Collaborative Intelligence Week in Review - 29Sep2023

Best business article(s) I read this week…

• How to Have Better Strategy Conversations About Monetizing Data

  This is an excerpt from a new book titled Data is Everybody’s Business. It describes what companies need to do to use their data to create new sources of revenue via the generation of valuable, actionable insights (often from AI/ML).

Best technical article(s) I read this week…

• Mastering Customer Segmentation with LLMs

  This article, which includes a lot of computer code, describes how to use Large Language Models (LLMs) to improve the performance of customer segmentation analysis.

• Why Retrieval Augmented Generation will become a cornerstone of system design using LLM capabilities

  Describes how the combination of LLMs and indexed data can be used to scale knowledge dynamically, allowing for much larger and more specific corpora to be used in intelligent systems.

Other item(s) of note…

• ChatGPT can now see, hear, and speak

  This past Monday, OpenAI announced that it will begin rolling out new ChatGPT capabilities that enable the technology to interact using voice and images in addition to text. They expect text prompting to be replaced by voice interactions in more natural language.

• IBM releases research insights on AI-augmented work

  This most recent publication from IBM’s Institute for Business Value contains several interesting insights about how to “boost performance with human-machine partnerships.”

• SAP adding a generative AI assistant to its software suite

  This past Tuesday, SAP announced Joule, an “AI copilot” that will be embedded directly into each of its solutions. According to the company, Joule will enable users to “ask a question or frame a problem in plain language and receive intelligent answers drawn from the wealth of business data across the SAP portfolio and third-party sources.”

Coolest thing I saw…

• Shape-changing smart speaker lets users mute different areas of a room

  This is related to AI in that the way that the speakers self-deploy is based on deep learning algorithms. As people return to work, this is a technology that could make the recording and translation of in-person meetings much easier.