Welcome to the AI Index Report, featuring practitioner analyst and entrepreneur Toni Witt. This series is intended to deliver the timely intelligence about artificial intelligence (AI) you need to get up to speed for an upcoming client engagement or board meeting.
Discover how AI has created a new ecosystem of partnerships with a fresh spirit of customer-centric cocreation and a renewed focus on reimagining what is possible. The Acceleration Economy AI Ecosystem Course is available on demand.
Amazon’s AWS re:Invent 2023 conference hosted notable announcements coming from the company.
Amazon has a generative AI platform called Bedrock, which is primarily geared toward enterprise customers already in its ecosystem. At the conference, Amazon announced new guardrail functionality built into Bedrock; it allows users to control what kinds of questions their models will answer, using natural language to limit the LLM response categories.
Another announcement was Amazon Q, a generative AI-powered assistant. With Q, Amazon has hit two major value propositions:
- It plugs into major systems, such as Dynamics 365, Salesforce, and more.
- It only uses information that a particular user is allowed to see, enabling interaction customization based on an individual user’s organizational identity, role, and permission levels.
AI21 Labs, an Israeli enterprise AI company, completed a series C funding round for over $200 million. At a high level, AI21 is a natural language processing company. Its first product was a text rephrasing tool that used generative AI for consumer use.
Now, its main product is the AI21 Studio, which lets enterprise customers easily deploy natural language processing-based applications using models that AI21 custom-trained. The company has a set of use case-specific models. For instance, one is for paraphrasing and another is for summarization.
Because the applications are accessible via an API, it’s easy to get started with AI21’s tools. Pricing is on a pay-per-request basis. Because of this, it may be more enticing to smaller businesses or non-technical teams. While AI21 has simplified the process, users will have to expose their data to an additional third party for the API access. This could pose a security threat.
Solution of the Week (07:04)
Smith.ai is a small firm that handles call answering and inbound sales for SMBs using AI. It uses real people as receptionists. When it comes to virtual agents, Smith.ai is taking a different approach. It understands the unique power of combining the human touch with machine intelligence. It built its own custom software and AI tooling that helps the agents go through conversations with a natural flow, using prompting instead of scripts, for instance.
Additionally, it built a custom AI chat tool specifically targeting FAQs. The feature takes in a document containing the FAQs and then provides a chat that can be put in its place. This lets the end user answer their own questions more efficiently. Smith.ai has noticed a reduction in cost and in time real reps are spending on answering basic questions.