What is the '8. NLP Grouped by Type' Report in CDR for wolkvox Manager
Written by Jhon Bairon Figueroa
Updated at September 16th, 2026
Table of Contents
Introduction
The "8. NLP grouped by type" report, available within the CDR report group in wolkvox Manager, allows you to query and consolidate credit consumption generated by different features using Artificial Intelligence or NLP processing within the operation.
Every time a feature makes a request to an Artificial Intelligence model, a consumption record is generated. This report groups these records by the TYPE_NLP field, allowing you to identify which feature generated the consumption, how many executions it performed, and how many credits corresponded to sent requests and generated responses.
In this way, the report can be used to audit the use of AI features, identify which components account for a higher amount of credits, and link consumption to specific features of wolkvox Manager, wolkvox Studio, wvx Copilot, Speech & Text Analytics, Voice of Customer, and agentic solutions, among others.
Important: the number of executions does not correspond directly to the billed amount. Billing for NLP features is calculated based on consumed credits. The monetary rate applied will depend on the commercial terms in effect for the operation.
How to view the report?
- Go to the Reports section in the left side menu of wolkvox Manager.
- Set the Start Date and End Date for the period you wish to query.
- Access the CDR report group.
- Select the "8. NLP grouped by type" report.
- Click Generate report.
The system will display consumption grouped by each TYPE_NLP value recorded during the selected period.
What information does the report display?
| Field | Description |
|---|---|
| TYPE_NLP | Identifies the Artificial Intelligence feature, component, or process that generated the consumption. Depending on the type, the value may include at the end the identifier for the routing point, IVR, diagram, or flow where execution occurred. |
| NLP_AMOUNT | Number of events or executions recorded for that TYPE_NLP during the queried period. This value represents usage volume and does not correspond directly to cost. |
| CREDITS_SEND | Total credits associated with the information sent to the Artificial Intelligence model to make the request, analysis, or processing. |
| CREDITS_ANSWER | Total credits associated with the response or result generated by the Artificial Intelligence model. |
| TOTAL_CREDITS | Total credits consumed by that feature type during the queried period. Corresponds to the sum of CREDITS_SEND + CREDITS_ANSWER and is the primary value for analyzing NLP consumption. |
How are NLP credits calculated?
Consumption is generated when a feature makes a call to an Artificial Intelligence model. The resulting credits depend on the amount of information processed and the model used to serve the request.
A distinction is also made between consumption generated by information sent to the model and consumption corresponding to the response produced. For this reason, two different requests may generate different amounts of credits even if they correspond to the same TYPE_NLP.
The selected model can also influence consumption. An execution requiring a model with higher processing capabilities may represent more credits than an equivalent request processed by a lighter model.
Special consumption cases
Cache
When a response can be reused from cache, it is not necessary to perform full processing with the model again. This consumption may appear identified as nlp-cache or by the original value accompanied by the -cache suffix.
For example:
conversations-chat-61074-cache
Consumption served from cache corresponds to 10% of the credits that would apply to standard processing.
wvx Copilot
Records whose TYPE_NLP begins with nlp-Copilot correspond to wvx Copilot features.
In addition to variable consumption by NLP credits, wvx Copilot may include a fixed activation charge per period, according to the commercial terms applicable to the operation. The fixed charge does not replace credit consumption: both concepts can coexist when using the service.
Values TYPE_NLP can display
The TYPE_NLP column allows identifying which Artificial Intelligence feature generated the consumption recorded in the report. Each value corresponds to a specific function of wolkvox Manager, wolkvox Studio, Speech & Text Analytics, Voice of Customer, wvx Copilot, or conversational and agentic agents.
In some values, XXXXX is used to represent the identifier of the routing point, IVR, diagram, or flow from which the feature was executed. For example, conversations-chat-61074 indicates that the consumption was generated by a conversational chat agent associated with routing point 61074.
When a value ends in -na, it means that execution was not associated with a flow identifier, which can occur during testing or configuration actions. When it ends in -cache, it indicates that wolkvox reused a previously processed response and only 10% of equivalent credits are recorded.
| TYPE_NLP | What feature generates consumption? | When does it appear in the report? | Consumption type |
|---|---|---|---|
nlp-agentic-mysqldb-XXXXX |
Corresponds to wolkvox Studio's Agentic MySQL DB. AI interprets the request and queries or processes information available in a MySQL database configured within the component. | Appears when Agentic MySQL DB is executed within a flow. XXXXX identifies the diagram or flow from which execution took place. | NLP Credits. |
nlp-agentic-mysqldb-na |
Corresponds to the same Agentic MySQL DB feature, but without an associated flow identifier. | May appear when testing the component from wolkvox Manager or during certain configuration actions, such as generating component information before using it in a flow. | NLP Credits. |
nlp-agentic-deepresearch-XXXXX |
Corresponds to the Deep Research component of wolkvox Studio. AI performs a multi-stage research process to search, analyze, summarize, and construct a response based on the information found. | Appears when Deep Research is executed within a flow. A single execution may require multiple queries to the AI model. XXXXX identifies the associated flow. | NLP Credits. |
nlp-agentic-deepresearch-na |
Corresponds to a Deep Research execution without an associated flow identifier. | Can be generated during component testing or when performing certain configuration actions before using it in a flow. | NLP Credits. |
nlp-mcp-XXXXX |
Corresponds to using an MCP component from wolkvox Studio. Allows an AI agent to use external tools or services connected via Model Context Protocol. | Appears when an agent needs to analyze which tool to use, prepare required information, execute an action via MCP, or process the obtained result. XXXXX identifies the associated flow. | NLP Credits. |
nlp-agentic-mcp-XXXXX |
Corresponds to processing the session memory of an MCP agent. | Appears when the session memory option is enabled in the MCP component and AI must summarize or preserve the context generated during execution. | NLP Credits. |
nlp-mcp-tool-call-metadata |
Corresponds to using AI to generate descriptive information for an MCP component, such as its name and description. | Generated when saving or configuring an MCP component in wolkvox Studio. This consumption is associated with flow design or configuration, not a final customer conversation. | NLP Credits. |
nlp-agentic-agentkey-XXXXX |
Corresponds to Agentic Key, wolkvox Studio's agentic orchestrator that interprets a goal and coordinates tools, integrations, or sub-agents to execute required tasks. | Appears when Agentic Key analyzes a request, decides which actions to execute, or delegates activities to other components. XXXXX identifies the associated flow. | NLP Credits. |
nlp-AgenticCampaing |
Corresponds to AI features used to query, manage, or make decisions about campaigns. | May appear when using campaign-related agentic features, such as Intelligent Dial or AI-driven campaign management tools in wolkvox Manager. | NLP Credits. |
conversations-chat-XXXXX |
Corresponds to the conversational AI agent handling written conversations, for example on WhatsApp, WebChat, or other digital channels. | Appears every time the Conversations component processes a customer message and uses an AI model to generate a response. XXXXX corresponds to the associated routing point. | NLP Credits. |
conversations-chat-XXXXX-cache |
Corresponds to a conversational chat agent response that could be reused from cache because an equivalent request was previously processed. | Appears when wolkvox finds a previously processed response that can be reused without performing full AI model processing again. | 10% of equivalent NLP credits. |
nlp-conversations-voice-XXXXX |
Corresponds to wolkvox Studio's voice conversational agent that understands customer speech and generates bot responses during a call. | Appears during different turns of an AI-handled voice conversation. XXXXX identifies the associated IVR or flow. | NLP Credits. |
conversations-voice-XXXXX |
Likewise corresponds to consumption by a voice conversational agent. It is an identifier used by certain executions or legacy versions of voice flows. | May appear in historical consumption or in flows that still record this identifier. Functionally corresponds to the same voice conversational processing type as nlp-conversations-voice-XXXXX. |
NLP Credits. |
nlp-model-router |
Corresponds to the automatic AI model selector. Analyzes customer requests to determine which model is most appropriate to respond. | Appears when a conversational agent has automatic model selection configured and wolkvox must decide which model to use before generating the primary response. | NLP Credits. |
nlp-getvars |
Corresponds to using AI to extract specific information from a conversation and store it in variables. | Appears when a flow uses variable extraction via AI, for example to identify data provided by the customer for subsequent use in the process. | NLP Credits. |
nlp-early-bridge |
Corresponds to a voice conversational agent feature that can generate a short hold phrase while processing what the customer said. | May appear during voice conversations when processing takes long enough for wolkvox to generate a brief response that avoids prolonged silence. | NLP Credits. |
nlp-barge-stop |
Corresponds to analysis determining whether the bot should stop its audio when the customer starts speaking while the response is still playing. | Appears in voice conversational agents that allow customer interruptions during response playback. | NLP Credits. |
nlp-chatbot |
Corresponds to the General NLP component used within chat flows or Agent Scripting. Allows sending a specific instruction to an AI model and using its response within the flow. | Appears when a General NLP component configured in a chat or scripting flow is executed. | NLP Credits. |
nlp-ivr |
Corresponds to the General NLP component used within a voice flow. | Appears when a Voice Flow uses an NLP component to send information to an AI model and subsequently use the result in the flow. | NLP Credits. |
nlp-webhook |
Corresponds to a General NLP query executed from a webhook. | Appears when an automation or external integration triggers an AI process configured in wolkvox Studio via webhook. | NLP Credits. |
nlp-interactions |
Corresponds to using General NLP within Interactions flows, such as processes related to emails or cases. | Appears when an Interaction Flow executes an NLP component to analyze, classify, or generate content using AI. | NLP Credits. |
nlp-wolkvox-studio |
Corresponds to a test of a General NLP component performed directly from wolkvox Studio. | Appears when someone uses the component testing option during flow design or configuration, without necessarily involving a customer interaction. | NLP Credits. |
nlp-wolkvox-inten |
Corresponds to wolkvox Studio's AI Intents component, responsible for identifying what the user wants or needs from their message. | Appears when a Voice Flow, Chat Flow, Interaction Flow, or Agentic Engine uses the intents component to classify received requests. | NLP Credits. |
nlp-classifications |
Corresponds to an AI text classifier that uses pre-configured examples to assign received content to a category. | Appears when a text classification component is executed within wolkvox Studio. | NLP Credits. |
nlp-voice-summary |
Corresponds to using AI to generate a call summary from its transcript. | Appears when a call is processed by Speech Analytics or Voice of Customer features requiring voice conversation analysis. | NLP Credits. |
nlp-voice-sentiment |
Corresponds to using AI to analyze sentiment expressed during a voice conversation. | May appear when processing calls via Speech Analytics or Voice of Customer when sentiment analysis is performed. | NLP Credits. |
nlp-text-summary |
Corresponds to using AI to generate a chat conversation summary. | Appears when a digital conversation is processed using Text Analytics or Voice of Customer features. | NLP Credits. |
nlp-text-sentiment |
Corresponds to using AI to analyze sentiment in a written conversation. | May appear when processing chats or other digital conversations using Text Analytics or Voice of Customer features. | NLP Credits. |
nlp-voc-general |
Corresponds to wolkvox Studio's Voice of Customer analysis component, which enables performing specific queries or analyses on conversation content using AI. | Appears when a flow executes a VOC analytics component to answer a question regarding a previously processed conversation. | NLP Credits. |
nlp-voc-autoqa |
Corresponds to Auto QA, which uses AI to automatically evaluate an interaction using a configured quality matrix. | Appears when a processed conversation is automatically evaluated using Quality with AI features. | NLP Credits. |
nlp-voc-analyzer |
Corresponds to wolkvox Manager's VOC Analyzer, used to generate executive analysis from gathered conversation data. | Appears when a user requests analysis generation using this feature from wolkvox Manager. | NLP Credits. |
nlp-voc-podcast |
Corresponds to AI processing used to generate Voice of Customer podcast content from conversation findings. | Generated during the periodic podcast-building process when the operation has sufficient analyzed data. Does not correspond to a per-call consumption. | NLP Credits. |
nlp-voc-best-actions |
Corresponds to using AI to build recommended best actions based on the analysis of successful and unsuccessful conversations. | Generated during the periodic process that builds recommendations, frequent scenarios, objections, and next actions to support subsequent customer service. | NLP Credits. |
nlp-vocrt |
Corresponds to Real-Time Voice of Customer. AI analyzes the conversation while the call is active to provide information and recommendations to the agent. | Appears while handling calls with Real-Time VOC enabled and generating analyses or suggestions during the conversation. | NLP Credits. |
nlp-customer-memory |
Corresponds to Customer Memory. AI analyzes previous interactions to build or update a customer profile with information such as context, preferences, discussed topics, and pending matters. | May be generated from conversations processed via Speech Analytics, Text Analytics, or voice/chat conversational agents. Does not correspond to session memory in wolkvox Studio agentic components. | NLP Credits. |
nlp-Copilot-default |
Corresponds to using wvx Copilot with the "default" knowledge domain. | Appears when an agent, user, or flow queries wvx Copilot using information stored in the default domain. | wvx Copilot fee, when applicable, plus usage NLP credits. |
nlp-Copilot-helpdesk |
Corresponds to using wvx Copilot with a specific knowledge domain, in this case a domain named "helpdesk". | Appears when a wvx Copilot query uses that domain. Other custom domains may appear following the format nlp-Copilot-domain_name. |
wvx Copilot fee, when applicable, plus usage NLP credits. |
nlp-Copilotchat |
Corresponds to using wvx Copilot from a chat interface, utilizing AI models and information retrieval from the knowledge base. | Appears when querying wvx Copilot from a written conversation experience. | wvx Copilot fee, when applicable, plus usage NLP credits. |
nlp-MrWizard-chat |
Corresponds to consumption generated by wvx Copilot when delivering a response in a chat channel. | Appears when an agent or automation queries wvx Copilot during a written conversation. | NLP Credits. |
nlp-MrWizard-voice |
Corresponds to consumption generated by wvx Copilot during a voice conversation. | Appears when an agent uses wvx Copilot assistance during a call to query information or receive a knowledge base-driven response. | NLP Credits. |
nlp-MrWizard-rtvoice |
Corresponds to using wvx Copilot within a call analyzed in real time. | May appear alongside nlp-vocrt when, in addition to Real-Time Voice of Customer analysis, a query is made to wvx Copilot's knowledge base. |
NLP Credits. |
nlp-rag-gate |
Corresponds to consumption associated with queries using AI models and RAG information retrieval to access knowledge used by wvx Copilot. | May appear during query processes where AI needs to retrieve information from a knowledge base before generating the response. | NLP Credits. |
nlp-QAi |
Corresponds to wolkvox Manager's QAi assistant, which allows asking questions via AI regarding operational information. | Appears when a user utilizes QAi in wolkvox Manager. A single question may require multiple AI executions to build the response and maintain query context. | NLP Credits. |
nlp-QAiDeepResearch |
Corresponds to the Deep Research feature of the QAi assistant in wolkvox Manager. | Appears when a user requests deep research from QAi and AI must analyze different parts of the information to build the result. | NLP Credits. |
nlp-ai-reports |
Corresponds to wolkvox Manager's AI Reports. Artificial Intelligence analyzes information and writes the requested result. | Appears when a user generates a report using AI. Generating a single report may require multiple AI calls to select appropriate processing, generate content, create titles, or maintain context. | NLP Credits. |
nlp-PHPFunction |
Corresponds to using AI to create a new PHP function from a user-written instruction. | Appears when using wolkvox Studio's AI PHP function generator. Corresponds to code creation, not the subsequent execution of the function within the flow. | NLP Credits. |
nlp-PHPFunction-Review |
Corresponds to using AI to review or modify an existing PHP function. | Appears when requesting AI to make changes to a previously created PHP function. | NLP Credits. |
nlp-PHPFunc-Meta |
Corresponds to using AI to generate descriptive information for a PHP function, such as its title, description, or structure. | May appear when generating or saving a PHP function using wolkvox Studio's AI tools. | NLP Credits. |
nlp-TurboSyntax |
Corresponds to TurboSyntax, an AI tool used to assist syntax and code creation or editing within wolkvox Manager or wolkvox Studio. | Appears when AI is used to generate or modify elements like code, forms, or structures used during solution design. | NLP Credits. |
nlp-TurboText |
Corresponds to the AI tool used to rewrite or transform text according to an instruction. | Appears when a user uses AI text editing features available in wolkvox Manager. | NLP Credits. |
wolkvox-studio-AIFomr |
Corresponds to using Artificial Intelligence to create or edit forms used from wolkvox Studio's scripting tools. | Appears during form generation or editing using AI. This consumption is associated with solution design, not necessarily a customer interaction. | NLP Credits. |
wolkvox-studio-DocumentGenerator |
Corresponds to the wolkvox Studio component that uses AI to generate PDF documents. | Appears when a flow uses the AI document generation component. Using this feature requires wolkvox Storage to be enabled. | NLP Credits. |
nlp-email-fix |
Corresponds to using AI to analyze an email address that appears invalid and propose corrected alternatives. | Appears when a flow validates an email and needs to use AI to suggest potential corrections. It is not generated for all valid emails. | NLP Credits. |
nlp-cache |
Indicates that wolkvox was able to reuse a previously processed AI response instead of re-querying the model entirely. | Appears when a request matches a cached response that can be reused. Although full processing is not performed again, cache usage does generate consumption. | 10% of equivalent NLP credits. |
How to interpret TYPE_NLP suffixes?
Some values function as a prefix to which wolkvox automatically adds an identifier. This identifier links consumption to the flow where it originated.
For example:
nlp-conversations-voice-61267
In this case:
-
nlp-conversations-voice-identifies the feature type that generated the consumption. -
61267identifies the associated flow, IVR, or routing point, as applicable.
Identifiers are part of the same TYPE_NLP shown in the report, so different flows may appear as separate rows even if they use the same component type.
-na suffix
The -na suffix indicates that the execution was not associated with a specific flow identifier. This can occur, for example, during a test conducted from Manager or when executing certain processes related to configuration or metadata generation.
-cache suffix
A value ending in -cache indicates that the response was served via NLP cache.
For example:
conversations-chat-61128-cache
In these cases, previously processed information is used, and consumption corresponds to the reduced cache scheme.
Considerations for interpreting consumption
- A single operation may show multiple TYPE_NLP values during the same period. Each corresponds to a different feature or process.
- NLP_AMOUNT represents the number of executions and should not be interpreted as credit volume or monetary value.
- To identify types with higher consumption, primarily check the TOTAL_CREDITS column.
- Two types containing the same numerical identifier do not necessarily represent the same feature. For example, an Agentic DB component and a voice conversational agent may be associated with the same flow, but they generate independent consumption.
-
nlp-customer-memorycorresponds to the processing used to build or update customer historical information. It should not be interpreted solely as enabling the Memory tab in wolkvox Agent. - Session memory used by certain wolkvox Studio agents may generate specific records associated with the agent type and the flow where it executes.
Report reading example
If the report displays a row with the following value:
conversations-chat-61074
it means that during the queried period, consumption was generated by a conversational agent used in a chat flow associated with routing point 61074.
The NLP_AMOUNT column will indicate how many executions were recorded for that type, while CREDITS_SEND and CREDITS_ANSWER will show the credits used in requests and responses, respectively.
Finally, TOTAL_CREDITS will show the total credits associated with that feature during the selected date range.
In this way, the "8. NLP grouped by type" report allows linking Artificial Intelligence consumption to the features that generated it and facilitates reviewing NLP credit usage within the operation.