18 Nov QA Workflow Planning with Sample Workflows
QA is, in part, meant to align customer facing service with company goals. An effective Quality Assurance workflow reflects those goals with respect to company size and resources. Questions to consider when developing a QA workflow includes:
- What is your company’s main focus? What is your company trying to improve?
- How can QA help identify areas of improvement?
- With the resources available, what is the best breakdown of how an inquiry escalates?
- Can an AI tool streamline your process making agents more autonomous, allowing supervisors and managers to maximize their effectiveness?
Companies vary in size, resources, maturity, mission statements and goals. All of these inform a workflow. Let’s start with focus. Identify what QA Metrics your company is looking to measure and improve. Examples:
- Having the smallest abandon rate in your sector
- Improve CSAT or NPS as compared to your industry
- Improve customer loyalty
Company size and resources inform the levels available in a QA workflow to escalate and monitor QA. Interactions are monitored by one or more of the following:
- Team Leads
- Supervisors, Managers or both
- QA Specialists (smaller companies can’t often afford this)
- Tools driven by AI and informed by customer behavior
Though it is best to have a dedicated QA person or department that can focus solely on this, not all companies will have the funds to do so. In those cases choosing detail oriented individuals and those who can balance QA with their other duties is beneficial.
Having more than one supervisor or manager review interactions offers more perspectives. This is important when agents disagree with scoring to avoid a harsh divide between agents and management. When agents are heard and their work is valued agent churn is reduced. This in turn saves on the costs of training as skilled agents are costly to replace, especially in this post pandemic market.
QA Workflow Sample with a Small Four Level In-House Team
In this workflow example there is a four level system of Agent, Team Lead, Supervisor and Manager. The company also has a knowledge database for agents to consult.
The Team Lead is often a seasoned agent with a deep grasp of company policies and knowledge of its customers.
The Supervisor(s) in this case are the level between agent and manager. They help to convey company policies to agents and elucidate customer service challenges to managers who may rarely engage these challenges directly. Supervisors in this hierarchy are also able to advocate for agents when they disagree with scoring.
When to Escalate
Identifying triggers for ticket escalation from one agent to the next available level is key to a smooth workflow. If a company has a knowledge base available to all agents it acts as a resource to reduce the need for escalation. This enables supervisors and managers to spend more time in review, improvement of processes and training. This also empowers agents, offering them more opportunity for higher scores.
In companies with multiple departments, allowing agents to communicate with other divisions, or their knowledge bases, also improves agent autonomy.
Train agents to listen carefully to the customer and ask clarifying questions. This is the easiest way to reduce escalation.
One area for immediate escalation or intervention is abusive language or legal threats. Verbal abuse, whether from agent to customer or customer to agent, is damaging. Intervening quickly improves the odds of redirecting the interaction to a positive outcome. Create a company culture of employee engagement supporting agents during very difficult interactions for lasting benefits.
Outsourcing QA to an AI
Whether your company is large or small, AI tools offer cost effective methods of reducing the effort of QA. The’re customizable to address your company QA focus, size, structure and industry.
These tools can assist with scaling customer service and QA by supporting agents in being more autonomous. Fewer escalations improve agent scores.
One key to this is real time scoring. When an agent can see their scores move with every response they offer, they are being trained in the moment to provide better CX. Real time scores are also a motivator for agents to excel. When an agent has metrics for empathy, excellent service and whether or not a customer query has been answered available to them during an interaction, they are able to modify their responses in the moment. This reduces the number of customer touch points needed, improving FCR.
Companies that do not have a dedicated QA specialist or team benefit from fewer escalations in more than just scores. Reducing strain on the supervisors and/or managers that perform QA oversight along with their other duties allows for a more comprehensive quality assurance picture.
Below is an example of a QA workflow with an AI tool in place.
QA Workflow sample with an AI assist using real-time scoring
In this illustration, the AI is continually scoring the interaction as it progresses. The agent is aiming for as close to five stars as possible. If any portion of the score is low and the AI has not indicated resolution, the agent is expected to try to improve the interaction.
Once the agent has exhausted the review, additional questions, other departments and alternative solutions, only then will the call be escalated.
One caveat to this is abusive language. In particular, abusive or threatening language by the customer will automatically cause the AI to notify a manager.
5 Stars is Not Always Achievable
It is important to understand that not all interactions are able to be five star level. There are some questions that cannot be answered, but customers still ask them. An example: “Why didn’t my sister like your product?” There is no way for an agent or anyone from the company to answer such a question.
There are also some situations where policy and customer expectation will conflict. The customer will be disappointed, even angry. An example: “We cannot replace your item as it is out of warranty.”
Deciding at what point your company will bend policy and how much latitude each member of the hierarchy has will inform your workflow escalations. Take these situations into account when reviewing agent scores and evolving company policies.
Below is a workflow of a small QA team of only Agent and Supervisor levels. A knowledge base is particularly valuable in such small teams. Effective communication between these levels is very important. Clear policies must inform the workflow. Define how many times an agent can flag a score. Be clear on the resources an agent is expected to consult prior to escalation.
With advanced planning a QA workflow is a vital tool to streamline processes and improve metrics.
For More on QA Check out these posts: