About & Method
Why "Press Zero"?
For years, pressing 0 was how you escaped the phone menu and reached a person. Now the person is often gone entirely. Press Zero is a free, non-commercial research project at the University of California, Riverside that measures what customer-service bots fail to provide, using your experiences.
The research team
Vivaan Rao
Student researcher · UC Riverside
Gary Zhang
Student researcher · UC Riverside
Dr. Aruhn Venkat
Faculty advisor · UC Riverside
Chris Thomas
Advisor · Former Chief Strategist at Intel · UC Riverside
The 8 dimensions
- Human Access
- Could I reach a real person when I needed one?
- Understanding
- Did it understand what I was asking?
- Resolution
- Did it actually fix my problem?
- Effort & Time
- Was it quick, or did I go in circles?
- Honesty
- Was it clear it was a bot, and honest about its limits?
- Accuracy
- Was the information correct?
- Respect & Care
- Did it treat me like a person, not a ticket number?
- Privacy & Control
- Did it respect my data and my choices?
How scores are calculated
Bot Gap Index (0–100, higher = bigger gap). For each rating we take every dimension the person scored and turn it into a gap: (5 − score) ÷ 4. Dimensions that person said matter most to them (their top 3) count twice. We average those into one number per rating, then average across all ratings for the bot and multiply by 100.
Failure rate. A dimension "fails" in a rating if it was scored 1 or 2. "Didn't come up" is ignored.
Importance. The share of all respondents who picked that dimension in their top 3.
Crowd Gap Score. Importance × failure rate, so problems that are both common and important rise to the top.
Small samples. We only show a bot's scores after 10 ratings, and we show a 95% confidence interval next to every index.
Quality. Very fast submissions and repeat ratings of the same bot within 7 days are flagged for review. Everything is crowd-reported and not verified by the companies.
Privacy
Your email is stored separately from your answers, is only visible to the research team, and is never shown publicly. We use it to recognize returning participants, send occasional research updates, and credit you if you opt in.
Public pages only show combined statistics and quotes the team has reviewed and stripped of personal details. Screenshots are private and never published. There are no ads and no tracking for marketing.
Request deletion
Email vrao019@ucr.edu from the address you used, and we'll delete your data.