Generative AI and Microsoft’s Strategic Competitive Advantage

Original Post on Medium: Generative AI and Microsoft’s Strategic Competitive Advantage | by Sam Bobo | Speaking Artificially | May, 2023 | Medium

Disclaimer: I am actively employed by Microsoft but am independently writing this blog post without association to Microsoft marketing or similar entitles

Since November 2022, an explosion of new Artificial Intelligence start-ups and products have hit the market, sparked by the Generative AI capabilities “debuted”?(debuted as viewed by general populus but not the technological capabilities itself which long preceded it)?by OpenAI’s , in an effort to ride the hype cycle wave, set standards, and achieve first mover advantages within a particular domain or solution set.

What resulted in aggregate were two dominating top-level use cases of Generative AI:

1. Drafting Assistance?— Generative AI capabilities, given its underlying, ever-growing multi-billion parameter large language models, span text-generating and image-generating center around text-based transformation capabilities such as text summarization, template creation, auto-completion, text-to-image or image-to-text, tone-rewriting, and much more!
2. Question Answering?— based in part by text summarization capabilities, question answering inputs a text or image query and seeks to provide summarized and contextual answers on the matter.

Framing the AI Applicability Problem in Solutions — 4 Steps to Determine AI’s Use within a Solution Set | by Sam Bobo | Speaking Artificially | May, 2023 | Medium

Most notably, the latter use case, question answering, is predominantly positioned with large search engine providers such as Google and Microsoft (via Bing) as well as chat-bot based capabilities whom seek to answer long-tail questions outside of the programmable confines of the bot itself. This space is riddled with AI-based hallucinations plaguing its acceptance at the moment, whereby the system is inherently trained on the former use case to prevail.

The focal point of this piece shall be on Drafting Assistance and why Microsoft is poised as an organization to strategically capitalize on its capabilities.

At the top most layer, Microsoft’s Mission Statement reads:

“To empower every person and every organization to achieve more”

Productivity is innate to Microsoft’s existence and drives decision making, commercialization, and product direction from the top down, imbued within the organization’s culture, and focal point of the influence it seeks to have on the world.

Let’s start with a summary of Microsoft's history, first, in part, by the rise of personal computers and the inheritance of the DOS framework from IBM back in the 1980s followed by the debut of Microsoft Windows in 1985. Moving into the 1990s, Microsoft unveiled its flagship Word, Excel, and PowerPoint applications bundled with the Windows Operating System and started penetrating the consumer market via the enterprise market with its capabilities. Since then, Microsoft has continued to evolve its office products, releasing ever-complex capabilities to attach to both every-day users and power-users alike. Over time, Office transformed into a suite of productivity tools, including Outlook (email), Access (Databases), Publisher (content-based layouts), Visio (charting), Teams (Messaging), OneNote (note-taking), Project (Product Management), OneDrive (document storage), etc and set the standard for document extensions globally. Today, office has nearly 1.2B users spanning personal and enterprise using its suite of productivity applications.

Delving more into Office 365, imaging the type of work handled by the applications Microsoft provides and the jobs to be done thereof. To name a few — internal communications, document creation (resumes, mortgage agreements, executive summaries, books, medical summaries, and more). I’d simply continue in near-perpetuity describing the basic jobs that office-related productivity applications service to users spanning all industry verticals and disciplines. Documents serve as artifacts of work, archives of record keeping, and facilitators of communication amongst parties, whether document-based, images, audio, and more! Simply put, work progresses society and is vital to the fabric of progress, not having AI integrated would be a mistake.

The Three Eras of Conversational AI — a framework for explanation | by Sam Bobo | Speaking Artificially | Medium

Circling back to Generative AI capabilities, specifically Drafting Assistance, its blatantly obvious why Microsoft would attach to the concept of Generative AI as it fits perfectly within its mission statement and downstream offering suite. Hence, when Generative AI’s popularity accelerated, Microsoft capitalized on the opportunity:

Introducing the Microsoft 365 Copilot Early Access Program and new capabilities in Copilot | Microsoft 365 Blog

What followed over the timespan of 4months was the announcement of 12+ Co-pilot integrations whereby Generative AI, powered by the long-standing partnership with OpenAI, powered productivity-based tasks to reduce the burden of mundane work, spark the creativity and iterative improvement cycle within knowledge workers, and shift the focus of work to more innovative thought by the common worker. Simply put, Microsoft commoditized the lowest level of work and sought to raise the general productivity of society. Take a look at the image below to see the series of announcements nearly back-to-back by Microsoft:

Why am I so bullish on Microsoft? Take a look at the large competitors in the space, namely the large hyperscalers clouds and noteworthy large tech companies:

• Google?— the business model of Google continues to be under attack by Generative AI, although, combatted intelligently by the recent announcement of the new reimagined Bard-based search experience announced recently at Google I/O. Google, while captures a segment of the productivity market with its Google Workspace business, focusing on the free-aspect of the market and locking in with gmail as a primary driver within the personal space, continues to rely on search as the largest part of the business and thus, the “Question Answering” dimension of Generative AI.
• Amazon?— Amazon, with AWS as the catalyst of componentized and modular SaaS services, is poised to offer generative AI but the primary use case focused on by Amazon is its storefront of Amazon.com, not a primary use case for Generative AI.
• Facebook?— Facebook is interesting as Generative AI provides a near-countless amount of generated content to be added onto the platform and litter its news feed (notwithstanding all of the controversy behind AI Generated content transparency, hallucinations and falsehoods with confidence, and more.) Facebook as taken an Open Source approach to Generative AI to promote its core business, less on owning proprietary models and hance the public release of models such as LLaMa.

Microsoft made the highly strategic move to partner with OpenAI and provide Cloud credits for hosting its models on Microsoft Azure and amortizing its fixed datacenter costs. Additionally, this partnership gave Microsoft early access to OpenAI capabilities which it leveraged into the aforementioned releases and will continue to benefit from ongoing developments by the organization. Yes, could Open Source outperform OpenAI? Yes, certainly. The main focus should be, however, that all users will eventually become accustom to drafting first with AI and applying subject matter expertise on top to compete the job. Moreover, the applicability of subject-matter-specific AI models could be a logical next step for Microsoft. In the interim, they are focusing on Semantic Index as a way of accessing company information to provide additional context to prompts.

I, personally, am lucky to be joining Microsoft at such as pivotal point in its AI journey and existence and look forward to helping to evangelize AI capabilities for the populus en masse to understand!