Personalization to support customer engagement and boost revenue
In addition, customers will share additional personal information if they get something back for doing so. “I will rate a restaurant because I want the app to know what I like and to make a good recommendation,” he opined. “If a customer uses WhatsApp to contact an organization, they will be willing to hear back on that channel as long as they get value.” The more interactions a customer has and the more datapoints that are available, the more personalization of offers and recommendations is possible.
Large language models (LLMs) are often used to train AI applications, but Cooper advised caution. If the training is on data that is universally available, the responses will not be tailored to the organization creating the AI. “Any AI is only as good as the data it’s trained on,” Cooper noted. “Enlighten has billions of calls to rely on. We redact personal information but have data that is labeled, such as ratings that someone gave a product, or whether a call was a complaint. This allows us to predict a likely NPS score and can guide the response from the agent using the AI-based recommendation.”
Such information also allows personalization by matching certain agents to the caller. “After just a few minutes of audio or text, we can map that individual as one of seven personality types,” Cooper continued. “We know how certain agents perform against those personality types and can route the call to the right agent.”
Several other modules support other customer engagement functions. Enlighten’s Customer Satisfaction module mea- sures and guides agents on the soft-skill behaviors that improve customer sentiment, such as demonstrating ownership, active listening, and building rapport. Enlighten Autopilot provides guidance in the self-service environment, and Enlighten Copilot guides what agents should offer. “If a customer says they dislike a certain product, 90% of the time, the agent will take the recommendation of Enlighten Copilot,” Cooper said. “And if the recommendation is for the customer to go to Enlighten Autopilot, then they can be steered to that solution and don’t need the agent for that interaction.”
Tailoring intranet communications
Although ecommerce and contact centers account for much of the current focus on personalization, other environments have been leveraging it as well. Search has been a critical component in KM for decades, and now personalized search is standard in online and enterprise search. Advances in personalization have also emerged in modern intranet solutions, which are now configured to deliver personalized content to employees in either push or pull mode.
“Organizations have been moving away from the ‘content dump’ in their homegrown intranets for the past 8 or 9 years because they knew that it was not working,” said Cheryl McKinnon, principal analyst at Forrester. Intranets built in-house often had a poor user interface, and the content lacked adequate metadata.
Now organizations are taking a closer look at purpose-built, state-of-the-art intranet packages. “The new solutions are mostly cloud native and have built-in analytics,” McKinnon stated. “In addition, they are managed by internal communication leaders who are responsible for employee engagement, rather than by IT, so there is more input from line-of-business managers and subject matter experts.”
According to McKinnon, a long-standing intranet issue is whether people are actually using the content. “One approach to maximizing engagement is to orchestrate content delivery so that people only receive content that is relevant to them. If people start to trust that corporate messaging is personalized and therefore relevant to their work, they will be more engaged and more willing to read the material than if they perceive it as junk mail.”