AT&T has always worked to provide excellent customer service to ensure the easiest and simplest solutions when challenges arise for its customers. The front-line support acts as the first support gate dealing with multiple lines of business each requiring their knowledge to navigate and decide if the next support gate was needed.
Over years of AT&T's business growth, the services expanded which in time created large workflows for agents to maneavor. These workflows were made by the Business Line Managers who then sent these new or modified workflows to the development team to update the workflow engine with the new process.
After 7 years of this cycle, the workflow engine became cluttered and hard to navigate, no history logs were kept, not one developer could tell you a process from a to b, and the agents who were the front line support were catching errors and glitches for loops of workflows being run together. This had become a mess of spaghetti and not the good kind.
Many agents had accepted that there would be no tool to support their needs, and it was with this, that I shared my goal to help make their job more efficient and enjoyable. Agents through many of the calls I listened to, were working in an ad-hoc style, finding solutions based on their previous knowledge and experience and asking others if they have the answers. Ultimately this sounds like great employees, and yes they were, but the challenge was how this delayed the call handling time. Agents were opening multiple browsers to find the answers they needed, jumping constantly to get what they needed.
This engine navigated similarly to a nested visio document, with embedded processes you click into to find the next process.
Normally, this is acceptable for a linear view of the flow, however, I found flows circling back to other flows tied into dead ends
and old services that were no longer active, but still in the tool and accessible. Given the number of workflows and the
overlapping processes, it became easier to enter the dry-erase room with sticky notes and map the whole thing out.
I learned many workflows were used multiple times in different business lines instead of using a single workflow passed through.
After going through them I was able to reduce the workflow count by 66%.
Agents had to use multiple tools to solve their cases, meaning this new design should incorporate each tool that is accessible in a single pane of sight instead of swapping screens actively. Six ideas were created and through discussions with the managers and technical leaders, we came to a fitting solution to the business goals.
With the final design, the prototype was built using Just In Mind. This allowed for the use of dummy data or real data if we decided to. Agents were excited about the new design, from two different test groups. One group was given no demonstration on where their tools were and the other group was given a demonstration. While the group without demonstration took about double the time to get acquainted with the new tool, they were able to find everything as they clicked each section.
Since this became a custom-built solution, a design library was pertinent for the future success of the new platform. Every component and instruction guide was added to the company Wiki for guidance.
Workflow engines are a beast to deal with if they are not managed and historical notes are kept. While they can produce challenges to fix, they can be done and ultimately this particular project reduced AHT by 4 minutes per call, and development of new workflows by half a day. While my priority for this role was to redesign the new platform, I was invited to help develop it as I have strong knowledge of HTML, CSS, JavaScript, JQuery, PHP, and SQL. This also allowed me to support the back-end team in coordinating where data should be connected to the new design and creating a decent product development experience.