As the first research project I am leading with the Oregon Reality Lab, I decided to combine my passions for AI, UX, and public relations (PR) into a single study. Funded by the University of Oregon's School of Journalism and Communication and the Oregon Reality Lab, this research evaluates the potential of Google Gemini and Meta AI to provide optimal PR experiences for younger stakeholders. The short version of this paper will be published in December by the Institute for Public Relations (IPR) as part of the Ketchum PRISM Fellowship I received earlier.

H1. Younger stakeholders enjoy more optimal PR experiences using AI agents than websites, as they have higher perceived usability.
H2. Younger stakeholders enjoy more optimal PR experiences using AI agents than websites, as they are richer in media features.
H3. As the anthropomorphic features of AI agents increase media richness, positive brand judgment rises among younger stakeholders.
H4. As content personalization by AI agents increases media richness, so does information accessibility for younger stakeholders.
H5. The more open younger stakeholders are to generative AI, the closer the relationship they feel to an organization.

To answer these hypotheses, I utilized the three frameworks of computer are social actors (CASA), media richness theory (MRT), and Morville's Honeycomb model. To collect empirical data, I chose a mixed-methods approach using comparative usability tests. Participants first complete a pre-test survey that assesses their baseline attitudes, then complete a series of tasks using AI agents and corporate websites, and finally wrap things up with a post-test survey. The usability tests have been created using Lookback, and the surveys using Qualtrics. The data collection process has already begun, where 20 participants will spend an hour of their time contributing to this research. Participants have been instructed to use the Think-Aloud Protocol, as their audio, alongside video and screen, will be recorded.

Problem: Evaluating AI agents as automated PR agents and their potential for delivering optimal experiences.
Approach: Mixed-methods using comparative usability tests and surveys
Execution: Quantitative data will be collected through the two surveys and analyzed in R, while qualitative data, such as audio transcripts and observational notes, will be examined in NVivo.
Top findings: Coming soon
Impact: Coming soon
Reflection: Coming soon