The Draper Family welcomes three companies to pitch their idea and reveals which company will make it to the semi-finals....
Problem
Every voice AI interprets human language differently
The concept of voice AIs is promising—simply ask a question and receive an answer immediately! Unfortunately, the execution isn’t as easy as it sounds. Users don’t know which voice commands are available, and AIs aren’t actually that smart.
There are two major AI bias problems. The first is input. Every user speaks slightly differently. For example, accents, slang, and vocabularies can all confuse an AI. The second issue is interpretation. Every AI is designed for a different use case, and there isn’t a universal AI for every need. We need a protocol to exchange information about voice commands that both humans and machines agree on. How do we connect AIs for 50+ input languages and 200+ use cases?
Solution
Teaching AIs to seek help from other AIs
whoelse.ai is a platform devoted to improving the performance and intelligence of AIs. We provide voice AIs with simplified grammar that allows them to exchange information in a standardized format. From there, voice assistants can look up other voice assistants if they cannot respond to a request themselves. Think of how browsers learned to look up websites—that’s exactly what we do for voice AI services.
The idea of whoelse.ai can be compared to the original World Wide Web (WWW) proposal by CERN scientist Tim-Berners Lee. He suggested an address system to access distributed files that later became known as the Internet.
We adopted this idea to create the "Internet of Voice." Unlike closed source AI (e.g. Alexa, Siri), whoelse.ai creates an interconnected network of voice assistants.
Product
Our API helps AIs process voice commands in standardized formats
whoelse.ai works like a directory for voice AIs. Users always speak to the best voice AI available, because we provide voice AI ecosystems with a shared e-commerce architecture. From there, whoelse.ai enables voice assistants to forward user requests to the most appropriate AIs available.
whoelse.ai makes natural language processing (NLP) privacy-by-design: Instead of full-text and audio files of voice commands, only a simplified version of the intent is exchanged between voice AIs.
The service directory currently includes the 200 most common intents in human language. The namespace API is available via an ISO-compatible API.
Traction
Completed DIN Standard and pilots with international industry partners
We've been working on our technology since our demo app launched in 2018, and have come miles since that first app. whoelse.ai has participated in accelerators and pilots at OEMs such as Next AI Toronto, Telekom Hubraum, and Creative Destruction Lab Oxford. In 2020, we published a DIN standard for NLP API interoperability. This makes it easier for industry clients to implement our API. Our company has garnered attention from the AI world, and we have been proud to present our work at e.g. O’Reilly AI and BOCOM.
As seen on:
Customers
With whoelse.ai, users will always speak to the best AI available
Our product was built with OEMs, telcos and smart cities in mind. We want to pivot from an outdated form of search to a newer, more efficient model. Our product makes it so that users always speak to the best AI available, with built-in privacy and infinite possibilities.
Our initial focus is on voice AI industry projects. In the future, we believe language interoperability will include a growing number of domains, not limited to just voice.
Business model
Like AdWords, but for voice: we built an exchange for spoken intent
With whoelse.ai, we build and exchange intent. Everything users say can be a potential e-commerce order, and whoelse.ai provides OEMs with an incentive-based business model. With our financial ecosystem, we project revenue of $15M+ by 2023.
Market
Voice is the future of search
The AI market is booming. Within 5 years, over 50% of search requests will start as a voice command, and it is not hard to envision a future where every initial customer service contact is with a bot. whoelse.ai provides a new market solution in the voice AI e-commerce industry, which is projected to reach $55B+ in 2025.
Competition
Marketplace for voice assistants
whoelse.ai establishes a shared registry of voice AIs. As a marketplace for voice assistant-based products and services, we solve the same issues faced by voice AI integrators across industries. Voice AI integrations are currently slow, expensive, and risky. Companies and organizations relying on third-party voice AIs don't know which technology will be the industry leader a couple of years from now. We enable B2B users to combine voice AIs from different vendors in a more flexible way. This makes whoelse.ai complementary to existing voice AI solutions in the market. Our value proposition is our ability to work as a language interoperability provider.
Vision and strategy
Becoming the universal leader for AI queries
whoelse.ai has big plans for the future. Our initial focus is on building a network of connected voice AI technologies that support our protocol. The long-term vision is for our brand to potentially become a universal wake word available across different voice AI ecosystems as well for consumers.
We believe users do not want to remember unique wake words (e.g. Alexa, Siri) for different voice AI ecosystems.
Instead, it's much more appealing to communicate with all AIs in a language that is comprehensive enough to capture intent quickly. This makes "who else?" questions particularly well-suited to become a universal query available across all voice AIs.
Funding
Bootstrapped to date
whoelse.ai has been bootstrapped since inception. We're currently raising $750K–$1M to co-finance B2B sales and the development of an enterprise version of the namespace technology. With your investment, we will go live with both a B2C and B2B product launch.
Founders
Geeks with a passion for branding and AI usability
whoelse.ai was founded by Tobias Martens and Sven KrĂĽger.
Tobias has 8+ years of experience working in startups, consulting, and government organizations. In his last role, he managed digital transformation projects and corporate ventures for international industry clients with a focus on corporate education.
Sven was previously CMO of T-Systems at Deutsche Telekom and has 20+ years of telecommunication industry experience. He was also a co-founder of AI Guild, an international network of data scientists.
Our team is currently based in Berlin, Toronto, and Singapore. We work with a fantastic advisory board that includes Internet TLD entrepreneurs, gaming industry veterans, and deep tech M&A experts.