As per yesterday or today, when the person hasn't spoken, it seems the analysis is not run (no summary, success evaluation nor structured data). It's a change of behavior, as it did send analysis results on this type of call previously. Behavior change also detected by another vapi user. Is it a bug ? Is it on purpose ? For example, assistant id cd37ff71-f729-4466-bec9-77dbe95daeb7 has calls today: - analysis on call id b2d5d958-6107-48a4-b66c-86e85abfab77 (2:34 pm) - no analysis on call id 2d17d778-163d-4c4a-b870-ce89c5d880ba (2:19 pm) - analysis on call id 6b81c052-9b8d-476e-9ddb-834b0700838e (1:45 pm) Similar for other assistants. No analysis if only assistant has spoken. https://cdn.discordapp.com/attachments/1330927184486993993/1330927185032249385/image.png?ex=678fc208&is=678e7088&hm=1d984684d2587a443c4b572372a7c9cc4e792dbef6530fb16a085b41ba2c5f27&

@Jeebs from eva.be I understand your concern regarding the change in behavior where analysis results are not being generated for calls where only the assistant has spoken. This could be due to a recent update or an unintended issue. To assist you further, I would need to check the system logs and configurations for the mentioned assistant ID and call IDs. However, as a support bot, I don't have direct access to the system or user data. Please reach out to our support team with the assistant ID and the specific call IDs you mentioned. They will be able to investigate this matter and provide you with a detailed explanation. If it turns out to be a bug, they will work on a fix, and if it's an intentional change, they will clarify the reasoning behind it. Would you like me to guide you on how to contact support or is there anything else I can assist you with?

Thanks @Shubham Bajaj for your msg yesterday during vapi-stage. Let me know if you need any more info

@Jeebs from eva.be Call Id: b2d5d958-6107-48a4-b66c-86e85abfab77 "analysisCostBreakdown": { "summary": 0.0024, "structuredData": 0.0051, "successEvaluation": 0.0027, "summaryPromptTokens": 328, "summaryCompletionTokens": 94, "structuredDataPromptTokens": 839, "successEvaluationPromptTokens": 878, "structuredDataCompletionTokens": 169, "successEvaluationCompletionTokens": 4 } Call Id: 2d17d778-163d-4c4a-b870-ce89c5d880ba "analysisCostBreakdown": { "summary": 0, "structuredData": 0, "successEvaluation": 0, "summaryPromptTokens": 0, "summaryCompletionTokens": 0, "structuredDataPromptTokens": 0, "successEvaluationPromptTokens": 0, "structuredDataCompletionTokens": 0, "successEvaluationCompletionTokens": 0 } Call Id: 6b81c052-9b8d-476e-9ddb-834b0700838e "analysisCostBreakdown": { "summary": 0.0037, "structuredData": 0.0057, "successEvaluation": 0.0032, "summaryPromptTokens": 484, "summaryCompletionTokens": 153, "structuredDataPromptTokens": 995, "successEvaluationPromptTokens": 1034, "structuredDataCompletionTokens": 183, "successEvaluationCompletionTokens": 4 } @Jeebs from eva.be can you check again only for one call where there is no user input the analysis plan was not generated else it has worked for other calls.

Hello @Shubham Bajaj ! That must indeed be the reason why many calls don't have an analysis anymore. Basically, my assistant is an intelligent voicemail - but many people drop the call when they get to the voicemail (prefer to call back). In that scenario, only the first message has been spoken, so messages <= 2. It had quite an impact as I relied on the analysis to tell me that no meaningful interaction occurred 😓

No, it is added to ensure unnecessary generations don't happen. We will try to add a force parameter which runs regardless of any conditions.