barriers (no, it isn't three)
Vendors pushing artificial intelligence solutions may be overpromising and underdelivering PHOTO: The Tire Zoo

Gartner issued a warning yesterday to any vendor busy marketing their artificial intelligence (AI) capabilities: your hard sell could be counterproductive. 

According to Jim Hare, research vice president at Stamford, Conn.-based Gartner, the push to build and market AI products has been so intense that many vendors have forgotten to do a basic analysis of enterprise needs and use-case scenarios.

AI Moves Up the Hype Cycle

"As AI accelerates up the Hype Cycle, many software providers are looking to stake their claim in the biggest gold rush in recent years," Hare said in a statement.

Jim Hare
Jim Hare
"AI offers exciting possibilities, but unfortunately, most vendors are focused on the goal of simply building and marketing an AI-based product rather than first identifying needs, potential uses and the business value to customers."

Hare was speaking ahead of the Gartner Application Architecture, Development and Integration Summit 2017 taking place later this month in Sydney.

The speed at which AI has captured the enterprise imagination cannot be overestimated. In January 2016, for example, AI didn’t even figure in the top 100 search terms at Gartner. By May of this year, it had risen to seventh position.

Big Vendors’ AI Plays

Microsoft: 'Help People Do Their Jobs Better'

But it's early days yet. Only last week, Microsoft announced the development of a new research lab at its headquarters in Redmond, Washington. Harry Shum, executive vice president of Microsoft’s AI and Research Group, outlined Microsoft’s AI ambitions at an event in London. The hub will focus, he explained, on working with departments across the company to support integration of AI advances into products and services.

Microsoft is developing AI technologies to create tools “that help people do their jobs better,” he said. However, he added a caveat:

“The people who use those tools should be able to understand how they work and what data they rely on. AI can be more useful if the people who created and use the tools can explain how they work and why decisions are made.”

AI, according to Gartner, refers to systems that can change their behaviors as they learn from data collected from apps, people or analytics applied to market places.

Google: The 'Human Side' of AI

Mountain View, Calif.-based Google is concerned AI is leaving workers behind. On July 10, it announced the creation of a new initiative aimed at making AI accessible to workers called PAIR (People + AI Research).

PAIR brings together researchers across Google to study and redesign how people interact with AI systems. In a blog post about the release Martin Wattenberg and Fernanda Viégas, senior researchers on the Google Brain team explained:

"The goal of PAIR is to focus on the 'human side' of AI: the relationship between users and technology, the new applications it enables, and how to make it broadly inclusive. The goal isn’t just to publish research; we’re also releasing open source tools for researchers and other experts to use.”

The objective is to put people at the heart of AI development:

“We believe AI can go much further — and be more useful to all of us — if we build systems with people in mind at the start of the process,” they wrote.

IBM: Baking Watson Into Its Software

For its part, Armonk-NY based IBM announced the restructuring of its services division to put Watson and its AI capacities at the heart of its software offerings. The new AI-capabilities will help enterprises predict problems in the infrastructure offering before they even happen.

This follows wide-ranging job cuts at IBM's Global Technology Services division last year, which the company said was the first step in a strategy shift that would redirect IBM towards cloud computing and AI operations, in much the same way Microsoft did a couple of weeks ago.

These are only three of the bigger vendors. There are others plus a whole range of new, smaller entrants into the game which could result in major market disruption by 2020.

3 AI Problems

The result of the manic drive towards AI, according to Gartner’s Hare, is the emergence of three problems that vendors need to overcome.

"AI offers exciting possibilities, but unfortunately, most vendors are focused on the goal of simply building and marketing an AI-based product rather than first identifying needs, potential uses and the business value to customers," Hare said.

1. Lack of differentiation

With a growing number of start-ups entering the field and larger vendors pushing their products, enterprises are finding it difficult to distinguish between one product and another.

There are, according to Hare, more than 1000 vendors on the market all of whom claim to be offering AI applications or platforms. However, many vendors are pasting the AI label on their products indiscriminately.

To overcome this, vendors need to be clear as to what AI capabilities they are offering, what problem they solve — and be able to back these claims up.

2. Introducing unnecessary complexity

If complex machine-learning capabilities are grabbing the headlines, Hare points out that enterprises should be looking at AI products that have already proven themselves.

Complex ML tools could well be the proverbial hammer for the AI nut, but they also risks alienating workers. Hare argues vendors should provide apps that do what is required as simply as possible and avoid cutting-edge technologies unless they are needed and useful.

3. AI skill shortage

Skills shortages are nothing new and impact many different areas of technology. According to Gartner's 2017 AI development strategies survey, more than half of the enterprises survey cited this as a major problem for AI adoption.

The survey also showed organizations would prefer to buy solutions with AI already embedded rather than having to pull AI into existing solutions or platforms through customized solutions.

His parting advice? AI vendors should focus on providing solutions that tackle business problems rather that delivering cutting edge AI that cannot be used.