The development of artificial intelligence and natural language processing has the potential to transform the way people interact with technology. SAP's Conversational AI approach encompasses most or all aspects of the machine-to-human language interface, but it also raises questions about the practicality of a voice-enabled world.
Dialogue, not dialog boxes
SAP's recent work in developing a more natural way to interact with computing devices is built upon its Leonardo AI and IoT services, the recent acquisition of Recast.AI, and its CoPilot NLP user interface technology. The intent behind development is to deploy NLP throughout SAP and its product line, encompassing 12 lines of business and 25 industries and allowing users to speak and chat with enterprise applications. Integration would reach deep into infrastructure, including access control and data security, up through enterprise systems like HANA, Hybris, and SuccessFactors, to end-user apps and social media.
SAP and Recast.AI have several case studies that showcase their successes with Conversational AI, including banking IT, auto sales, telco customer account management, and pharmaceuticals procurement. The applicability of AI and NLP technology, whether SAP's or that of its competitors, is not much in question. What is not often considered in the discussion of new user experiences is the potential social impact.
Changing the conversation
Lessons from the past show us that new technologies lead to new complications. The automobile led to sweeping changes in urban planning, industrial methods, and social norms; more recently, mobile communications technology has caused nothing short of a revolution in business, media, and (again) social norms. Neither case, nor any of the many others throughout history, has been entirely positive in its effects. The common complaints that people spend more time staring at their phones than in face-to-face conversation, that spelling and grammar have collapsed under the weight of abbreviations and emojis, or that conversations that were once private and/or quiet are now conducted loudly in public with no regard to people nearby are sometimes exaggerated but nonetheless real. IVR is an example of something whose intended convenience was somewhat overshadowed by the annoying realities of using it.
Ovum sees the rise of speech-capable AI apps as potentially another iteration of the problem with phones and IVRs – will the utility exceed the noise factor? In this case, noise factor is meant literally as well as figuratively; talking to a computer is fine in isolation, but it quickly leads to chaos in more crowded environments. Mobile phones have shown that we cannot count on individual discretion when it comes to private conversations in public spaces. The combination of annoyance and the possibility of exposing sensitive information must be considered in any implementation.
On the other hand, followers of this discussion should not be quick to alarm either. There is nothing inherent to SAP's technology, or any other, that is an especial threat to productivity or social order. All NLP systems can understand text as well as speech – possibly even better – even if the option is not touted. SAP should carefully consider the appropriate use cases for voice enablement and let that guide their deployments and marketing, with a thought to privacy as well as etiquette. Ovum recommends that developers and those considering deploying an NLP-enabled interface emphasize the use of keyboards over microphones in public places.
Market Radar: AI-Assisted Chatbots for Customer Service, IT0020-000287 (June 2017)
Building Predictive Capability and Use into Self-service Analytics, INT002-000077 (February 2018)
IoT and UC: Points of Intersection and Convergence, ENS001-000013 (January 2018)
MWC 2018: What to Expect, GLB007-000032 (February 2018)
Marshall Lager, Senior Analyst, Customer Engagement