# CSAT and NPS surveys in contact centers

Canonical: https://warmtransfer.net/knowledge/contact-center-surveys-csat-nps

Last verified: 2026-10-02

The Net Promoter Score (NPS) question asks how likely a respondent is to recommend the company to a friend or colleague, and the respondent answers on a 0 to 10 scale[^37]. Genesys Cloud, Amazon Connect, Webex Contact Center and Dynamics 365 Contact Center each document post-interaction surveys that run through voice flows, disconnect flows or digital channels[^27][^7][^50][^49][^14].

## How NPS is defined and calculated

NPS classifies respondents who answer 9-10 as promoters, those who answer 7-8 as passives, and those who answer 0-6 as detractors[^34]. The score is the percentage of respondents who are promoters minus the percentage who are detractors, and passives count only in the denominator[^35]. An NPS value ranges from -100 to +100[^38].

Frederick F. Reichheld introduced the would-recommend metric that became NPS in the Harvard Business Review article "The One Number You Need to Grow", in the December 2003 issue[^36]. The Harvard Business Review article "Stop Trying to Delight Your Customers", by Matthew Dixon, Karen Freeman and Nick Toman, appeared in the July-August 2010 issue[^11]. That article argued against the assumption that delighting customers drives loyalty[^11].

## Scales in use

| Question type | Scale | Source |
|---|---|---|
| NPS (standard definition) | 0 to 10 |[^37] |
| Webex Experience Management IVR CSAT | 1-5 |[^52] |
| Webex Experience Management IVR CES | 1-5 or 1-7 |[^51] |
| Webex Auto CSAT (applicable CSAT scales) | 3-point or 5-point |[^45] |

WarmTransfer's reading of the sources is that the Webex IVR NPS option uses 0-9 rather than the standard 0-10 scale[^54]. We infer that NPS collected through that option is probably not directly comparable with standard NPS unless the answers are re-banded[^54].

## Platform differences

### Genesys Cloud

A Genesys Cloud customer survey form can contain only 1 NPS question[^22]. That question can request an optional follow-up explanation[^20]. It can also control whether later questions display, based on whether the respondent is a promoter, passive or detractor[^20].

Genesys Cloud displays the NPS result as an integer rather than a percent[^21]. It calculates that result separately from the other survey questions and shows it in the interaction's quality summary[^21]. Survey results are linked to the interaction[^25]. They appear under Quality in the interaction's Quality Summary view, and in the Surveys Performance Summary and Detail views[^25].

**Web surveys.** Genesys Cloud sends web survey invitations by email from a configured domain and sender[^29]. A web survey policy requires a published web survey form and a published Architect survey invite flow[^29].

Genesys Cloud sends only 1 web survey per interaction, and when several web survey policies match, only 1 of them sends[^23]. When an interaction has several customer participants, only the most recent one receives the survey[^23]. Recipients have a 90-day window to respond before the survey expires, and reminder invitations can be spaced up to 90 days apart[^30].

**Voice surveys.** Genesys Cloud supports post-call voice surveys only for inbound calls[^27]. Invoking a voice survey flow requires 1 of the following licences[^28]:

- Genesys Cloud CX 1 WEM Add-on II[^28]
- Genesys Cloud CX 2 WEM Add-on I[^28]
- Genesys Cloud CX 3[^28]

The Ask Survey Question action exists only in voice survey flows[^17]. It handles multiple choice, Range, Yes/No and NPS questions, and it treats Range and Yes/No as multiple choice[^17].

Genesys Cloud documents these limits for voice surveys[^19][^26][^24][^18]:

- The agent (callee) cannot be transferred to a voice survey flow[^19].
- The post-flow does not trigger if the caller abandons before reaching an agent[^19].
- If a callback has been scheduled, the interaction goes through the voice survey before the callback can happen[^26].
- Invocation data is not available in the post-flow if the post-flow launches 10 days after the Set Post-Flow action ran[^24].
- If the voice survey flow disconnects before the Ask Survey Question action times out, the participant's response may not be fully captured[^18].

### Amazon Connect

**Voice.** Amazon Connect documents post-call surveys as a Set disconnect flow pattern[^7]. The pattern runs in this order[^7]:

- The agent asks the customer to stay on the line.[^7]
- The agent hangs up.[^7]
- The disconnect flow asks the survey questions with Get customer input.[^7]
- An AWS Lambda block uploads the answers to an external feedback database.[^7]

**Chat.** For post-chat surveys, set DisconnectOnCustomerExit to ["AGENT"] in StartChatContact[^3]. With that setting, the agent is disconnected automatically when the customer leaves, so the disconnect flow runs whichever participant leaves first[^3].

When the chat customer disconnects under this setting, the contact record shows disconnectReason CUSTOMER_DISCONNECT and a nextContactId[^4]. The survey then runs on that new contact, with initiationMethod DISCONNECT[^4].

Amazon Connect documents 2 ways to build the post-chat survey in the disconnect flow[^5]:

- a Show view block with a custom survey view[^5]
- an Amazon Lex bot for text-based collection[^5]

Responses appear as contact attributes in Contact Search and Contact Details[^5].

**Sample solution.** The aws-samples Contact Surveys for Amazon Connect solution stores survey definitions and results in Amazon DynamoDB[^8]. It selects the survey through a surveyId contact attribute and supports inbound voice, outbound voice and chat[^8]. If a flagged question receives a score below a set threshold, the solution automatically creates an Amazon Connect Task routed for supervisor review[^9].

### Webex Contact Center

**IVR surveys.** Post-call IVR surveys run through the Flow Designer Feedback V2 activity[^50]. That activity is valid only after the AgentDisconnect event, and Global_FeedbackSurveyOptin must be set to true[^50]. IVR surveys record partial responses when the caller disconnects or gives invalid input within the configured timeouts[^55].

**Digital-channel surveys.** Digital-channel surveys are delivered through Webex Connect flows[^49]. They are supported on these channels[^49]:

- Facebook Messenger[^49]
- live chat[^49]
- email[^49]
- SMS[^49]
- Apple Messages for Business[^49]
- WhatsApp[^49]

**Reporting and retention.** Survey responses are reported in Analyzer through the Survey Response Report, under Historical Reports, Post Interaction Surveys[^57]. Webex Experience Management keeps contact-center survey data for 400 days and then deletes it[^58].

**Auto CSAT.** Webex Contact Center Auto CSAT currently supports only voice interactions[^47]. It requires the AI Assistant, AI QM or AI Assistant QM Bundle add-on SKU A-FLEX-AI-ASST[^46].

Auto CSAT cannot run without the tenant's own CSAT survey data, because the model is trained on it[^42]. Its data requirements and training cycle are[^41][^39][^44]:

- at least 250 interactions with CSAT survey responses before it starts predicting[^41]
- at least 50 survey responses per CSAT scale point[^39]
- retraining every 30 days on the preceding 180 days of data[^44]

Cisco states that Auto CSAT predictions that deviate from actual survey scores fall within plus or minus 1 of the actual score[^43]. Auto CSAT scores are published in these places[^48]:

- Call Session Records, readable through the Search API[^48]
- the AI Assistant dashboard in Analyzer[^48]
- the Supervisor Desktop Interaction Details page[^48]

### Dynamics 365 Contact Center

**Voice.** From August 1, 2025, the post-call survey option on the Language tab of a voice workstream is deprecated[^16]. Microsoft recommends feedback surveys configured with Copilot Studio instead[^16]. The documentation describes 3 post-call survey consent modes[^15]:

- automatic-implicit[^15]
- automatic-explicit, which keeps consent in the va_SurveyConsent Boolean[^15]
- representative-initiated transfer to a survey queue[^15]

**Post-conversation surveys.** Post-conversation surveys are built in Customer Voice[^14]. They are not available in Government Community Cloud[^14]. They can be enabled for these channels[^14]:

- live chat[^14]
- Twilio and TeleSign SMS[^14]
- Facebook[^14]
- LINE[^14]
- WhatsApp[^14]
- custom channels[^14]
- Microsoft Teams[^14]
- Apple Messages for Business[^14]

Responses are logged as Anonymous (anonymous-post-conversation@omnichannelengagementhub.com)[^12]. The survey data is not hosted or stored in Dynamics 365 Contact Center[^12].

Dynamics 365 Contact Center can also send post-conversation surveys for conversations handled by AI agents[^13]. Those surveys can use the same settings as representative-handled conversations or separate ones[^13].

### Zendesk and NICE

By default, Zendesk sends its CSAT survey 24 hours after a ticket is solved, not closed[^59]. A system automation sends the survey, and admins can change its delay[^59].

NICE markets Enlighten AI for CSAT as scoring agent soft-skill behaviours on all interactions, such as empathy, acknowledging loyalty, ownership and effective questions[^31]. NICE does not publish an accuracy figure for it on its product page[^31].

## Response bias and interpretation

A 2016 BMC Health Services Research study covered 16,779 orthopaedic outpatients who were sent the Press-Ganey satisfaction survey[^33]. It found a 16.5% response rate, with older, female and privately insured patients more likely to respond[^33]. The study concluded that the survey showed non-response bias[^33].

Cisco's CSAT insights article states that customer survey response rates are low and skewed toward extreme experiences[^56]. The same article presents Auto CSAT as the answer to that gap[^56].

WarmTransfer's reading of the sources is that contact-center post-interaction surveys with low response rates are likely subject to the same kind of non-response bias[^32]. We infer that raw CSAT or NPS from respondents should therefore not be read as the score of all customers[^32].

We infer that because Webex Auto CSAT is trained on the tenant's own survey responses, any non-response or extreme-response bias in those surveys is likely to carry into the predicted scores[^2].

This is inferred: some surveys are offered only when the agent asks or transfers, as in the agent-initiated mode in Dynamics 365 and Amazon Connect's documented pattern[^1]. We infer that in those designs, agents control who is surveyed, which can bias results toward favourable calls[^1].

## See also

See also [Contact center KPI definitions and calculation](https://warmtransfer.net/knowledge/contact-center-kpis).
See also [Genesys Cloud Architect flows](https://warmtransfer.net/knowledge/genesys-architect-flows).
See also [Amazon Connect contact flows and routing profiles](https://warmtransfer.net/knowledge/amazon-connect-flows).
See also [Troubleshooting Webex Contact Center post-call surveys that never play](https://warmtransfer.net/knowledge/wxcc-post-call-survey-troubleshooting).
See also [Webex Contact Center Analyzer and the legacy reporting split](https://warmtransfer.net/knowledge/webex-cc-analyzer-reporting).
See also [Microsoft Dynamics 365 Contact Center](https://warmtransfer.net/knowledge/dynamics-365-contact-center).

## Applicability

Applies to: Bain Net Promoter System, Genesys Cloud, Amazon Web Services Amazon Connect, Cisco Webex Contact Center, Microsoft Dynamics 365 Contact Center, Zendesk Support, and NICE CXone. Deployments: multi-tenant. Sources checked 2026-10-02. Invoking a Genesys Cloud voice survey flow requires Genesys Cloud CX 1 WEM Add-on II, Genesys Cloud CX 2 WEM Add-on I, or Genesys Cloud CX 3[^28]. The Dynamics 365 Contact Center voice workstream survey option is deprecated from August 1, 2025[^16]. Dynamics 365 Contact Center post-conversation surveys are not available in Government Community Cloud[^14]. Webex Auto CSAT supports English interactions, and its multi-language support was in beta as of the 2026-05-17 article[^40]. The Press-Ganey non-response study covers US orthopaedic outpatients, not contact-center customers[^33].

## What remains uncertain

Whether Genesys Cloud offers a predicted CSAT or NPS score, and what its prerequisites would be, is not covered by the sources below. FCC TCPA consent rules for outbound or callback survey calls that use prerecorded voice or autodialers are not covered by the sources below. Microsoft recommends feedback surveys configured with Copilot Studio in place of the deprecated voice survey option[^16]. How those Copilot Studio surveys are configured is not covered by the sources below. How Webex Experience Management bands a 0-9 IVR NPS answer into promoters and detractors is not covered by the sources below. Post-call survey configuration and result linkage on platforms other than those documented above is not covered by the sources below. The only NICE claim here concerns Enlighten AI for CSAT[^31]. NICE survey and feedback management admin documentation is not covered by the sources below. In the documented Amazon Connect voice pattern, the agent hangs up before the disconnect flow runs[^7]. Whether a voice disconnect flow runs after the customer hangs up is not covered by the sources below. The lawful basis and consent for survey processing and survey recording under GDPR are not covered by the sources below.

## Sources

[^1]: Where a survey is offered only when the agent asks or transfers (agent-initiated modes in Dynamics 365 and Amazon Connect's documented pattern), agents control who is surveyed, which can bias results toward favourable calls (inferred). Source: [Configure a post-call survey in the voice channel](https://learn.microsoft.com/en-us/dynamics365/customer-service/administer/voice-channel-survey), introduction, 'Agent-initiated' consent mode; compare aws-connect-set-disconnect-flow-block Description example. Checked 2026-10-02.
[^2]: Because Webex Auto CSAT is trained on the tenant's own survey responses, any non-response or extreme-response bias in those surveys is likely to carry into the predicted scores (inferred). Source: [Measure customer satisfaction with Auto CSAT](https://help.webex.com/en-us/article/l59wt8/Measure-customer-satisfaction-with-Auto-CSAT), section 'Frequently Asked Questions (FAQ)' (training on survey data); read with nonresponse-bias claims. Checked 2026-10-02.
[^3]: For Amazon Connect post-chat surveys, setting DisconnectOnCustomerExit to ["AGENT"] in StartChatContact disconnects the agent automatically when the customer leaves so the disconnect flow runs whichever participant leaves first. Source: [Enable post-chat survey](https://docs.aws.amazon.com/connect/latest/adminguide/enable-post-chat-survey.html), intro paragraph and section 'For Custom Chat Widget'. Checked 2026-10-02.
[^4]: When an Amazon Connect chat customer disconnects with DisconnectOnCustomerExit configured, the contact record shows disconnectReason CUSTOMER_DISCONNECT and a nextContactId, and the survey runs on that new contact with initiationMethod DISCONNECT. Source: [Enable post-chat survey](https://docs.aws.amazon.com/connect/latest/adminguide/enable-post-chat-survey.html), section 'Update contact flow to add post-chat survey as a disconnect flow', Contact Trace Records note. Checked 2026-10-02.
[^5]: Amazon Connect documents two ways to build a post-chat survey in the disconnect flow: a Show view block with a custom survey view, or an Amazon Lex bot for text-based collection; responses appear as contact attributes in Contact Search and Contact Details. Source: [Enable post-chat survey](https://docs.aws.amazon.com/connect/latest/adminguide/enable-post-chat-survey.html), section 'Update contact flow to add post-chat survey as a disconnect flow'. Checked 2026-10-02.
[^6]: The Amazon Connect Set disconnect flow page's Tip states that no flow can be invoked when the customer disconnects; the flow ends and the agent enters After Call Work (disputed). Source: [Flow block in Connect Customer: Set disconnect flow](https://docs.aws.amazon.com/connect/latest/adminguide/set-disconnect-flow.html), section 'Description', Tip box. Checked 2026-10-02.
[^7]: Amazon Connect documents post-call surveys as a Set disconnect flow pattern: the agent asks the customer to stay on the line, the agent hangs up, the disconnect flow asks questions with Get customer input and an AWS Lambda block uploads answers to an external feedback database. Source: [Flow block in Connect Customer: Set disconnect flow](https://docs.aws.amazon.com/connect/latest/adminguide/set-disconnect-flow.html), section 'Description', post-contact survey example. Checked 2026-10-02.
[^8]: The aws-samples Contact Surveys for Amazon Connect solution stores survey definitions and results in Amazon DynamoDB, selects the survey through a surveyId contact attribute, and supports inbound voice, outbound voice and chat. Source: [Contact Surveys for Amazon Connect](https://github.com/aws-samples/amazon-connect-contact-surveys), README architecture and usage sections. Checked 2026-10-02.
[^9]: In the aws-samples contact survey solution, a score below a set threshold on a flagged question automatically creates an Amazon Connect Task routed for supervisor review. Source: [Contact Surveys for Amazon Connect](https://github.com/aws-samples/amazon-connect-contact-surveys), README feature description (flagging low scores). Checked 2026-10-02.
[^10]: The Amazon Connect Set disconnect flow page lists a customer hang-up on a voice call as a disconnect event that runs the configured disconnect flow (disputed). Source: [Flow block in Connect Customer: Set disconnect flow](https://docs.aws.amazon.com/connect/latest/adminguide/set-disconnect-flow.html), section 'Description', first bullet list (disconnect event definition). Checked 2026-10-02.
[^11]: The Harvard Business Review article 'Stop Trying to Delight Your Customers' by Matthew Dixon, Karen Freeman and Nick Toman appeared in the July-August 2010 issue and argued against the assumption that delighting customers drives loyalty. Source: [Stop Trying to Delight Your Customers](https://hbr.org/2010/07/stop-trying-to-delight-your-customers), article metadata and summary line. Checked 2026-10-02.
[^12]: Dynamics 365 Contact Center post-conversation survey responses are logged as Anonymous (anonymous-post-conversation@omnichannelengagementhub.com) and the survey data is not hosted or stored in Dynamics 365 Contact Center. Source: [Configure post-conversation survey using Customer Voice](https://learn.microsoft.com/en-us/dynamics365/customer-service/administer/configure-post-conversation-survey), section 'Runtime experience of post-conversation surveys', final paragraph. Checked 2026-10-02.
[^13]: Dynamics 365 Contact Center can send post-conversation surveys for conversations handled by AI agents, with the same or separate survey settings from representative-handled conversations. Source: [Configure post-conversation survey using Customer Voice](https://learn.microsoft.com/en-us/dynamics365/customer-service/administer/configure-post-conversation-survey), section 'Configure the post-conversation survey', step 6 (Bot conversation survey). Checked 2026-10-02.
[^14]: Dynamics 365 Contact Center post-conversation surveys built in Customer Voice can be enabled for live chat, Twilio and TeleSign SMS, Facebook, LINE, WhatsApp, custom channels, Microsoft Teams and Apple Messages for Business, but not in Government Community Cloud. Source: [Configure post-conversation survey using Customer Voice](https://learn.microsoft.com/en-us/dynamics365/customer-service/administer/configure-post-conversation-survey), section 'How post-conversation surveys work' and top Note. Checked 2026-10-02.
[^15]: Dynamics 365 Contact Center documents three post-call survey consent modes: automatic-implicit, automatic-explicit (consent kept in the va_SurveyConsent Boolean) and representative-initiated transfer to a survey queue. Source: [Configure a post-call survey in the voice channel](https://learn.microsoft.com/en-us/dynamics365/customer-service/administer/voice-channel-survey), introduction, consent list; section 'Set up agent-initiated survey'. Checked 2026-10-02.
[^16]: From August 1, 2025 the post-call survey option on the Language tab of a Dynamics 365 Contact Center voice workstream is deprecated, and Microsoft recommends feedback surveys configured with Copilot Studio instead. Source: [Configure a post-call survey in the voice channel](https://learn.microsoft.com/en-us/dynamics365/customer-service/administer/voice-channel-survey), Important box at top of article. Checked 2026-10-02.
[^17]: The Genesys Cloud Ask Survey Question action exists only in voice survey flows and handles multiple choice, Range, Yes/No and NPS questions, with Range and Yes/No treated as multiple choice. Source: [Ask Survey Question action](https://help.genesys.cloud/articles/ask-survey-question-action/), page body, supported question types. Checked 2026-10-02.
[^18]: If a Genesys Cloud voice survey flow disconnects before the Ask Survey Question action times out, the participant's response may not be fully captured. Source: [Ask Survey Question action](https://help.genesys.cloud/articles/ask-survey-question-action/), page body, response capture note. Checked 2026-10-02.
[^19]: Genesys Cloud does not support transferring the agent (callee) to a voice survey flow, and the post-flow does not trigger if the caller abandons before reaching an agent. Source: [Set Post-Flow action](https://help.genesys.cloud/articles/set-post-flow-action/), page body, limitations. Checked 2026-10-02.
[^20]: A Genesys Cloud NPS question can request an optional follow-up explanation and can drive conditional display of later questions based on whether the respondent is a promoter, passive or detractor. Source: [Net Promoter Score customer survey question overview](https://help.genesys.cloud/articles/net-promotor-score-web-survey-question-overview/), section 'Use an NPS question in a survey form'. Checked 2026-10-02.
[^21]: Genesys Cloud displays the NPS result as an integer rather than a percent, calculated separately from the other survey questions and shown in the interaction's quality summary. Source: [Net Promoter Score customer survey question overview](https://help.genesys.cloud/articles/net-promotor-score-web-survey-question-overview/), section 'NPS survey question scores and meaning'. Checked 2026-10-02.
[^22]: A Genesys Cloud customer survey form can contain only one NPS question. Source: [Net Promoter Score customer survey question overview](https://help.genesys.cloud/articles/net-promotor-score-web-survey-question-overview/), section 'Use an NPS question in a survey form'. Checked 2026-10-02.
[^23]: Genesys Cloud sends only one web survey per interaction; when several web survey policies match, only one sends, and multi-customer interactions survey only the most recent customer participant. Source: [Create a web survey policy](https://help.genesys.cloud/articles/create-a-web-survey-policy/), page body, notes on policy matching. Checked 2026-10-02.
[^24]: In Genesys Cloud, invocation data is not available in the post-flow if the post-flow launches 10 days after the Set Post-Flow action executed. Source: [Set Post-Flow action](https://help.genesys.cloud/articles/set-post-flow-action/), page body, invocation data note. Checked 2026-10-02.
[^25]: Genesys Cloud survey results are linked to the interaction and appear under Quality in the interaction's Quality Summary view and in the Surveys Performance Summary and Detail views. Source: [About customer surveys](https://help.genesys.cloud/articles/about-customer-surveys/), page body, survey results section. Checked 2026-10-02.
[^26]: When a callback has been scheduled, Genesys Cloud routes the interaction through the voice survey before the callback can happen. Source: [Set Post-Flow action](https://help.genesys.cloud/articles/set-post-flow-action/), page body, callback note. Checked 2026-10-02.
[^27]: Genesys Cloud supports post-call voice surveys only for inbound calls. Source: [Set Post-Flow action](https://help.genesys.cloud/articles/set-post-flow-action/), page body, voice survey notes. Checked 2026-10-02.
[^28]: Invoking a Genesys Cloud voice survey flow requires Genesys Cloud CX 1 WEM Add-on II, Genesys Cloud CX 2 WEM Add-on I, or Genesys Cloud CX 3. Source: [Set Post-Flow action](https://help.genesys.cloud/articles/set-post-flow-action/), page body, licence requirement note. Checked 2026-10-02.
[^29]: Genesys Cloud web survey invitations are sent by email from a configured domain and sender, and a web survey policy requires a published web survey form and a published Architect survey invite flow. Source: [Create a web survey policy](https://help.genesys.cloud/articles/create-a-web-survey-policy/), Prerequisites and invitation settings. Checked 2026-10-02.
[^30]: Genesys Cloud web survey recipients have a 90-day window to respond before the survey expires, and reminder invitations can be spaced up to 90 days apart. Source: [Create a web survey policy](https://help.genesys.cloud/articles/create-a-web-survey-policy/), page body, reminder and expiry settings. Checked 2026-10-02.
[^31]: NICE markets Enlighten AI for CSAT as scoring agent soft-skill behaviours (such as empathy, acknowledging loyalty, ownership and effective questions) on all interactions, without publishing an accuracy figure on its product page. Source: [NiCE Enlighten AI for Customer Satisfaction](https://www.nice.com/products/enlighten-ai-customer-satisfaction-csat), product page body. Checked 2026-10-02.
[^32]: Contact-center post-interaction surveys with low response rates are likely subject to the same kind of non-response bias, so raw CSAT or NPS from respondents should not be read as the score of all customers (inferred). Source: [Evidence of non-response bias in the Press-Ganey patient satisfaction survey](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4972948/), Conclusions; read with cisco-help-q2tw3s-wxcc-csat-insights 'Challenges with limited data and feedback'. Checked 2026-10-02.
[^33]: A 2016 BMC Health Services Research study of 16,779 orthopaedic outpatients found a 16.5% response rate to the Press-Ganey satisfaction survey, with older, female and privately insured patients more likely to respond, and concluded the survey showed non-response bias. Source: [Evidence of non-response bias in the Press-Ganey patient satisfaction survey](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4972948/), Abstract (Results and Conclusions). Checked 2026-10-02.
[^34]: NPS classifies 9-10 respondents as promoters, 7-8 as passives and 0-6 as detractors. Source: [Measuring Your Net Promoter Score](https://www.netpromotersystem.com/about/measuring-your-net-promoter-score/), page body, promoters/passives/detractors list. Checked 2026-10-02.
[^35]: NPS is calculated as the percentage of respondents who are promoters minus the percentage who are detractors; passives count only in the denominator. Source: [Measuring Your Net Promoter Score](https://www.netpromotersystem.com/about/measuring-your-net-promoter-score/), page body, calculation paragraph. Checked 2026-10-02.
[^36]: The would-recommend metric that became NPS was introduced by Frederick F. Reichheld in the Harvard Business Review article 'The One Number You Need to Grow', December 2003 issue. Source: [The One Number You Need to Grow](https://hbr.org/2003/12/the-one-number-you-need-to-grow), article metadata (author line and issue date). Checked 2026-10-02.
[^37]: The Net Promoter Score question asks how likely the respondent is to recommend the company to a friend or colleague and is answered on a 0 to 10 scale. Source: [Measuring Your Net Promoter Score](https://www.netpromotersystem.com/about/measuring-your-net-promoter-score/), page body, question and scale paragraph. Checked 2026-10-02.
[^38]: An NPS value ranges from -100 to +100. Source: [Measuring Your Net Promoter Score](https://www.netpromotersystem.com/about/measuring-your-net-promoter-score/), page body, calculation paragraph. Checked 2026-10-02.
[^39]: Webex Auto CSAT requires at least 50 survey responses per CSAT scale point. Source: [Measure customer satisfaction with Auto CSAT](https://help.webex.com/en-us/article/l59wt8/Measure-customer-satisfaction-with-Auto-CSAT), section 'Frequently Asked Questions (FAQ)'. Checked 2026-10-02.
[^40]: Webex Auto CSAT supports English interactions, with multi-language support in beta as of the 2026-05-17 article. Source: [Measure customer satisfaction with Auto CSAT](https://help.webex.com/en-us/article/l59wt8/Measure-customer-satisfaction-with-Auto-CSAT), section 'Prerequisites to enable Auto CSAT'. Checked 2026-10-02.
[^41]: Webex Auto CSAT needs at least 250 interactions with CSAT survey responses before it starts predicting. Source: [Measure customer satisfaction with Auto CSAT](https://help.webex.com/en-us/article/l59wt8/Measure-customer-satisfaction-with-Auto-CSAT), section 'Frequently Asked Questions (FAQ)'. Checked 2026-10-02.
[^42]: Webex Auto CSAT cannot run without the tenant's own CSAT survey data because the model is trained on it. Source: [Measure customer satisfaction with Auto CSAT](https://help.webex.com/en-us/article/l59wt8/Measure-customer-satisfaction-with-Auto-CSAT), section 'Frequently Asked Questions (FAQ)'. Checked 2026-10-02.
[^43]: Cisco states that Webex Auto CSAT predictions that deviate from actual survey scores fall within plus or minus 1 of the actual score. Source: [Measure customer satisfaction with Auto CSAT](https://help.webex.com/en-us/article/l59wt8/Measure-customer-satisfaction-with-Auto-CSAT), section 'Frequently Asked Questions (FAQ)'. Checked 2026-10-02.
[^44]: Webex Auto CSAT models are retrained every 30 days on the preceding 180 days of data. Source: [Measure customer satisfaction with Auto CSAT](https://help.webex.com/en-us/article/l59wt8/Measure-customer-satisfaction-with-Auto-CSAT), section 'Frequently Asked Questions (FAQ)'. Checked 2026-10-02.
[^45]: Webex Auto CSAT applies to customers that measure CSAT on a 3-point or 5-point scale. Source: [Measure customer satisfaction with Auto CSAT](https://help.webex.com/en-us/article/l59wt8/Measure-customer-satisfaction-with-Auto-CSAT), section 'Prerequisites to enable Auto CSAT'. Checked 2026-10-02.
[^46]: Webex Auto CSAT requires the AI Assistant, AI QM or AI Assistant QM Bundle add-on SKU A-FLEX-AI-ASST. Source: [Measure customer satisfaction with Auto CSAT](https://help.webex.com/en-us/article/l59wt8/Measure-customer-satisfaction-with-Auto-CSAT), section 'Prerequisites to enable Auto CSAT'. Checked 2026-10-02.
[^47]: Webex Contact Center Auto CSAT currently supports only voice interactions. Source: [Measure customer satisfaction with Auto CSAT](https://help.webex.com/en-us/article/l59wt8/Measure-customer-satisfaction-with-Auto-CSAT), section 'Prerequisites to enable Auto CSAT'. Checked 2026-10-02.
[^48]: Webex Auto CSAT scores are published in Call Session Records (readable through the Search API), the AI Assistant dashboard in Analyzer, and the Supervisor Desktop Interaction Details page. Source: [Measure customer satisfaction with Auto CSAT](https://help.webex.com/en-us/article/l59wt8/Measure-customer-satisfaction-with-Auto-CSAT), section 'View Auto CSAT scores'. Checked 2026-10-02.
[^49]: Webex Contact Center digital-channel surveys are supported on Facebook Messenger, live chat, email, SMS, Apple Messages for Business and WhatsApp, and are delivered through Webex Connect flows. Source: [Experience Management - Configure surveys for IVR and Digital Channels for Webex Contact Center](https://help.webex.com/en-us/article/nlu4x20/Experience-Management---Configure-surveys-for-IVR-and-Digital-Channels-for-WebexContact-Center), digital channel survey section. Checked 2026-10-02.
[^50]: In Webex Contact Center, post-call IVR surveys run through the Flow Designer Feedback V2 activity, which is valid only after the AgentDisconnect event and needs Global_FeedbackSurveyOptin set to true. Source: [Experience Management - Configure surveys for IVR and Digital Channels for Webex Contact Center](https://help.webex.com/en-us/article/nlu4x20/Experience-Management---Configure-surveys-for-IVR-and-Digital-Channels-for-WebexContact-Center), IVR survey configuration section, Feedback V2 activity. Checked 2026-10-02.
[^51]: The Webex Experience Management IVR survey CES question type offers a 1-5 or a 1-7 rating scale. Source: [Experience Management - Configure surveys for IVR and Digital Channels for Webex Contact Center](https://help.webex.com/en-us/article/nlu4x20/Experience-Management---Configure-surveys-for-IVR-and-Digital-Channels-for-WebexContact-Center), IVR question options, 'CES: Set the following parameters'. Checked 2026-10-02.
[^52]: The Webex Experience Management IVR survey CSAT question type uses a 1-5 scale. Source: [Experience Management - Configure surveys for IVR and Digital Channels for Webex Contact Center](https://help.webex.com/en-us/article/nlu4x20/Experience-Management---Configure-surveys-for-IVR-and-Digital-Channels-for-WebexContact-Center), IVR question options, 'CSAT: This option populates a 1-5 scale.'. Checked 2026-10-02.
[^53]: The Webex Experience Management IVR survey NPS question type uses a 0-9 scale entered as a single keypad digit. Source: [Experience Management - Configure surveys for IVR and Digital Channels for Webex Contact Center](https://help.webex.com/en-us/article/nlu4x20/Experience-Management---Configure-surveys-for-IVR-and-Digital-Channels-for-WebexContact-Center), IVR question options, 'NPS: This option populates a 0-9 scale.'. Checked 2026-10-02.
[^54]: Because the Webex IVR NPS option uses 0-9 rather than the 0-10 scale Bain defines, IVR-collected NPS from that option is probably not directly comparable with standard NPS without re-banding (inferred). Source: [Experience Management - Configure surveys for IVR and Digital Channels for Webex Contact Center](https://help.webex.com/en-us/article/nlu4x20/Experience-Management---Configure-surveys-for-IVR-and-Digital-Channels-for-WebexContact-Center), IVR question options, NPS option; compared with bain-nps-measuring-your-nps bands. Checked 2026-10-02.
[^55]: Webex IVR surveys record partial survey responses when the caller disconnects or gives invalid input within the configured timeouts. Source: [Experience Management - Configure surveys for IVR and Digital Channels for Webex Contact Center](https://help.webex.com/en-us/article/nlu4x20/Experience-Management---Configure-surveys-for-IVR-and-Digital-Channels-for-WebexContact-Center), IVR survey section, partial response handling. Checked 2026-10-02.
[^56]: Cisco's CSAT insights article states that customer survey response rates are low and skewed toward extreme experiences, and presents Auto CSAT as the answer to that gap. Source: [Enhance performance and customer service using CSAT insights](https://help.webex.com/en-us/article/q2tw3s/Enhance-performance-and-customer-service-using-CSAT-insights), section 'Challenges with limited data and feedback'. Checked 2026-10-02.
[^57]: Webex Contact Center survey responses are reported in Analyzer through the Survey Response Report under Historical Reports, Post Interaction Surveys. Source: [Experience Management - Configure surveys for IVR and Digital Channels for Webex Contact Center](https://help.webex.com/en-us/article/nlu4x20/Experience-Management---Configure-surveys-for-IVR-and-Digital-Channels-for-WebexContact-Center), reporting section. Checked 2026-10-02.
[^58]: Webex Experience Management retains contact-center survey data for 400 days and then deletes it. Source: [Experience Management - Configure surveys for IVR and Digital Channels for Webex Contact Center](https://help.webex.com/en-us/article/nlu4x20/Experience-Management---Configure-surveys-for-IVR-and-Digital-Channels-for-WebexContact-Center), data retention statement. Checked 2026-10-02.
[^59]: Zendesk sends its CSAT survey by default 24 hours after a ticket is solved (not closed), through a system automation whose delay admins can change. Source: [Customizing when your customer satisfaction (CSAT) survey is sent](https://support.zendesk.com/hc/en-us/articles/4408886194202-Customizing-when-your-customer-satisfaction-CSAT-survey-is-sent), page body, default timing paragraph. Checked 2026-10-02.