The rationale for voice-based guidance
Students should not have to know their story before they can begin discovering it. Yet a college application, a major selection, or a job interview can ask them to present a clear account of who they are before they have had space to explore the answer.
ESAI’s approach begins with that exploration. Brandi, ESAI’s conversational AI guide, is designed to help people talk through their experiences, examine what matters to them, and connect that understanding to opportunities. The starting point can be a messy explanation, an unfinished thought, or an experience someone has never considered relevant.
Published research offers several reasons to take this approach seriously. Both speaking aloud and guided conversations can support self-awareness and career confidence. Additionally, emerging speech-based AI tools are showing how people capture loose thoughts first to organize and refine them afterward.
Together, these findings support a practical design principle: give people room to explore their experiences before asking them to present a polished story or commit to a pathway.
The problem with starting at the polished answer
A student asked to describe a strength may offer a familiar label: hardworking, organized, good with people. The experience behind that label can be more revealing. Perhaps they handled difficult customers, coordinated responsibilities at home, or helped a group recover from a failed project.
Consider a hypothetical student who begins: “I don’t know if this counts, but when my manager was away, everyone kept asking me what to do.” There is no polished leadership story yet. There is an experience worth exploring: What did people need? What decisions did the student make? Which parts felt natural or rewarding?
ESAI’s premise is that useful guidance should make room for those questions before asking for a finished narrative. The student can first describe what happened, then consider what it might reveal about their strengths, preferences, and possible next steps.
1. Speaking can change what makes it into the conversation
In Expression Modalities, Berger, Rocklage & Packard (2022) reported six laboratory and field studies in which writing produced less emotionally expressive attitudes than speaking. Greater deliberation helped explain the difference; encouraging speakers to deliberate reduced it.
The distinction is consequential: composing can affect what people choose to share, as well as how they phrase it. They raise a relevant hypothesis for ESAI: a spoken first response may bring forward details a student would otherwise dismiss as ordinary, irrelevant, or difficult to explain.
2. An imperfect first draft can become clearer afterward
Research on AI-assisted writing offers a close parallel to the “talk first, refine later” process.
Lin (2024) introduced Rambler, a tool that lets users dictate thoughts to review, reorganize, and revise material. In a comparative study, it better supported iterative revision and user control than other tools. A follow-up, Yang (2025), tracked participants who used both outline-first and free-speaking strategies. Spontaneous speech followed by restructuring was particularly evident in personal and reflective writing; participants described positive benefits for organizing thoughts and introspection.
These small studies support the feasibility and perceived value of separating expression from refinement. For ESAI, the opportunity is to let users begin with their own imperfect words, then help clarify their meaning while checking that the resulting account remains theirs.
3. Responsive conversation can support self-discovery
Voice alone is only part of the experience. The way a conversation responds to someone’s thoughts also matters.
Across five experiments, Itzchakov (2018) found that attentive, empathic, nonjudgmental listening increased speakers’ self-awareness and clarity about their attitudes. Related work by Shen & Sengupta (2018) found support for greater expression of self-related thoughts when speaking about liked brands.
These findings point toward the quality of engagement alongside the communication format. They inform Brandi’s intended role: invite examples, ask follow-up questions, and help users examine their own interpretations.
4. Guided exploration has a connection to career development
Career-specific research brings the discussion closer to the decisions students face. Gai (2022) studied career-focused motivational peer interviews delivered to first-year college students with low career adaptability. The intervention involved engaging, focusing, evoking, and planning, and produced significant improvements in career control and confidence supporting the value of structured career conversations.
The ESAI connection is from exploration to action: examine an experience, identify a possible strength or preference, and use that insight to investigate a course, major, internship, or career pathway. An insight becomes even more useful when the student can test it against real options and discuss it with an adviser, mentor, or employer.
5. Helping people see connections requires more than transcription
Capturing words is different from helping someone examine them. Li (2026) studied two AI voice tools with participants working through career uncertainty. The findings highlight how organization and follow-up questions support exploration, development of ideas, and clarification.
Brandi’s opportunity is to help users examine connections more deeply and more often across experiences.
6. Articulation includes being able to say it aloud
Self-understanding also needs to travel into real conversations. Schroeder & Epley (2015) found that evaluators, including professional recruiters, rated job candidates as more competent, thoughtful, and intelligent when hearing their pitches rather than reading them, and expressed greater hiring interest.
These findings highlight the importance of helping students communicate their experiences aloud.
ESAI’s approach: explore, understand, articulate, act
In ESAI’s application of the research, we use voice to help users explore and articulate their experiences while maintaining our commitment to keeping their story and judgment at the center of the process. The referenced studies inform a sequence for guidance in which discovery comes before presentation:
Explore
Share an experience without needing perfect wording
Examine
Describe actions, choices, and personal meaning
Connect
Consider themes across experiences and possible pathways
Refine
Build an accurate account the user recognizes as their own
Practice
Prepare to explain that account to another person
Act
Investigate options and take a concrete next step
Early results at Marymount University
ESAI’s initial pilot at Marymount University offers an early view of this approach in practice. Over 5–10 weeks, students used Brandi to explore their strengths, connect their experiences, and translate that understanding into career narratives. ESAI’s published pilot results report that 91% applied more efficiently, 90% felt more motivated in their job search, and 75% improved their interview skills. These encouraging early findings inform our next phase of evaluation: understanding how guided reflection translates into clearer articulation, more informed choices, and sustained career progress.
What the evidence supports and what remains to be tested
The evidence supports several parts of the rationale: communication format can shape expression; responsive human listening can support self-awareness; structured career conversations can improve aspects of career adaptability; and speech-based AI tools can help users organize and revise rough thoughts.
Building on this foundation, our next step is to evaluate how these benefits connect within the full experience of Brandi: speaking with AI → deeper self-knowledge → clearer articulation → more informed college and career choices. Through direct evaluation, we aim to understand whether and for whom this progression improves student outcomes, and how voice and text can each support that work.
Of course, more spontaneous language is not necessarily more truthful. A polished written response can be deeply personal, and an unfiltered spoken response can be incomplete or mistaken. Privacy, accessibility, language, surroundings, and comfort can affect how someone wants to engage.
At ESAI, authenticity means an account grounded in real experiences that the user recognizes, verifies, and can stand behind. Students deserve space to explore who they are before being asked to decide where they belong. Voice offers a promising way to begin that work: capture the experience in imperfect words, examine its meaning, and develop a clearer account of what comes next. Published research provides a foundation; direct evaluation will establish how far that promise translates into student outcomes.
References
- Berger, J., Rocklage, M. D., & Packard, G. (2022). Expression modalities: How speaking versus writing shapes word of mouth. Journal of Consumer Research, 49(3), 389–408. https://doi.org/10.1093/jcr/ucab076
- Berger, J., & Iyengar, R. (2013). Communication channels and word of mouth: How the medium shapes the message. Journal of Consumer Research, 40(3), 567–579. https://doi.org/10.1086/671345
- Lin, S., Warner, J., Zamfirescu-Pereira, J. D., Lee, M. G., Jain, S., Huang, M. X., Lertvittayakumjorn, P., Cai, S., Zhai, S., Hartmann, B., & Liu, C. (2024). Rambler: Supporting writing with speech via LLM-assisted gist manipulation. Proceedings of the 2024 CHI Conference on Human Factors in Computing Systems. https://doi.org/10.1145/3613904.3642217
- Yang, X., Li, W., Lee, M. G., Li, Z., Zamfirescu-Pereira, J. D., & Liu, C. (2025). Rambler in the wild: A diary study of LLM-assisted writing with speech. Extended Abstracts of the 2025 CHI Conference on Human Factors in Computing Systems. https://doi.org/10.1145/3706599.3706676
- Itzchakov, G., DeMarree, K. G., Kluger, A. N., & Turjeman-Levi, Y. (2018). The listener sets the tone: High-quality listening increases attitude clarity and behavior-intention consequences. Personality and Social Psychology Bulletin, 44(5), 762–778. https://doi.org/10.1177/0146167217747874
- Shen, H., & Sengupta, J. (2018). Word of mouth versus word of mouse: Speaking about a brand connects you to it more than writing does. Journal of Consumer Research, 45(3), 595–614. https://doi.org/10.1093/jcr/ucy011
- Gai, X., Gu, T., Wang, Y., & Jia, F. (2022). Improving career adaptability through motivational interview among peers: An intervention of at-risk Chinese college students majoring in foreign language. Journal of Vocational Behavior, 138, 103762. https://doi.org/10.1016/j.jvb.2022.103762
- Li, W., Tian, J., & Liu, C. (2026). Orality: A semantic canvas for externalizing and clarifying thoughts with speech. Proceedings of the 2026 CHI Conference on Human Factors in Computing Systems. https://doi.org/10.1145/3772318.3791713
- Schroeder, J., & Epley, N. (2015). The sound of intellect: Speech reveals a thoughtful mind, increasing a job candidate’s appeal. Psychological Science, 26(6), 877–891. https://doi.org/10.1177/0956797615572906
- ESAI. (2026, October 14). Marymount University × ESAI. https://www.esai.ai/blog-posts/marymount-university-x-esai
An ESAI perspective based on selected published research; not a systematic review or a report of an ESAI efficacy study. Findings from human interactions, consumer research, and small prototype studies are distinguished from ESAI’s product hypotheses. The cited researchers and institutions are not represented as endorsing ESAI.
