Designing for Trust DagliMeric 2019 <div>Trust can be thought of as one of the currencies that humans</div><div>use to accept a technology in their everyday lives. One of those</div><div>technologies are conversational agents (also known as chatbots)</div><div>like digital assistants in our smartphones. To provide a more</div><div>personal experience, increasing number of businesses are</div><div>developing virtual agents with conversational interfaces by</div><div>personifying their products and services through automation.</div><div>While some agents are generalists like voice-based personal</div><div>assistants, many of them are specialist agents that are designed for</div><div>specific tasks in different domains including e-commerce. As user</div><div>expectations get more complex each day, a collaboration between</div><div>specialist agents is needed. In such a scenario, users’ trust level</div><div>may change dynamically due to the agent hand-offs. Additionally,</div><div>despite many offered benefits, many people hesitate to trust</div><div>conversational agents with complex tasks in certain e-commerce</div><div>scenarios such as travel booking.</div><div><br></div><div>This design research project explores trust issues with virtual</div><div>e-commerce agents in several collaboration scenarios between</div><div>generalist and specialist agents and provides design guidelines for</div><div>more trustworthy conversational agents.</div>