Opportunity:
The concept of a virtual personal assistant has progressed rapidly in recent years. Historically, real-world implementations have included a virtual receptionist directing customers (the digital phone tree), voice-typing software transcribed audio recordings. Apple’s release of Siri in 2011, was the first commercially viable and dynamic personal assistant product directed towards consumers. Since Siri’s debut, large technology companies have released their own versions of broad, voice-based consumer facing personal assistants, including: Cortana (Microsoft), Alexa (Amazon), and Google Voice/Now (Google). On the startup side, leading technologies are text-based and focused on specific verticals – notable companies include: X.ai, Clara Labs, and Julie Desk.
IT research firm Gartner predicts that many touch-required tasks on mobile apps will become voice activated within the next several years. The voices of Siri, Alexa and other virtual assistants have become globally ubiquitous. Siri can speak 21 different languages and includes male and female settings. Cortana speaks eight languages, Google Assistant speaks four, Alexa speaks two. For the first time ever, parity with the human voice has been reached amongst conversational speech recognition and AI systems at an accuracy level of 93.4%. The consumer opportunity for voice technologies is huge, as the number of people worldwide using AI voice assistants is projected to increase to 1.8 billion by 2021.
The even larger opportunity is the market for voice AI products within enterprises. In a joint announcement, Amazon and Microsoft have said that the companies will be combining their technologies to unlock new opportunities within the enterprise. Research firm Tractica estimates that more than 23,000 AI voice assistants will be deployed for customer service applications between now and 2022. The most impactful technologies will deployed upon datasets that are constantly streaming and standardized.
Solution (james.ai):
James.ai is a enterprise AI assistant that leverages artificial intelligence to become the center of the efficient enterprise. By focusing in this space, our company could become the dominant enterprise interface, an claim a powerful dataset. Voice has yet to be digitized and the product could leverage team meetings as a beachhead to build a valuable data asset centered on voice before expanding into broader enterprise use cases.
Specifically, james.ai is an AI-powered note-taking personal assistant that attempts to optimize business meetings by joining scheduled calls and taking note of key action items and follow ups. The technology then circulates a summary of those key moments to participants following the meeting. Over time, the technology will integrate with and automatically update enterprise systems (e.g., Salesforce, slack, trello, etc.), addressing the ongoing challenge of end-user adoption. In the long run, the company will become an open platform for other vertically-focused AI agents to integrate with and the product will become capable of emotion and behavior analysis to further optimize enterprise workflows.
Commercial value:
The Company’s initial product experience would be straightforward. The user journey involves four main steps:
– A subscriber schedules a meeting and adds james.ai (via an email address) to the meeting invite
– james.ai dials into the group meeting and announces his presence
– james.ai silently participates in the meeting by taking note of key action items and follow ups, leveraging natural language processing
– james.ai emails a summary of key moments noted to meeting participants and updates enterprise systems
From a go-to-market perspective, we would leverage a consumerized B2B sales model and focus on sales teams. Client meetings are core to the function and sales professionals are highly dependent on CRM systems, making the group an attractive initial target for an AI personal assistant. The company would greatly benefit from the natural network effects and virality elements driven by group meetings – one subscriber invites multiple attendees to a meeting, naturally exposing other potential customers to james.ai.
Product promise:
To jumpstart the product development process, we could leverage public domain NLP technologies and deploy these technologies through group meetings that are tied to classes that we are taking at Booth. Upon receiving feedback from this process, we can begin to leverage the learning to deploy NLP technologies on larger datasets – for example club meetings.
Upon establishing an accuracy baseline, we can then run A-B tests, applying the NLP technology to different types of meetings that stretch the accuracy of the product. As we experience when and how accuracy declines, we can then better understand which vertical within types of enterprise meetings is best to build an MVP.
Sources:
https://www.engadget.com/2018/04/09/in-pursuit-of-the-perfect-ai-voice/
https://blogs.3ds.com/northamerica/the-rise-of-ai-voice-assistants-in-the-enterprise/
