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 illustrated several consumer pain points– difficulty in identifying company contact channels, curating case information, tracking issue resolution progress, and desire of assistance in constructing arguments to effectively advocate for their rights when interacting with businesses. This week our student team at the Human-Computer Interaction Institute at Carnegie Mellon University will tell you about our guiding framework informed by these pain points, before introducing three prototypes designed to resolve these problems. Finally, we will share how these concepts come together as one product to be validated and tested in the following weeks.
Guiding Design with Research-Inspired Framework
Findings and pain points of consumers’ daily experience from the research stage were formulated into this statement:
How might we streamline and empower consumers’ post-purchase interactions with businesses to reduce cognitive load, emotional strain, and time investment to ensure transparency and fairness in the marketplace?
Transparency and fairness became the primary focus since we aim to bridge the information asymmetry between consumers and businesses and ameliorate the power imbalance. We hope to address cognitive overload of facing overly complex terms and conditions, alleviate stress, apathy and fear that prevent consumers from advocating for their rights, and finally– streamline customer service interactions to be less time consuming so as to encourage and empower consumers to argue for what they deserve. To consolidate, we established the “IGA” framework with three guiding principles which our design would follow:
Inform. The agent will “inform” consumers about relevant information at appropriate timing. Curating information in a digestible, concise, and timely manner, the agent can reduce information overload consumers experience due to the dense and technical company policies.
Guide. Through step-by-step guidance and resources, the agent will “guide” consumers to help them navigate the complex process of interacting with companies in the post-purchase phase. Such a form of guidance may alleviate the emotional overwhelm associated with customer service.
Act. Last but not least, the agent can take it one step further, and act on behalf of consumers. In doing so, we remove all the labor from them, reducing the time and effort invested to advocate for consumer rights.
Following the IGA framework, we proposed three design concepts of an agent that would inform, guide, and/or act on behalf of consumers.
Concept #1 – Negotiation Helper
Inspired by one consumer’s story, we imagined a fairy godmother that sits in and provides tips during a call with a customer service representative. Negotiation Helper would do exactly that– except that instead of a fairy godmother, it’s a “smart agent” leveraging advanced technologies like Large Language Models (LLM). Negotiation Helper would enable real-time alerts based on feedback from customer service representatives and provide post-call tips to strengthen argument and advocacy. It would also centralize businesses’ contact channels and call history for easy access.
This idea, while not without risks, would be justified and perhaps overdue in the sense that customer service calls are often recorded by companies without explicitly obtaining consent. Negotiation Helper would be recording and interpreting these calls under the assumption that consumers should have the right to record, especially if they are being recorded already.
Tying back to the IGA framework, Negotiation Helper would “inform” with company policy alerts, “guide” users from providing company contact before call, suggesting questions to ask during call, to showing tips and summary after call. But it would not “act”– at least not initially. Certain consumer archetypes may hold reservations to the idea of AI autonomy:
“I don’t think a smart agent is capable of advocating for me… you can be forceful in a situation, a robot can’t.” [Female, 20s]
We thus provide information and guidance as tools to empower consumers to argue better for themselves, without having to trust a robot to advocate on their behalf.
Concept #2 – Policy Assistant
Attempting to offset the harm of businesses’ deceptive practices, Policy Assistant– a smartphone widget in the control center– would alert users to dark patterns or beneficial company policies, sift through emails and browser content on demand, and present relevant policy segments clearly. It would also draft emails and formulate arguments based on these policies.
By providing user flexibility and control over when to engage the agent to capture their on-screen content, Policy Assistant would cater to people whose comfortability with AI may be lower while holding higher privacy concerns. The agent would only see what it is given permission to see, adhering to CR’s core value to always put consumers first.
When contacting company customer service, consumers are often offered something seemingly more rewarding than what they initially requested for, but in the long-term these offers might only benefit the company. Policy Assistant would leverage LLMs to “inform” consumer risks or benefits, before “guiding” users by drafting arguments to advocate for their desired outcomes.
Concept #3 – CR Wallet
Our third concept, CR Wallet, would be an app-based smart agent assisting consumers throughout their purchase journey. It would use Augmented Reality (AR) technology to identify product issues, guide users through resolutions, alert them to relevant company policies for repairs or replacements, and submit and track requests on their behalf.
True to its name, CR Wallet would integrate with CR Everywhere to track purchases, ensuring convenience and peace of mind. Incorporating AR with CR TV Screen Optimizer to identify issues, CR Wallet would “guide” users through troubleshooting, then– with permission– “act” and contact companies for replacements, refunds, or other requests on users’ behalf.
Moving Forward for Consumers, in Larger Marketplace
To create effective and delightful customer service experiences, we began merging our designs. The primary concept, Negotiation Helper, would offer real-time coaching and incorporate elements like explaining company policies and constructing case arguments from Policy Assistant, and documenting purchases from CR Wallet. Our new design further identified key transactions– billing issues, warranty claims, and subscription modifications as focal consumer pain points. But completing these requests requires handling personal identifiable information (PII); that’s why we conducted Fairness, Accountability, Transparency, and Ethics (FATE) analysis to mitigate privacy risks by ensuring secure PII storage.
Another concern that emerged from this analysis was the burden of multi-sensory stimuli on users. But what if…the paradigm is reversed? We proposed an innovative approach; the smart agent makes customer service calls on consumers’ behalf whereas they listen and supplement information as needed, while maintaining the option to take over the call. Excited yet cautious about this innovative concept, our team focused on careful execution to ensure convenience, user agency, and consumer advocacy while validating and iterating on the current version of Negotiation Helper.
Tune In Next Month
Stay tuned for the following posts in this series, as we begin concept validation and rapid prototyping to learn more about consumers’ attitudes and behaviors upon which our design continues to evolve.
