VENDORiQ: Why Microsoft GPT Superpowers Go Far Beyond Bing

The Latest

14 February 2023: Microsoft’s GPT3.5 (and future) generative AI will be embedded in its search engine, Bing, and web browser, Edge. For Bing, Microsoft claims that it is already capable of providing more accurate results, summarised answers, interactive chat experience and AI-generated content. Edge, on the other hand, now has AI-enabled chat and compose features.

Why it’s Important

After investing significantly in OpenAI (developers of ChatGPT, Dall-e) and with billions more come, IBRS expected Microsoft to rapidly leverage these AI technologies throughout the Microsoft 365 and Business Apps. However, the speed at which Microsoft is moving is impressive, though not without risks, as Google’s fumble with Bard demonstrated spectacularly. 

A 2018 advisory paper, DIY or ready-made? Choose your AI adoption path carefully, predicted the advent of AI at a cost-performance ratio that enabled vendors to bundle the services into existing productions, would be met with zero resistance by users. We anticipated that as soon as AI services became available within existing Cloud-based products, the adoption would be rapid and (importantly) largely outside the control of enterprise IT departments. This is exactly what we are seeing now. 

The upgraded Bing is currently available on limited preview and is expected to be scaled to millions of users in the coming weeks. It heralds a wave of AI-powered applications not just for enterprise users but for mass adoption. IBRS expects many more announcements of this kind from Microsoft in the coming months.  

In particular, the limited release of the newly revamped Azure AI services will create new opportunities for organisations currently developing on the Azure platform and those using the Power Platform, which IBRS sees as a low-code front-end to Azure.  

Despite the speed at which Microsoft is leveraging AI in its existing products, it is actually taking a cautious approach. It is articulating (and evolving) it’s terms of use for AI. No doubt, it has learned from its painful experiences with Tay, a chatbot that became a racist powerhouse. At the heart of this lesson is that AI is only as good as the data it consumes (the corpus) and can accentuate and perpetuate the (all-too-human) bias inherent in the data. For this reason, Microsoft is being cautious in how it approaches AI ethics, while also speeding ahead with (hopefully) less-controversial technical enhancements.

For enterprises, as vendors will eventually embed AI in their products, it will be common to have these solutions. However, being competitive in the field will still require organisations to invest in their workforce’s AI skills – both in technical and business areas, since the real cost of AI is in the people. 

Who’s Impacted

• C-Suite
• CIO and CTO
• Cloud architects
• Low-code teams
• End user computing/digital workspace teams
• IT teams

What’s Next?

• Be ready for users to adapt AI to their day-to-day operations as the services become available in the tools they already use. 
• Advise senior executives of the coming changes, without falling into the current AI hype cycle.
• Be ready with your enterprise’s AI governance framework that integrates AI and ML projects and IT operations.

Related IBRS Advisory

1.DIY or ready-made? Choose your AI adoption path carefully

2. Five Things to Consider When Evaluating AI… and Five Dangerous AI Misconceptions