EkaLore has predicted increased conceptualization, development, implementation, and deployment of Artificial Intelligence (AI) based agent technology in customer inquiry, customer service delivery, status/expediting, order management, and pre-order interactions. The increase in technology deployment results from many coinciding factors, such as Covid-19, workforce demographics, labor costs, hybrid/remote work models, technology maturity, and user interface sophistication. Of particular note is the growth of user expectations accepting AI-based agents.
In this release, EkaLore briefly looks at some success factors enabling broad user acceptance of AI-based agents.
Success with “Smart Speakers” from Amazon and Google; compute-device based functions on Android, iOS, macOS, and Windows; and, generally, “hands-free” implementations (such as in vehicles) have had uneven success. Amazon and Google have publicly reallocated some of their “smart device” resources. Microsoft has seen little acceptance of Cortana on the desktop. Vehicle “hands-free” implementations have had uneven acceptance. Various anecdotal and reviewer comments have focused on the difficulties of converting user interactions to valuable interactions (monetized or data producing), growth in functional uses from unaware users, and highly-tested –and ultimately clumsy – implementations. In short, the experience with devices and voice-driven functions has been convenient for many users while financially unrewarding for technology developers.
Narrowly defined smart device interactions drive satisfaction. Current users identify specific situations that prove gratifying and are satisfied with those specific interaction patterns. Accepted interactions driven by specific situations provide user interface experience improvements (“turn off the lights”, “look for a family restaurant”, “how long to drive to The Mall”).
Greater sophistication of AI applications results in simplicity for the end user. Instead of having to navigate a menu-tree-based inquiry line Natural Language Processing/Understanding applications can shortcut to specific use cases and improve user experience (cell phone support and payment applications are one example out of customer-inquiry-bill-pay applications in general). The tens of millions of devices and billions of smartphones are creating user competency and critical acceptance of AI-agent deployment.
The weakness of current research is also a strength for innovative and capable deployments. Recent research https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4225917, “Bots with Feelings: Should AI Agents Express Positive Emotion in Customer Service?”, also Han, E., Yin, D., and Zhang, H.; “Bots with Feelings: Should AI Agents Express Positive Emotion in Customer Service?”, Information Systems Research, 2022, DOI: 10.1287/isre.2022.1179) was reviewed as expressing, “The results across the studies show that using positive emotion in chatbots is challenging because businesses don’t know a customer’s biases and expectations going into the interaction. A happy chatbot could lead to an unhappy customer.” (Science Daily, Releases, 2022-12, 221221185751) Simply asking the user what makes them comfortable for user interaction is as acceptable as a human-agent asking how someone would like to interact with the servicing agent in a high-touch luxury environment and context.
A look at the broader contexts of dialogue-driven applications and user interface design sees multiple paths for successful deployments. Tedious or over-structured interactions (such as interactions using type-wheels or hierarchical menu systems) are disliked compared to free-form voice interactions with confirmation and notification styles set to user preferences. Cultural, experience, or sophistication of user interactions can all be accommodated by applications of sufficient sophistication and flexibility. Clearly the day of “programmer ego dictates” (with limitations by culture and emotional intelligence) is coming to a close with flexibility driven by users the choice. The excuses of hard-to-program, limited compute resources, and programmer simplicity are affronts to too many likely users. Enterprises looking to meet their objectives will see a front-end loaded cost to a successful deployment with long term rewards.
This post makes the case that there are compelling reasons to for enterprises to use AI Agents for customer service. We will expand on this idea in future posts.
If you’d like to read more posts like this, you can find them at www.ekalore.com/alien-invaders
Comentários