Frequently Asked Questions

How does Phebi make survey and market research better?

Phebi uses speech recognition, respected behavioral science, machine learning, and AI to unlock what’s in people’s heads and hearts at scale. Its flagship software is used by agencies and brands in quantitative and qualitative health and market research, customer experience, and other applications to increase understanding, enhance storytelling, better predict behavior, and help inform better decisions that improve business results.

The software solves a strategic problem. Because people don’t always say what they really mean and feel, despite its widespread use, generative AI for text analysis can miss important context, engagement levels, and other factors needed to tell the full story. Reseachers who rely on it may realize cost savings at scale while missing out on deeper, differentiated information that is likely to lead to better business decisions.

By tapping into voice signals, Phebi introduces a new signal layer that creates relevant, in-depth intelligence that text-only systems simply can’t.

Behavioral sciencists know that capturing and analyzing how people say things can provide a richer understanding of people’s responses, intentions, wants, and needs in ways that correlate with better decisions and real-world outcomes. For example, it can help answer common questions like:

• How do patients and providers feel about a course of treatment?
• Which product or concept resonates most resonates with each type of consumer?
• Do their responses indicate strong or perfunctory engagement?
• Which issues are people really concerned about?
• Is this response from a person or a bot?

For example, imagine a voice-enabled survey of patients about a new medical treatment. Phebi scores the respondent’s spoken responses, identifying moments where, based on a comparison of their voice signals to trusted behavioral science, whether people are engaged or not, and if they likely feel calm, anxious, happy, or sad. Responses are transcribed, reported on, and visualized. They can be played back; filtered by subgroup, topic, emotion, or other variables; and downloaded. All of this opens up new levels of understanding and enables better outcomes.

How does Phebi work technically?

Phebi is based on well-accepted behavioral science research about the relationship between characteristics of speaking voices, such as volume, rate, and tone, and the speaker’s underlying emotion. Phebi uses a combination of voice tech, AI, and advanced machine learning techniques to collect and analyze audio. Each speaker’s vocal characteristics are analyzed every 3 seconds to produce reliable, unbiased results.

Phebi’s innovative technology ‘stack’ was built from the ground up, trained, and tested at Phebi Labs, where it is continually being improved as our team works with new customers and in new applications, and proven in ‘research-on-research’ by research professionals at the leading agencies and global brands that are Phebi clients.

What data does Phebi make available to me? How long is it available?

Phebi gives you the full picture. With it, you can learn:

• what people said (speech-to-text transcription with separate results for each speaker),
• the implications of the words they used (via sentiment analysis), and
• how they likely felt about the topic (via analyzed and quantified voice signals).
  

You may listen to and download spoken responses and segments of recorded audio and video and view intensity scores. If you’d like to, you may further explore the data in a portal to which your team has unlimited access. There you may do further analysis by tagging/coding segments and groups, using interactive visualizations, quickly creating showreels, and more.

If you use Phebi to voice-enable a survey, transcriptions of spoken or videoed responses are automatically fed back into your survey platform in near real-time. Phebi also exports data for use in other software applications, and data, visualizations, and showreels for presentations.

How can Phebi be deployed?

Phebi is like Swiss Army knife. It’s valuable in many ways across MR and CX use cases, including:

• You may create a seemless, voice-enabled audio or video experience by adding Phebi to a survey created in your favorite survey platform.
• If you have existing long-form audio or video as files from IDIs, groups, or other sources, you may upload them into Phebi.
• You may dial Phebi into an on-line focus group, interview, or confernce call to record, transcribe, translate, tag, etc. the event in real-time.
• If you’d like to put Phebi to use in your solution, you may use the Phebi API.
• You may add dashboards with information from Phebi to you contact center or other operational system.

Phebi also offers software for automated voice surveys (IVR), AI-moderated interviewing (AIMI), real-time response capture in CATI interviews, and more.

How do I add Phebi to a survey?

If you’re doing quant research, Phebi makes audio and video data collection simple via a no-code add-on to any market research platform called the Phebi plugin. It creates a seamless user experience, and automatically puts transcriptions of what people say back into your platform as if the responses had been typed. By enabling respondents to respond by speaking instead of typing, Phebi improves each respondent’s experience and generates longer, richer answers because when people talk instead of type they say more, especially on mobile.

Phebi easily integrates with any survey software or online form so you can realize its value with no disruption to your current processes or project timelines. And, it works with all the supported versions of the major browsers. Your team simply answers a few questions online and receives a small JavaScript and CSS plugin that is configured to your preferences for how your survey will deploy voice. Fast, friendly expert support is available.

Typically, what percentage of survey respondents choose to respond via voice?

Each survey is different in terms of the type of person who takes it, the ground it covers, where it is fielded, how the opportunity to use spoken responses is presented, and more. With those factors and others in play, Phebi’s customers typically find they can expect 25% to 50% of respondents to choose to respond via voice when given the option of talking or typing. Phebi will work with you to use voice best practices to help you collect more, richer data and to provide a brand-building user experience.

How do I use Phebi to analyze longer-form audio or video?

Phebi transcribes and analyzes audio and video files from focus groups, in-depth interviews, concept tests, call centers, etc. You simply select the files from wherever you’ve stored them to upload then into Phebi, and the software does the rest.

Are there any limits or size restrictions related to using Phebi?

With the exception of a limit of 100 clips in a showreel, there are no limits or size restrictions related to using Phebi.

In which countries is Phebi available?

Phebi is available in any country not restricted by U.S., U.K., or EU law.

What can we expect in terms of software availability? 

Providing customers ongoing access to all our software offerings is a high priority for us. We maintain high platform availability and reliability, with a monthly uptime percentage (the total number of minutes in a calendar month minus the number of minutes of downtime, divided by the total number of minutes in the calendar month) of at least 99.9% for the production Phebi platform, measured 24 hours per day, 7 days per week.

That uptime calculation excludes scheduled maintenance windows, which shall not exceed four hours per calendar month. We’ll communicated about scheduled maintenance in advance, and we’ll perform it outside of standard business hours whenever reasonably practicable. Emergency maintenance may be performed as necessary to preserve platform security, stability, or performance.

How is Phebi priced?

Software licenses for Phebi’s flagship product are based on the volume of audio or video processed and the type of processing (transcription only or transcription with analysis of voice signals); there are no ‘per seat’ charges for portal access. Phebi offers two pricing options: (1) pay-as-you-go, project-based pricing and (2) enterprise licenses with monthly payments. Value-added reseller and referral arrangements are available.

Two levels of software technical support and access to Phebi’s Knowledge Base are included in Phebi’s software licenses.

Phebi’s flagship software is DIY, there are no implementation fees associated with it. Other offertings, including Phebi AIMI, Phebi AVS, and Phebi Live have required set up fees.

Professional services, including Phebi’s FastStart service that adds expert resources to your team, helps create efficiencies, reduces delivery timing risk, and can deliver client-ready analysis, are available for an additional fee.

For more information at pricing, contact us at marketing@phebi.ai.