Atlassian Implementation for Health Insurance Startup

Employer: Adaptavist
Role: Senior Consultant
Duration: July 2015 to July 2016

In July of 2015, I was invited to spend four days with Collective Health (CH) at their office in San Mateo, California, to talk about knowledge management with Confluence. They had just closed a successful round of funding and over the next 6 months would need to scale up from 100 members to 30,000.

Back in 2015, Collective Health’s key performance indicator was customer satisfaction. Health insurance customers are typically unhappy with the service they receive, and CH was shooting for a 70% customer satisfaction rating. With the help of Adaptavist, they surpassed 95% and provided us a case study: How Adaptavist helped Collective Health empower people and processes.

I have two goals when I start a discovery engagement. The first is to understand what the business really needs, not just what they asked for. The second is to deliver a blueprint so the customer can execute on my recommendations themselves, or even hire someone else, if they prefer. Of course, I’d love for them to hire me, and I will be able to deliver on the work more quickly and efficiently than someone else, but I don’t want to lock someone in to needing me; I always want to work my way out of a job.

At the end of those initial four days, I delivered my blueprint. This document shared my findings, recommendations, and the initial steps for building the solutions that Collective Health needed. In presenting it to CH’s leadership, I also talked about other opportunities to help them scale up and improve how their people and teams worked together. Consequently, CH extended our contract and I began flying out to San Mateo every other week to help them launch.

Over the next six months, I learned a lot about health insurance plans and regulations. We built and delivered three key systems.

Knowledge Management Systems

Collective Health selected Atlassian’s Confluence as their knowledge management system and intended to store all details of the various insurance plans on Confluence. The goal was to enter the data once, then use it in multiple places.

I built a system of templates and pages using excerpts and macros to create plan pages and glossaries. Working with CH’s teams, I structured the data and created many of their root pages with a consistent excerpt and macro naming scheme so that people could easily build plan pages. These plan pages pulled in data from many other source pages.

The outcome was many pages that could be linked to from other systems. When accessing one of these pages, all the plan information was presented consistently and quickly. And when something changed, the data only had to be edited in one place–one of the root pages–and that change would be visible on every other page referencing the data.

Customer Service Integration

With the plan information in place, the next step was to integrate it with CH’s telephone and customer relations management systems. The goal was, when a user calls in, the person answering the phone should know everything about that customer including details on their health insurance (stored in a custom in-house built software), previous calls (logged in Jira), and plan information (compiled in Confluence).

I helped scope this system and managed an Adaptavist engineer who built the solution. We delivered it early and under-budget.

HIPAA-Compliant Bug Tracking

The last major software deliverable was to manage bugs in CH’s customer portal and mobile app. Because bug reports might include or link to HIPAA-protected information, it was important to manage access carefully while still letting engineers fix the problems.

We used Adaptavist’s ScriptRunner to build a complex permission automation system. In brief, it let a customer service agent create a bug ticket and then link that ticket to other relevant tickets in Jira. Only the agent could create these links, meaning they managed the initial access to HIPAA information. The bug ticket would be assigned to an engineer, and only that assignee would be able to access the protected information through permissions granted via link traversal. Engineers couldn’t reassign issues on their own, a team lead had to do that. Anyone could delete links, and upon deletion, permissions would be revoked automatically.

This meant no one had to mess with permissions directly, just create a ticket and a link to grant and delete a link to revoke. It was an elegant solution for the users, but quite complex in terms of building the automations, and took about six weeks to deliver on its own.

Business Consulting

Throughout the year I was engaged with Collective Health, I also performed a significant amount of business consulting. One of my specialties is inter-team communication, and I spent time talking with people across different teams to better understand how they needed to interact and then coach them on communication approaches. I also handled the configuration of their Jira projects and trained people on how best to use Jira to communicate and collaborate.

Create a website or blog at