Genesys Cloud - Main

Hi everyone,

We've been using Genesys Cloud for ~2 years, we only us it for calls but we are still struggling to build a consistent, row-level call dataset that accurately reports call volumes and handle times across inbound, outbound, queues, ring groups, and campaigns.

What we're trying to achieve is an old school single row per call (or per agent-handled call) that includes: Call Start Time, Call Connect Time, Call WrapTime, Call End Time, Direction (Inbound / Outbound), Queue / Ring Group, Agent who made it or it was offered to and the Wrap Outcome. Essentially, a single dataset that can reliably answer: "How many calls did we receive / make?", "How long did they take?", "How were they wrapped?".

What we are currently doing is pulling data from the Conversation Details API, flattening data across Conversations / Participants / Sessions / Session Segments / Session Metrics. While we can see all the raw data, we are struggling to identify what represents the "actual call" with a single reliable rule

Challenges we keep running into are that a single conversation can have multiple participants and sessions resulting wrap outcomes appear in different places depending on call types and outcomes. If internal support calls are made on the call we cannot break these calls out separately but double count the AHT.

What is the recommended grain for call reporting from the Conversation Details API? Are there Genesys-recommendations or rules to achieve what we're after so that we do not have multiple data sets each tailored for a single question. Are we overcomplicating this, or is this a known limitation of the raw API data?

Any guidance, examples, or confirmation of best practice would be greatly appreciated. Every time we feel like we're close a new type of call comes up and we find we're missing a key concept or rule that ties everything together.

Thanks in advance!

#Reporting/Analytics

------------------------------
Sam Merrey
------------------------------

Hi Sam,

Its an interesting question and I'm not sure that overcomplicating is the correct term or that its a limitation of the API.  The conversation details API is capturing all the activity related to a conversation from an analytics perspective.

It is possible to exclude legs(e.g. consult) or participants (the consulted agent), but creating a rule that will do it for each scenario could be tricky.  It can be better to adapt to the full conversation details and include all the details for a conversation than to exclude data. e.g. if an agent consults another agent, although it will double count this time, it reality that is the time spent by agents dealing with the the call, so excluding that time or part of that time would potentially misrepresent the actual time spent to resolve the customer query.

I am interested to see others in the communities view on this and if they have a solution that would work for you

------------------------------
Sam Jillard
Online Community Manager/Moderator
Genesys - Employees
------------------------------

Original Message:
Sent: 12-29-2025 11:48
From: Sam Merrey
Subject: How Do You Build a Reliable Row-Level Call Dataset (Inbound & Outbound)

Hi everyone,

We've been using Genesys Cloud for ~2 years, we only us it for calls but we are still struggling to build a consistent, row-level call dataset that accurately reports call volumes and handle times across inbound, outbound, queues, ring groups, and campaigns.

What we're trying to achieve is an old school single row per call (or per agent-handled call) that includes: Call Start Time, Call Connect Time, Call WrapTime, Call End Time, Direction (Inbound / Outbound), Queue / Ring Group, Agent who made it or it was offered to and the Wrap Outcome. Essentially, a single dataset that can reliably answer: "How many calls did we receive / make?", "How long did they take?", "How were they wrapped?".

What we are currently doing is pulling data from the Conversation Details API, flattening data across Conversations / Participants / Sessions / Session Segments / Session Metrics. While we can see all the raw data, we are struggling to identify what represents the "actual call" with a single reliable rule

Challenges we keep running into are that a single conversation can have multiple participants and sessions resulting wrap outcomes appear in different places depending on call types and outcomes. If internal support calls are made on the call we cannot break these calls out separately but double count the AHT.

What is the recommended grain for call reporting from the Conversation Details API? Are there Genesys-recommendations or rules to achieve what we're after so that we do not have multiple data sets each tailored for a single question. Are we overcomplicating this, or is this a known limitation of the raw API data?

Any guidance, examples, or confirmation of best practice would be greatly appreciated. Every time we feel like we're close a new type of call comes up and we find we're missing a key concept or rule that ties everything together.

Thanks in advance!

#Reporting/Analytics

------------------------------
Sam Merrey
------------------------------