With Retell, we’ve increased scheduling NPS by 38%, and filled underutilized provider capacity, allowing our team to focus on meaningful patient care instead of phone tag.
-p-130x130q80.avif)
Phone Survey Software for CSAT, NPS, and Research Calls
5.0
ON G2
Run phone surveys with an AI voice agent instead of a recorded IVR survey or a call team. The agent asks your questions in a natural voice, takes spoken or keypad answers, follows your skip rules, and sends every response to your systems as structured fields. Use it right after support calls, for outbound CSAT and NPS, or for research and verification calls.
Trusted by over 4,000 businesses powering 100+ million calls per month.
.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
.svg)
How It Works
Launch AI phone surveys in four stages, from question design to responses landing in your database.
Write questions and skip rules
Write each question with its answer type, such as a one to five scale, zero to ten, or yes or no, add the skip rules, and define a post-call field for every answer you want stored.
Test every answer path
Run simulated calls with off-scale answers, keypad entries, respondents who ask questions back, and people who refuse, and grade every run before a real person is called.
Connect your calls and lists
Take post-call transfers from your contact center over SIP, or upload a list as a batch call with a calling window, and choose the number the survey calls from.
Go live and read the results
Each call returns its answers as structured fields by webhook and API, so scores, skipped questions, refusals, and completion rates reach your database without anyone keying them in.
Use Cases
Phone surveys run from the first question to the record in your database
Post-call CSAT after a support call
Need feedback while the call is still fresh, not an email the customer ignores two days later. When your support agent finishes, your contact center transfers the caller to the survey agent's number or SIP address, and the survey starts on the same call. Pass the case ID or the support agent's name in a custom SIP header, so the greeting refers to the call they just had and every score is tied to the right interaction. Teams that also put a voice agent on the support call itself use AI customer service.
NPS and CSAT scales, by voice or keypad
Need scores people can give however they prefer. The agent asks the question and accepts a spoken "eight" or a key press, with keypad capture set to the number of digits the scale needs, then stores the score as a number field and the reason as a text field. Skip rules follow your logic: skip the recommendation question when the caller says the issue was not solved, or when the call ended in a transfer. For what a score means once you have it, see what a good NPS score is.
Outbound surveys after a visit, delivery, or ticket
Need to hear from customers who never open survey emails. Upload a list with batch calling: a phone number column plus columns such as first name, visit date, or ticket number, which the agent uses on the call. Set the calling window, schedule the batch, and choose whether the agent hangs up or leaves a short message when it reaches voicemail. A metadata column keeps your own record ID on every call, so each response maps back without guesswork.
Verification callbacks for completed surveys
Need to confirm that a survey submitted online came from the real person, without a team dialing every one. On the schedule you choose after the form comes in, the agent calls the number on file, asks whether they took the survey, and confirms identity with details you pass in, without asking for sensitive data such as a date of birth. The verified flag, the date confirmed, and any re-asked scores go back to your database by webhook. Calls nobody answers become a short list for your staff.
Research and opinion surveys with screening
Need market research or public opinion interviews in the respondent's own language. The agent says it is an AI at the start, asks your screening questions such as age eligibility, and ends politely when someone does not qualify. To avoid order bias, pass a different question order for each respondent as a variable in your list, and the agent asks in that order. When a respondent asks who is running the study or how the data is used, it answers from a knowledge base you approve, in Spanish, English, or many other languages.
Calling windows, refusals, and do-not-call requests
Need a survey program that respects when, and whether, people want to be called. Each batch runs only inside the time window you set, a refusal is recorded as its own field so you drop that number from the next list, and do-not-contact handling confirms a stop-calling request and marks the contact. Calls placed through the API can skip marked numbers automatically, while batch calls do not check the flag, so filter those lists first. Outbound survey calls with an AI voice carry consent rules in the US; the TCPA guide for AI voice calls covers them.
Built for Contact Centers, Not Demos
Retell is the AI Voice Platform Enterprise Contact Centers Run On
Retell lets you build, test and deploy AI phone agents that handle real call volume, so your team spends its time on the conversations that need a person.
Feature
Why CX and research teams choose Retell AI for phone surveys
Questions, scales, and skip rules you control
Survey logic breaks when someone outside your team has to change it. A conversation flow gives each question its own node, and logic split nodes branch on what the respondent said or on data you passed in, so a skip rule is a setting, not a script rewrite. You edit wording yourself and publish a new version when the change is tested.
Answers stored as fields, not buried in transcripts
A survey is only useful if the answers are countable. Post-call extraction writes each answer to the field type you choose: a number for a score, a choice for a category, yes or no for a screen, and text for the reason behind a rating. Fields arrive in the call_analyzed webhook, ready for your database or dashboard.
Context from your systems on every call
Respondents notice when a survey knows nothing about them. Columns in a batch list, values in an API request, and custom SIP headers from your contact center become variables the agent can use, such as a first name, a visit date, or a case ID. A metadata field stores your record ID on the call for clean matching later.
Inbound transfers and outbound lists on one agent
Post-call CSAT and outbound NPS often end up in two different tools. Here one survey agent can answer transfers on its number and place outbound batches from it. When a batch runs, you reserve concurrency for other calls so transferred callers are not left waiting, and branded caller ID and verified numbers are available on US lines.
Pickup, completion, and score trends
You need to know how many people answered and finished, not only the average score. Each batch shows calls sent, picked up, and successful, with success defined by you as a completed survey. The analytics dashboard charts results by agent and version, and alerts tell you when completion rate drops.
-p-130x130q80.avif)
-p-130x130q80.avif)
-p-130x130q80.avif)
- Hubspot
- Twilio
- Vonage
- Go High Level
- n8n
- Zapier
- Salesforce
- Hubspot
- Twillio
- Vonage
- Go High Level
- 8n8
- Zapier
- Salesforce
- Avaya
- Genesys
- Five9
- Amazon Connect
- Telnyx
- Make
- Cal.com
.svg)
- Avaya
- Genesys
- Five9
- Amazon Connect
- Telnyx
- Make
- Cal.com
Pricing
Cut Costs by 90% with AI phone Agents
Pay as you go
$0 to start.
$0.07–$0.31 per minute – Pay only for what you use.
60 mins of free access.
20 concurrent calls.
10 Free Knowledge Bases.
Enterprise Plan
For companies with large volumes, data or deployment requirements, or support needs.
White glove service for your use case.
Premium support with dedicated solution engineer teams.
Custom pricing and concurrent calls based on your needs
FAQ
Frequently asked questions.
What is phone survey software?
Phone survey software runs surveys over phone calls instead of email, web, or text. Older tools play recorded questions and collect key presses, or give live interviewers a script and a form. An AI voice agent asks the questions in conversation, takes spoken or keypad answers, handles off-script questions, and stores each answer as a structured field, on inbound transfers and outbound lists alike.
Can respondents answer by keypad as well as by voice?
Yes. Keypad input is captured by default, so a respondent can say "four" or press 4. You can tell the agent when an answer is complete by a digit limit, a termination key such as pound, or a short timeout after the last key. Keypad entry also helps in noisy places where speech is hard to hear.
How do survey answers get into our database or CRM?
You define post-call extraction fields for each answer, as a number, yes or no, text, or a choice from a list. After the call, Retell sends them in the call_analyzed webhook and returns them from the Get Call API, so your system stores them without parsing transcripts. HubSpot, Salesforce, Microsoft Dynamics 365, GoHighLevel, and Zoho CRM connections can write the results back to contact fields.
Can the survey start automatically after a call in our contact center?
Yes, if your contact center can transfer the caller when your agent drops. The survey agent sits on a Retell number or SIP address, and the transfer is set up in your platform's routing. Retell's docs include SIP setup guides for Genesys Cloud, Five9, Avaya, and Amazon Connect. Pass the case ID in a custom SIP header to link each response to its interaction.
Are outbound survey calls with an AI voice legal?
It depends on where you call, who you call, and why. In the US, the FCC has said AI-generated voices count as artificial voices under the TCPA, so outbound survey calls to mobile numbers generally need the person's prior consent, and the call should say who is calling at the start. Other countries and US states add their own rules. Check your list and script with counsel before you launch.
How much does phone survey software cost on Retell?
Retell voice agents are priced per minute of conversation, from $0.07 per minute, depending on the language model, voice, and telephony you choose. Pay-as-you-go needs no contract, and Enterprise plans have custom pricing for high volume or custom deployment needs. A short CSAT survey costs only the minutes the respondent is on the line. See pricing for your mix.
