11 reasons why your Quality Assurance (QA) process is letting you down and how to use technology to turn it around

Written on
byHayfa Bukhari

Our latest whitepaper offers a deep dive of how you can improve your QA process, and how technology can be used to ensure that the voice of every customer is heard by your organisation. Download it here today.

Quality Assurance (QA) often has a bad reputation with  many organisations seeing  it as a costly box ticking exercise,rarely understanding  the true value of QA. It’s ironic that the very process that is meant to ensure quality within a business, is in itself, a flawed one which can result in a wealth of insight falling between the cracks. In this post, consider the ways QA processes fall short and explain how advances in technologies such as Artificial Intelligent (AI), Machine Learning (ML) and Natural Language Processing (NLP) can unlock the true value of QA.

We’ll argue how this, often clunky process can be turned into a data goldmine, forming the central nervous system of your business, proffering insights and automation to all areas of your organisation. 

Why are QA processes so inefficient? 

At a high level, the QA practices that go on today are geared very much towards an internal criteria of quality which are assumed to be directly related to a customer’s idea of quality.  This isn’t always the case because QA practices are weighted heavily towards improving efficiency and being compliant to whatever regulatory body governs your industry, be it the Financial Conduct Authority (FCA) or Ofgem for example.  These legacy practices were designed at a time when customer retention and experience were often sidelined for cost-savings and cuts.


A good example of this might be an over reliance on a QA scorecard in call centres.  Solely relying on making sure you’re scoring well on one metric might tick a QA box but in reality, it often ends up leading to poorer outcomes for customers.  QA scorecards should be used as a minimum requirement for call quality and staff should be trained and coached to ensure they achieve well above that bare minimum standard.      


Here are 11 reasons  why your QA processes are failing you:


1. Assessing different facets of a call that impact different areas of a business


As part of the QA process, assessors and customer service representatives are asked to gather huge quantities of data such as whether someone is vulnerable; if there has been a change in their financial situation; whether they voiced a complaint, all the way up to collection of affordability information such as how much they spend on groceries, petrol or their mortgage payments each month and so on.


It’s a challenge for even the most skilled assessor to untangle the sheer volume of information, determine whether the agent adhered to a specific regulatory script and if all the required information has been gathered, not to mention sorting this data into areas that could benefit different parts of the business.  For example, it would be useful for the product development team to receive feedback about specific products and services, regardless if it’s positive or negative.  Likewise, information around whether a particular communications channel or process was frustrating for a customer would be valuable for the customer experience team to investigate.


2. Non-representative sampling


It is often the case that the sample of calls QA teams analyse aren’t representative of what’s actually going on across your call centre.  For example, an assessor  may randomly take 10 calls to analyse.  There may be no issues found with any of those 10 calls and they could tick the QA box and conclude, on the whole, customers are happy and no complaints have been made.  We all know this is unlikely to be the case and that out of the thousands of calls firms take, there will be a proportion of those that hold complaints, missed flagging of vulnerability or risk or even issues with agent conduct and lack of adherence to regulated scripts. 


3. Assessment is not risk targeted


Before they analyse a call, the assessor doesn’t know the profile of that call so it’s not risk targeted.  In other words, agents don’t have the insight upfront to know whether it’s script adherence or identification and verification issues etc they should be looking out for.  This makes assessment really inefficient, and to the point above, they might not even find anything in the call making its review an inefficient use of time.


4. Free format reporting


Assessors often have to record their findings from a call in a free format using a rough template and it’s usually at a very high level which is harder to do data analysis on.  This kind of data capture offers companies little value other than ticking the QA box.  The insights and information gathered from calls are too high level to do anything meaningful with and they’re not in a format that’s easy for other areas of the business to draw actionable outcomes from.   Once the assessor has taken a quote out of context to be used somewhere else, it loses some of its value.  It’s inherently difficult to capture the essence of a conversation succinctly. 

5. Manual reporting 

QA agents have to fill out forms and capture data the old fashioned, hand-written way.  There are several problems with this.  Firstly, it’s time-consuming for the assessor, not just to write things by hand but for someone to take that information and type it into other systems. Secondly there’s more margin for error as someone might not be able to read the handwriting or it might be inputted into the system incorrectly. Thirdly, the cost of paper processes is not just high, it has an environmental impact too.


6. Whole call assessment 

QA agents have to listen to the whole call to assess it which can last anywhere between 10 and 90 minutes.   Not only do they have to pay keen attention to what’s going on in a lengthy call but the process is made even longer by the fact that assessors have to pause the call every time they need to then type their findings.  Some calls don’t even need to be assessed.  An agent will often spend 20 minutes listening to a call only to realise at the end of it that it’s not a high risk call and there’s nothing to report.  And that’s valuable time wasted.  


7. Siloed information that impacts learning and development (L&D)

In many QA processes today,  learnings aren’t shared with customer facing representatives to assist L&D.  For example, an assessor might write a report and highlight that the call was poor but nothing is fed back to the team or individual agent as to why and how to improve should they come up against this situation in the future.  Valuable learning opportunities are lost.


QA assessors can also face the same challenges as the advisers.  For example, they need to know the business process for every situation an agent might encounter to check that the agent has handled it correctly. 


Using speech analytics and automation, learnings from calls can be shared with teams of agents as well as individuals for a more effective coaching experience.


8. Outsourced contact centre model

It’s not uncommon for large companies to have 50-60 different contact centres spread out across the globe.  These de-centralised, wide-spread teams means there’s often a lack of consistency with customer experience.  Not only that, in terms of efficiency, if your company is mandated to QA at least 10% of those calls, the process suddenly becomes logistically difficult, costly and a huge time drain on the assessment teams.


9. High staff turnover 

Being a customer service representative can often be very challenging.  Agents are asked to deal with a wide range of very sensitive, complex calls yet training for these types of calls can be quite narrow.  Also, it can sometimes feel like quite a negative experience for staff.  They might feel like they can’t ‘win’ unless they adhere strictly to a script which isn’t a good experience for either the agent or the customer.  This challenging environment plus low morale has led many call centres to have a high turnover in staff, meaning not only increased costs of recruitment and training but also a lack of consistency in customer experience and quality.     


10. Physical document checks

In the advice sector, a lot of QA is based on physical document checks and not, listening to calls.  This means that the assessor is assuming the advisor has accurately captured what the customer said which may not be the case, resulting in inaccurate assessment of quality.  


11. Multiple sources of quality assurance

A major inefficiency in the QA process for some organisations is that the sources for QA can come from multiple places.  For example, the sales team may compile a list of ‘sold’ cases, the pre-sales team might create a list from new advisers, plus a list of cases where risky products might have been sold.  They are then manually ranked high, medium or low risk.  Finally, the QA team only reviews the high risk calls, meaning the majority of these compiled lists are actually ignored. 


The benefit of implementing digital automation in your QA process 

Automation has such huge potential in regulated industries freeing up time for agents, offering a more accurate, consistent means of gathering and processing information and importantly, reducing costs. Some areas of automation companies are looking at include:


Robotic Process Automation (RPA) 

RPA is a software that mimics human interaction with systems, allowing firms to automate the audit of financial statements for example or speed up customer onboarding through automated verification. 


Business Process Management (BPM)

These are models that manage end-to-end business processes which allows for a more integrated and seamless experience for customers when applied to omni-channel interactions for exemple. 


Artificial Intelligence (AI) 

This technology is about to automate more complex, statistical and machine learning activities. 


Product and Sales 

NLP can help to automatically collect information around what customers are saying about certain products and services to improve and develop new ones.

NLP can help automatically capture feedback around the sales experience and products and services to better tailor sales approaches to customers in the future.


Customer Experience

Better understand customers and use it to focus the CX team’s efforts on parts of the experience that need most improvement.


How you can take action 

There are many good reasons why, if companies implement speech automation technologies, they could significantly enhance their QA process, improve their customer experience as well as their agents’ performance. But that’s not all AI and NLP have to offer. The insights and automation that can be driven by the voice of the customer have the potential to impact almost every part of your business

To find out more how technology can improve Quality Assurance, read the full whitepaper now 

Find us on LinkedIn and Twitter to learn more…

Other recent posts

AI, Machine learning, Hands of robot and human touching on big d

A-Z Journey of Adopting Aveni Detect

Like us, you’ve probably noticed that generative AI is causing a productivity revolution. In order for this to be successful, your business needs to adopt domain specific solutions built for your industry, your challenges,...

Analyse Customer calls

Optimising Customer Interactions:  How Aveni Detect Analyses Customer Calls

We’re living in a world of tightening  regulations and ever-changing business environments, where understanding and enhancing customer interactions has taken centre stage. If you analyse customer calls, you have an opportunity to deepen relationships,...

Customer review satisfaction

3 ways generative AI could boost your financial advisors today

It wasn’t so long ago that the thought of receiving financial–or any–advice from a machine would feel like something from a film, and likely heavily caveated with warnings about the robot uprising (we’re looking...

Robot is learning flat illustration Ai Generate

Accurate AI for Financial Services: Aveni's AI Models Lead the Way

Financial services institutions operating in today’s regulatory landscape face a myriad of challenges in ensuring compliance and quality assurance in their operations. To tackle these hurdles, Aveni has been at the forefront of developing...

Consumer Duty compliance

 Brokers James Sharp implement Aveni’s AI technology to meet Consumer Duty vulnerability monitoring needs

The broker and investment firm James Sharp has deployed  technology from Aveni.ai to ensure Consumer Duty compliance. It has moved quickly to adopt Aveni Detect, the AI and Natural Language Processing (NLP)-based technology platform...


 Aveni commits as member of Consumer Duty Alliance

We are excited to announce our membership of the Consumer Duty Alliance. Since our inception in 2018, we have been developing solutions utilising AI technology and incorporating the latest advances in Natural Language Processing...

AI Assistant

Aveni Assist

Aveni Assist, the GPT-powered AI assistant has been designed specifically for the financial advice market to transform administrative tasks, allowing advisers to focus on supporting a significantly greater number of consumers. This new product...

AI assistant for financial advisers

Aveni launches Aveni Assist to transform productivity in Financial Advice market 

Today, we announced the release of a new platform: Aveni Assist. The GPT-powered AI assistant has been designed specifically for the financial advice market to transform administrative tasks, allowing advisers to focus on supporting a...

AI Emergence

AI'll Grow Arms & Legs: Emergence