Editor’s note: this is a guest post from students in the Master of Human-Computer Interaction (MHCI) program at Carnegie Mellon University, who have been working with CR to imagine and design AI agents that help solve common consumer problems. Read on to learn about their design process and check out other installments in the series.
Our previous blog post introduced our project and a long-standing problem for consumers: the lack of information and power when dealing with businesses. It discussed how information asymmetry has historically favored businesses, and how organizations like Consumer Reports (CR) have worked to empower consumers. It also introduced some challenges that consumers face in navigating customer service in the digital age and how Agentic AI could be a potential solution.
This week, our student team at the Human-Computer Interaction Institute at Carnegie Mellon University explores the ways in which Agentic AI could be of service to consumers. Why exactly do consumers need to be empowered in the post-purchase stage? What pain points do they currently face in the marketplace? And what do they think of AI agents? The following blog post presents the research we conducted that led to the development of our agentic AI solution. Read more to learn about what we heard from consumers!
Discovering Consumer Pain Points
Through a series of guerilla interviews talking to consumers, we made several discoveries about consumers’ current pain points when interacting with businesses.
Our research confirmed that consumers generally lack awareness of their rights outlined in privacy policies and terms of service agreements. This results in reactive behaviors, where individuals only seek this information when encountering product/service issues, hindering their ability to effectively advocate for themselves. This got us thinking: how can we proactively equip consumers with the knowledge of their rights before problems arise?
Furthermore, the technical and opaque language used in these agreements emerged as a substantial barrier. This complexity contributed to feelings of helplessness and apathy among consumers, discouraging them from engaging with these documents. This presents another opportunity: how can we present terms of service agreements in a clear, concise, and user-friendly manner?
Opportunity in the Post-Purchase Stage
Our conversations with consumers revealed that consumers’ greatest frustrations arose during the post-purchase stage. Customer service interactions specifically appeared to be the most frustrating aspect of their journey. In many cases, frustration can arise when consumers lack a knowledge of their rights and the power to hold businesses accountable. And businesses may lack incentives to uphold their end of the bargain and assist consumers after a purchase. This presents a significant opportunity for CR to step in and empower consumers during this critical phase.
Our research underscores the value of empowering consumers with relevant, actionable information precisely at the moment of need. Equipping them with the knowledge and tools to navigate challenging interactions with businesses would significantly level the playing field, fostering a more fair and transparent marketplace. Beyond timely information provision and guidance, we also explored the possibility of an even more proactive approach – acting on behalf of consumers. We identified an opportunity for CR to act on consumers’ behalf in their post-purchase interactions with businesses via human or AI agents. Given the recent advancements in agentic AI technology and its scalability that allows it to simultaneously help millions of consumers 24/7, we decided to further explore employing this emerging technology in our project.
How Might Agentic AI Help?
In order to better understand how agentic AI technology can be employed in the post-purchase phases to empower consumers, we created storyboards that depict potential future post-purchase experiences mediated by agentic AI across various domains (ISPs, airlines, car insurance and loans).
We presented these storyboards to consumers to assess their reaction to these futuristic interventions in order to gain insights on consumer pain points, motivations, and comfort levels with this type of technology. From these interviews, we found that:
#1: Companies’ poor customer experience design necessitates an agent capable of identifying contact channels, curating case information, and tracking issue resolution progress on behalf of consumers.
There is inherent inefficiency within many companies’ customer service processes. This inefficiency creates significant obstacles for consumers seeking to resolve issues. Our consumer interviews pinpointed specific points of interaction that are particularly troublesome. These include the time-consuming and laborious task of locating contact information (e.g., email addresses, phone numbers) for companies. Furthermore, participants expressed frustration with the burden of gathering individual case information (such as personal details, order numbers, and purchase dates) from disparate sources when filing service requests or complaints. Finally, a common theme emerged regarding the lack of transparency within service request processes. Many participants described feeling lost within a convoluted system, with little to no progress updates.
To address these shortcomings, we propose the implementation of an AI agent capable of identifying contact channels, curating case information, and tracking issue resolution progress. This agentic AI system would streamline the customer service experience, eliminating the time-consuming and confusing aspects that currently plague consumers.
“Contacting customer service feels like a chore. It’s anxiety-inducing. And in general talking to strangers is scary. It’s frustrating because a lot of them aren’t helpful.” [Female, 20s]
“I was looking for the email to contact customer service but that took me forever.” [Female, 20s]
“…You are prompted for ticket number, confirmation number, reservation number, flight number, all of those things live in different places… you’re like going in and out of apps…it’s torment. “ [Female, 40s]
“I would experience anxiety due to the unknown — in between processes I would like to have updates for the entire process, each step…” [Male, 20s]
#2: Consumers desire assistance in constructing arguments to effectively advocate for their rights when interacting with businesses.
Company policies are often perceived as “dense” and “overly technical,” hindering consumer comprehension. This lack of understanding fosters feelings of lack of confidence, unwillingness to confront, apathy and a sense of powerlessness. Consequently, consumers express a desire for an advocate, likened to “a little Karen” by one participant, who is assertive, logical and can fight for their rights. With recent advances in conversational and voice AI technology, this is possible using agentic AI. An AI agent can play the role of this “advocate” that consumers desire, sorting through lengthy, complex company policies and drafting personalized negotiation arguments that consumers can use to get what they deserve in their customer service interactions.
“Preparing for the call, figuring out what I need. Why did this happen? Do I have a case? It takes time…” [Female, 20s]
“I want a ‘little Karen’ in that situation that comes with a certain level of comfort and confrontation [because] I think that empowerment as a consumer sometimes gets lost, whether it’s because of a lack of confidence or comfort.” [Male, 20s]
#3. The likelihood of a consumer adopting an agent to represent them depends on their confidence in the agent’s ability to advocate as persistently and effectively as they would for themselves or better.
Our research found that consumers are hesitant to use agentic AI unless they believe it can be as persistent and effective as they would be themselves, with some doubting the agent’s ability to advocate, especially in situations requiring human emotions like anger or forceful persuasion. This illustrates consumers’ desire for the agent to vigorously advocate on their behalf until the desired outcome is achieved and highlights their current lack of trust in AI agents to do so. This finding shows that we need to be cognizant that the agent we design is competing with the user themself, underscoring the importance of designing an AI agent that competes effectively with human capabilities, reassuring users of its competence and effectiveness.
“I wish I could argue for my money back. I don’t think an agent can persuade a person as effectively as I can… through anger, niceness, and persistence… you can be forceful in a situation, a robot can’t.” [Female, 20s]
#4: It is essential to cultivate consumer trust in AI agents
Building trust in agentic AI requires addressing several key hurdles identified in our research. We found that consumers rely on recommendations from trusted sources before feeling comfortable using the AI agent. Once they do interact, user control is important – while some consumers appreciate proactive agent outreach for low-risk scenarios, others prioritize control and expect to initiate contact with the AI agent themselves. Regardless of who initiates, clear confirmation that the AI agent understands their request is essential. Additionally, an appropriate level of personability from the AI agent, avoiding overly human-like behavior, is desired. Finally, fostering trust requires the AI agent to provide transparent and consistent responses throughout their interaction. Consumers need to know the source of the information being provided and receive regular updates on the progress of their service requests. This level of transparency builds confidence in the agent’s effectiveness and lays the foundation for a trusting relationship.
“The trust builds each time the agent conveys to me that it understands my circumstances. And so the more granularly you can do that, the more I will trust and engage because each time that happens, it’s like, oh my God, I feel seen. It totally understands my circumstances.” [Female, 40s]
Our research shed light on the significant potential of agentic AI to empower consumers in the post-purchase phase. By addressing the pain points of navigating convoluted customer service processes, deciphering complex company policies, and advocating for their rights, agentic AI can transform the consumer experience. However, building trust is paramount. To achieve this, agentic AI systems must be designed with transparency, user control, and an appropriate level of personability in mind. By addressing these concerns and building upon user trust, agentic AI can usher in a new era of empowered consumers and efficient customer service interactions.
Next time…
In our next post, we’ll explore the initial designs and concepts that inspired our final agentic AI solution.
