significantly, with a compound annual growth rate (CAGR) of 21.3% from 2023 to 2030. The market size is expected to increase from $19 billion in 2022 to $92 billion by 2030. Despite this growth, currently, no websites have functional voice control, highlighting a vast untapped potential.
Verbose AI leverages recently matured technologies such as large language models (LLMs) for intent parsing and fast voice synthesis. The platform ensures up-to-date site content is used as LLM context, while offering seamless user interface (UI) interoperability. There are no existing comprehensive solutions on the market, giving Verbose AI a unique competitive edge. The rise in consumer expectations, with 58% of U.S. adults using voice technology daily, underscores the timeliness of getting into this market.
Verbose AI envisions an end-to-end natural language chain encompassing all stages of customer interaction, from discovery and checkout to customer service. In the long term, the core technologies created will integrate intelligent agents and in-store AI assistants, supporting natural language based interactions across the digital-physical divide.
Founded by Manu Raivio, a seasoned tech leader with 15 years of experience in high-tech companies like Google and Amazon, Verbose AI is on a mission to make voice control a standard feature on websites.
Verbose AI is a fully remote company incorporated in Estonia through the e-Residency program, and is currently raising €1.5M in pre-seed funding, aiming to secure runway until Q4/2025. The roadmap includes stages from MVP to finding PMF, and scaling the product and team through revenue and subsequent funding rounds.
Are you competing with GPT-4o?
No - it’s just one of many of the models that can be used to answer voice queries. This leaves the business the control over choosing this technology to use in the quickly evolving AI landscape.
Are you creating proprietary technology?
At the outset no, but it’s on the horizon. In particular, fine-tuning smaller language models for the specific purpose of detecting on-screen and off-screen entities is a discrete and largely unsolved problem. These papers (1 and 2) by Apple suggest an early approach that is powering Apple Intelligence for Siri - and we can go in this direction after early success using hyperscaler’s models.
What’s your defensibility strategy?
In addition to first-mover advantage, targeting the developer and product communities with an open-source project that is easy to evaluate, modify, contribute to and audit is a highly defensible approach in comparison with proprietary and vendor-specific or SaaS solutions. Every man-hour of code on the project and functionality is a moat against other similar products. Creating and fine-tuning a dedicated language model that performs extremely well for this particular task creates further defensibility.
Are you just a wrapper for LLMs?
No, but also a wrapper. And a great one at that. The ability to wrap any LLM with comparative performance measurement while gathering 100% privately held intent and issue resolution data from your customers goes way beyond a simple wrapper.
Who’s your target market and what are the use-cases?
We’re unsure at this stage. The hypothesis is that there are two major use cases:
Our developer and community led go-to market will give us strong inbound signalling, while outbound outreach will allow us to be “smelling what’s selling”.
Where does the client value come from?
There’s a number of value hypotheses. The most exciting one is discovering user intent at scale, which has never been possible before. This is a key signal for matching user expectations with their service offering.
Furthermore, developing a keen understanding of customers through their use of language helps segment them better between novice-to-expert and offer targeted support/services.
You will be eaten by Alexa / Siri?
No. Voice based UIs are 100% complementary to on-screen information and largely rely on the elements presented on screen. Pure voice players such as Alexa have struggled to move beyond simple “more toothpaste” types of restocking types of transactions.
Why can’t Apple or Microsoft do this?
They could and sometimes large companies also get behind and start contributing to industry-leading open source projects. The
What’s your long-term vision?
There are multiple opportunities around natural language in the future that excite us. The foundational technologies we’re developing beyond intent to feature/ product matching will be useful for:
Can you share some examples of how this would work?