As 2023 draws to a close, it’s hard to believe that generative AI has been a “thing” in the public consciousness for barely over a year. Since OpenAI publicly launched ChatGPT 13 months ago, it feels as if there’s been talk of nothing else in the tech world – in the media, among analysts and consultants, in boardrooms, and within IT departments across all corners of business and industry.
Generative AI for enterprise knowledge management is now a reality
Indeed, generative AI promises to provide a valuable technological aid in a wide range of business functions and use cases. One early example is knowledge management, where generative AI, properly implemented, can help users find specific information and mine insights from vast knowledge stores by answering direct questions in plain language. It has effectively become the new user interface for search engines.
Let’s quickly review generative AI’s wild ride and how we got to the point where, today, a business researcher can ask the computer a market and competitive intelligence question and get an accurate response to jump-start their investigation.
Just a year ago, generative AI wasn’t ready for the enterprise…
November 30, 2022 – OpenAI releases ChatGPT, based on GPT-3.5 Turbo, and touches off a publicity tidal wave. However, at that time, GPT-3.5 Turbo was not available in the Enterprise API, and no pricing was available for the Enterprise API. A further limitation to ChatGPT’s utility for enterprise applications was that the model was trained on circa 2021 Internet content and generated text based on Internet content, with all its inherent variability, and it offered no citations for the information it included in its auto-generated text, so no statements could be readily verified.
January 23, 2023 – Microsoft invests $10 billion in OpenAI.
February 7, 2023 – Microsoft releases Bing Chat based on GPT-4 with citations and updated content.
The two events that enabled generative AI in enterprise knowledge management
March 1, 2023 – OpenAI announces Enterprise API access to GPT-3.5 Turbo and pricing.
March 9, 2023 – Microsoft announces Azure API access to OpenAI models using the same pricing as OpenAI.
These last two events are what opened the door to the embedding of generative AI capabilities into enterprise applications, and many software providers (Northern Light included) jumped at the opportunity. The advantage Northern Light had in getting to market quickly – we announced our generative AI solution for market and competitive intelligence question answering, based on GPT-3.5 Turbo, a month after the Enterprise API was released – was that we’d been developing a question answering solution using Google BERT (a neural network using a technical structure that predates large language models) for well over a year when GPT-3.5 Turbo became “enterprise ready”; it was relatively easy for us to switch from BERT to a large language model on the fly and quickly release our offering to our SinglePoint clients.
Generative AI is catching on in the enterprise
Now, fast-forward to a year after OpenAI first released ChatGPT to the public. As everyone not living under a rock knows, there’s been considerable corporate drama at OpenAI over the past month (the potential ripple effects of which are unclear) to close out the “year of generative AI”. But this much is certain: there is no turning back the clock on 2023 and the impact that generative and conversational AI is having on the enterprise, in knowledge management and many other business functions. And, with proper data security and accuracy guardrails, that impact will only grow in the years ahead. We can’t wait to see what 2024 has in store!