Applied AI Moves into the Spotlight at Build 2021

Microsoft makes artificial intelligence industry-specific

Among the 100 updates announced at Microsoft’s developer event Build last week, an area that stood out above the noise was artificial intelligence (AI) and in particular, Microsoft’s growing push into higher-level services for applied AI and business scenarios.

Redmond has been gradually embedding more of its AI with business logic in an attempt to help enterprises get out of the “pilot purgatory” that has so often characterized AI projects over the past few years. But at Build this year, it took a big leap forward.

Let’s take a closer look at this in the context of Build’s major moves and what they mean for the market.

The Need for More Business-Centric AI

According to our data, more than 80% of companies are now trialling or putting AI into production in their organizations, up from 55% in 2019, and we’ve seen adoption accelerate dramatically in several narrow areas such as contact centres, chat bots and fraud detection.

However, fewer than 20% deploying AI have fully put it into company-wide processes and at most, only one in five AI solutions currently becomes operational. In effect, enterprises can’t scale AI, held back by many challenges including the time it takes for it to bear fruit.

Over the past few years, cloud providers have celebrated advancements in what I call general-purpose AI, areas such as speech, language or image classification, for example. These have no doubt propelled the technology into the limelight, but the progress has been in many ways rather meaningless to the average company, which is often saddled with the complexity of customizing the technology for their particular business purpose.

Microsoft Responds

This is why there’s a growing need for more business-centric AI, which clusters general-purpose AI to solve specific business problems. Along with the aim of speeding up business scenarios for developers, this was the prime motive for some of Microsoft’s major AI updates at Build 2021.

There was a wave of announcements focussed on Cosmos DB, Azure Machine Learning and Power Platform, but an area that stood front and centre in applied AI was Cognitive Services, Microsoft’s suite of machine learning algorithms and API services that help developers embed AI into their apps.

Microsoft released into general availability Azure Metrics Advisor, which was announced at Ignite in March 2021. Metrics Advisor ingests time-series data and uses machine learning to proactively monitor metrics to detect anomalies and diagnose issues in business operations in sales or manufacturing processes. It also released Azure Video Analyzer, which brings Live Video Analytics and Video Indexer into a single service to help developers quickly build AI-powered video analytics from both stored and streaming videos. According to Microsoft, the new service will enable business uses in workplace safety, in-store retail or digital asset management.

The new capabilities also complement its Spatial Analysis AI service, launched in 2020. This aggregates information from multiple cameras to assess how many people are in a room and how close together they are to help with social distancing measures.

The Missing Piece of the Puzzle

Along with its Cognitive Search, Form Recognizer and Immersive Reader AI services, Microsoft is clearly positioning its AI as a set of turnkey services that offer more built-in business logic and address common business scenarios such as document processing, customer service and workplace safety.

This emphasis has been a missing piece in Microsoft’s AI puzzle, which in the past has largely focussed on platform and horizontal AI technologies for developers. Together with its recent Nuance acquisition, the moves form a pattern of concentrating more on business applications for AI and helping developers with limited machine learning experience get over the hurdles of AI and with more task-specific solutions.

A Warning Shot across Rivals’ Bows

What also stands out is that these moves are a strong competitive attack on Amazon Web Services (AWS) and Google Cloud, which have been similarly investing in the area for several years now.

After re:Invent in December 2020, I highlighted the steps AWS was taking to expand up the stack into higher-level services and solutions for businesses and vertical markets, including in the fields of business operations, business intelligence and contact centres. AWS’ headline launches over the past 12 months have targeted precisely this area, most notably solutions for industrial sectors aimed at improving assembly line production, quality management and remote operations in factories and warehouses.

Similarly, Google Cloud has released several business solutions in its AI portfolio over the past few years, targeted at contact centres, document understanding, demand forecasting and product recommendations and search, among others. These have become critical spearheads in a committed vertical strategy that has emerged under CEO Thomas Kurian.

Accelerating along the Path to Business Value

For machine learning to reach its potential in the enterprise market, it needs to be far more pervasive among businesses and users who have little to no expertise with the technology. It’s this gap in applied and business-centric AI that growing competition from Microsoft, AWS and Google Cloud is starting to bridge.

One of the hallmarks of the pandemic is that AI is now no longer viewed as an experimental, longer-term source of innovation for companies; rather, it’s a technology that can deliver quick transformational and business value. But companies can no longer afford to have investments tied up in longer-term AI projects and proofs of concept that yield limited business value as many did in 2019.

Microsoft is tapping into this trend and facing the competition head on as it continues to answer the challenges of the developer community with its portfolio of AI developer tools.

It will be fascinating to see how the market responds to the announcements from Build 2021 in the coming months.

A version of this article was first published by VentureBeat on 27 May 2021.