We're seeing some strange activity when our agents are presented with a call. When they click on answer the call does not connect and they go into a not accepting interaction state.

We're looking at network, local browser settings, clearing the cache etc, ZScaler etc.. but we cannot seem to find a solution. We're seeing this across fidderent agent persona - office based, home worker, vpn, cannot find a pattern.

Anyone experienced this and have any suggestions ?

thanks,

Hi Paul,

Are you on On-premise or AWS edges?

If On-Premise and agents are on same site or corp network, ensure workstations have reachability to edges WAN network interface.

This ensures webRTC voice routes thru LAN (via HOST candidates) instead of open internet.

CIDR ranges must be whitelisted on outbound traffic for both Edges and workstations too as seen here:
Cloud media services CIDR IP address range - Genesys Cloud Resource Center (mypurecloud.com)

If your using AWS virtual edges, ensure the sites where the webRTC phones are assigned to are using the correct media regions.

Base it on where the workstations are, not your org region e.g. agents sitting in Asia on a us-west-1 org should have asia pacific media regions.

You can also enable persistent connection and auto-answer for these agents. Disable any packet sniffing on purecloud traffic.

For browser settings, ensure cookies and microphone permissions are enabled.

Avoid use of wireless headsets that idles when not in use as sometimes, the browser cant detect microphone as the call is routed.