Inside Advanced Scale Challenges|Thursday, April 27, 2017
  • Subscribe to EnterpriseTech Weekly Updates: Subscribe by email

Forrester: Businesses Beginning to Jump on AI Bandwagon, but Barriers Abound 

Source: Forrester

The state of AI today reminds us of the old observation about teenagers and sex: there’s more curiosity than knowledge, more talk than action, more failed attempts than actual achievement. Also, it’s been around a long time, but each generation has to learn it anew.

As Forrester Research observes in its new report "Artificial Intelligence Technologies, Q1 2017," AI dates back to the 1950’s and has gone through several rebirths over the decades, notably around “expert systems” in the 1980s (at which time an office wag was heard to say, “artificial intelligence is better than none at all”). Today, another AI rebirth is taking place and it seems the availability of low-cost, increasingly powerful compute infrastructures (processors, data storage, networking) will bring it to fruition this time, putting impressive AI capabilities with the grasp of the enterprise.

Forrester’s copyright policies restricts us from quoting freely from its report, but senior analyst and lead Forrester author Brandon Purcell attempts to dispel misconceptions about AI as well as highlight its true potential and correct functional role.

Brandon Purcell, Forrester Research

One of his core points is that AI in the enterprise will not so much replace workers but “amplify human intelligence.” They key is to place AI in “hybrid” (people and machines working together) roles where it can excel by “providing contextual knowledge from data that the human mind alone can’t access and process,” such as “support combinatorial use cases like predictive lead scoring, combined with a training chatbot to make salespeople smarter; automated process guidance and knowledge search for customer support associates; and analysis of best practices and pitfalls of development decisions for software developers.”

One of the most effective roles for AI is taking on mundane tasks that require little human intelligence but consume much time, making work “onerous” and preventing workers from focusing on higher-level tasks. For example, AI can deliver rapid feedback based on analysis of large volumes of image and video content (such as surveillance video from surveillance cameras to detect shoplifting), as well as handling routine customer service needs using speech/text analytics and natural language processing.

Yet another role for AI: “robotic processes for self-healing and self-correcting systems” that operate without human involvement, including autonomous software agents that manage complex technology management infrastructures.

The design, knowledge engineering and model building required to put in place these workloads, no matter how mundane they may be, is highly complex. A good rule of thumb in selecting a use case, according to Forrester: the narrower the the better. Likewise, one of the primary barriers to successful adoption includes failure to develop a clear AI business case: many organizations don't understand how to apply AI… businesses are just beginning to jump on the bandwagon.”

And because there’s a dearth of AI ROI, it’s difficult to build a case for investing in AI technology.

Other challenges included scarcity of AI skill specialization, the need for a robust data management platform that can consume large volumes of training data from a variety of sources, and change management process concerns and the impact of AI on the existing organization.

In fact, Forrester found in a survey that AI has led16 percent of organizations to restructure their workforces and another 13 percent are in the planning stages.

To request a copy of the full report, go here.

Add a Comment

Share This