Approaching successful digital journeys with AI: Questions to consider

As customer and employee expectations continue to demand rapid access to information, enterprises must evolve to accommodate these desires to maintain a competitive edge. Siloed information, however, remains a significant challenge for inviting successful digital experiences.

Joe Pairman, director of product management Tridion at RWS, and Rahel Anne Bailie, executive consultant at ICP, joined KMWorld’s webinar, Delivering Better Digital Journeys with AI, to explore how AI holds the key toward crafting digital experiences that build loyalty and engagement, increase productivity, and overcome information silos.

Pairman began by explaining that content delivery and content management tools have focused on providing a snapshot of information, not so much user journeys.

The focus on giving the “right” answer, despite the complexity of some industries, ignores the fact that the true right answer does not boil down to one snippet of text. Rather, a connected journey that travels through a variety of sources provides the most relevant information.

How should organizations start thinking about user journeys, then? According to Bailie, it begins by acknowledging the personas interacting with an organization’s touchpoints. Depending on the generation, technological skill, and other components that define a demographic, approaching each persona may require different strategies.

“If we know more about the persona, we get to know more about the journey as well,” Bailie explained. “When we map out the journey, we’re thinking about how a persona interacts with our digital properties.”

User journeys can vary infinitely, Pairman added. This is where AI can enable organizations to act on this infinity, where humans may fail to recognize all the journeys that a user can take.

Part of this infinity that is often ignored is disability access; when organizations fail to consider the user journey of, for example, a colorblind user, lawsuits and fines are incurred. Bailie suggests that instead of treating these journeys as “edge cases,” these journeys can be made possible and easy through AI.

With the advent of generative AI (genAI), however, a large question looms: Is metadata out-of-date?

Bailie stated plainly: Yes, you do still need it, and without it, you will net significantly poorer results.

“Let’s say 20% of your content sources have metadata on it. When you ask a question and your answer is in that 20%, you will get a better answer,” she continued. “Having said that, you can actually use AI to assign metadata, so there’s no reason not to use it.”

While asking “What is our AI strategy?” remains top-of-mind for many organizations, Bailie believes it may not be the right question to ask. “Like any other sector of business, you have to understand your business problem before throwing software at it. You have to consider what is the right kind of AI, as well as if AI is even appropriate for solving a particular business problem,” Bailie said.

In the realm of large language models (LLMs), Bailie warns viewers to not confuse LLMs with humans. When an LLM hallucinates, it is not “making things up,” as a human might if it hallucinates. Rather, the LLM is simply generating from existing sources, even if its answer is not “good enough.”

Therefore, genAI is not going to be apt for every problem; for medical and health industries, “good enough” is certainly not good enough, and could lead to extremely negative outcomes.

Despite this, organizations are looking toward AI and genAI to solve all their modern problems. Recognizing quality solutions, then, becomes an extremely necessary skill.

While many organizations will source external expertise to educate themselves on the matter, depending on the resources they have available to contribute to that effort, there are other ways to test for quality. Pairman suggested sandboxing any new technology, testing it out, and interacting with the people behind it to ensure you’re tackling the right problem, and tackling the right problem with the right solution.

For an in-depth discussion of AI strategies to guide successful digital experiences, you can view an archived version of the webinar here.