Prcvd is a software company founded in 2020 that is creating novel technologies to keep track of information from human speech and conversations.

While our communication habits have changed rapidly in recent decades, we still rely heavily on our default adaptations for face-to-face communication that have been around for millions of years. Our work at prcvd is guided by our unique ability to leverage knowledge about the evolution of human communication, behavioral science, and state-of-the-art artificial intelligence (AI).

The main goal of prcvd is to improve our communications by helping people understand how speakers are perceived, and by helping people revisit the variety of information produced in conversation. We aim to provide this information in real-time and through searchable archives by keeping track of who speaks, when they speak, what they say, and how they say it.

Troves of audio data are generated whenever people have spoken conversations, yet we rarely record, analyze, or revisit these. As AI voice recognition and transcription tools improve, organizations are realizing the power and importance of leveraging this conversation-based information to improve their interactions. By keeping track of conversations details and dynamics, organizations can improve efficiency and communication in meetings, better recruit and train staff, and enhance customer experience. Likewise, behavioral scientists, educators, journalists, and entertainers can use this technology for data collection and to improve their interactions.

Learn more about the current focus of software development at prcvd.

Learn more about how prcvd technologies will improve communication.

Sign up for **early access.

Software development at prcvd is currently focused on overcoming three major challenges to developing practical voice and speech recognition models for human conversations in the wild:

  1. We are finding computationally efficient solutions. Currently, many voice and speech recognition models available are difficult to train and deploy, making them inefficient for real-time solutions and sparse samples.

  2. Many systems perform poorly when audio quality is bad and samples sparse, however we are finding that if humans can understand speech from audio, then AI can too.

  3. Current systems can transcribe single speakers, but fail miserably well when multiple speakers overlap. Real-life human conversation is complex and often complicated by noise, cross-talk, and interruptions.

In our modern world many of us suffer from interpersonal stresses and communication fatigue. While some of this strain is due to the systematic ways that phone and computer conversations constrain human interaction, the majority of our communication problems are ancient –stemming from ageism, sexism, and other dominance-based inequities. For example, studies show that males contribute the majority of speech to conversations in professional meetings (1) and in classrooms (2), people perceive women to talk more than they actually do, people interrupt women more often than men (3–7) – especially in group conversations (8, 9), and that when doing the interrupting - women are often perceived negatively for it (10–12). Unfortunately, we often fail to notice and correct these problem behaviors (13–15). But there’s hope, if we can raise awareness of conversational dynamics and use tools to regulate our interactions, we are not bound to repeat our problems. If you want to improve communications, prcvd is here to help!


  1. C. F. Karpowitz, T. Mendelberg, The Silent Sex: Gender, Deliberation, and Institutions (Princeton University Press, 2014).

  2. M. Sadker, D. Sadker, Failing at Fairness: How America’s Schools Cheat Girls (Simon and Schuster, 2010).

  3. A. B. Hancock, B. A. Rubin, Influence of Communication Partner’s Gender on Language. Journal of Language and Social Psychology. 34, 46–64 (2015).

  4. C. West, Not just doctors’ orders: Directive-response sequences in patients’ visits to women and men physicians. Language and gender: A reader, 328–353 (1998).

  5. C. West, When the Doctor is a Lady: Power, Status, and Gender in Physician-Patient Encounters. Language and gender: A reader, 396–412 (1998).

  6. C. West, When the Doctor Is a “Lady”: Power, Status and Gender in Physician-Patient Encounters*. Symbolic Interaction. 7, 87–106 (1984).

  7. C. West, D. Zimmerman, Small Insults: a Study of Interruption in Cross-sex Conversations between unacquainted Persons”. In Henley, N., Kramarae, C., Thorne, B.(eds), Language, Gender and Society. Rowley, MA, Newbury House (1983).

  8. K. J. Anderson, C. Leaper, Meta-Analyses of Gender Effects on Conversational Interruption: Who, What, When, Where, and How. Sex Roles. 39, 225–252 (1998).

  9. C. W. Kennedy, thesis, The Ohio State University (1979).

  10. V. L. Brescoll, Who Takes the Floor and Why: Gender, Power, and Volubility in Organizations. Administrative Science Quarterly. 56, 622–641 (2011).

  11. E. R. Burris, The Risks and Rewards of Speaking Up: Managerial Responses to Employee Voice. AMJ. 55, 851–875 (2012).

  12. M. LaFrance, Gender and Interruptions: Individual Infraction or Violation of the Social Order? Psychology of Women Quarterly (2016) (available at

  13. J. D. Orcutt, D. L. Mennella, Gender and Perceptions of Interruption as Intrusive Talk: An Experimental Analysis and Reply to Criticism. Symbolic Interaction. 18, 59–72 (1995).

  14. A. Cutler, D. R. Scott, Speaker sex and perceived apportionment of talk. Applied Psycholinguistics. 11, 253–272 (1990).

  15. D. James, S. Clark, in Gender and Conversational Interaction. (Oxford University Press, New York, 1993).