HCF chases near real-time intelligence on members' digital experience

Health insurer HCF intends to use a suite of monitoring tools to finetune its digital experience, with the aim of being alerted “the minute a member’s having a bad experience”.

HCF's Anton Eksteen.

Head of networks and reliability engineering Anton Eksteen told Cisco Live in Melbourne that the insurer has spent the past 18 months upgrading and modernising its network and observability stack.

It has deployed a Cisco stack of technology for both.

On the observability front, it is using Splunk, AppDynamics and ThousandEyes, with all data brought together in Splunk Cloud Platform.

“We've uplifted monitoring mainly around the infrastructure so far and a lot of that data is coming into Splunk,” Eksteen said.

“Instead of having to log onto three or four different systems to figure out what's going on, all the data is there [in Splunk] and the data is democratised.

“You can see that data in Splunk and you can then start correlating data so it's a lot less time spent on finding out what's going on and finding someone who's got access to the data to figure out what's going on”.

Given the focus on improving member experience, HCF has turned the observability stack toward monitoring core member-facing systems such as the one used to lodge claims.

“We can see the transactions coming in. We know how many transactions happen at any time, down to the minute. We can see how many of those claims are approved or declined, and how many of those claims error out or time out,” Eksteen said.

“It's a very sensitive system. Claims is a very important business process for us. We need to know when that system goes down. 

“So whereas before, it took us a couple hours to figure it out, we can now figure things out within minutes.”

Where an error pattern has been previously observed, the insurer has been able to automate some responses, such as to restart a system or server.

“For some of the problems, we know what the fix is. We've been able to use Splunk to automate it,” Eksteen said.

“It fixes the problem within two minutes instead of waiting 20 minutes or half an hour to get [engineers] on a call to restart it anyway”.

The observability stack is also used by engineering teams to see the user-facing impact of any backend changes they make.

“For instance, if our database team does some tuning on the databases, they can immediately see the impact – and hopefully it's a positive impact – on the transactions,” Eksteen said.

“They can see there' are less transactions erroring out or timing out, and they can see the transaction time, if it's increasing or decreasing, so you can immediately see that impact. Whenever they make a change in the evening, the next day [they] will know what the result is – we don't have to wait for that”.

The ultimate goal that HCF is working towards is to have near real-time alerting to any degradation in digital experience for members.

“We are working very closely with our digital experience team, and they are very focused – like everybody at HCF  – on the member experience,” Eksteen said.

“We want to monitor every member interaction [and] business process, make sure we understand it, so we can pinpoint issues early on. 

“It's not good when a member’s experience is bad and then they don't say anything, but eventually the calls come into the call centre, and then the call centre phones the service desk, and then it gets to us. By that time you've lost two [or] three hours". 

“So, we want to know down to the minute when the member's having a bad experience and we'll be focusing on that over the next six to 12 months”.