ChatGPT and AI have both become buzzwords in the contact center industry, with many advancements enabling automation in various areas. The recent buzz around ChatGPT has many businesses excited about the prospect of increasing their quality assurance coverage (and for good reason) however we wanted to share a dose of truth having tested ChatGPT for the last few months. After our testing, we believe that AI is not a replacement for human oversight, your QA team are still more valuable than ever.
It’s crucial to avoid jumping into full-scale auto-QA before you’re ready. Even a small increase in coverage from 5% to 10% can make a significant impact. While everyone is eager to reap the benefits of automated QA, it’s important to take things step by step.
This is why we’ve developed SmartScore powered by ChatGPT, which acts as a co-pilot to increase productivity and support you, much like a new person joining your QA team.
What we’ve learnt from testing Chat-GPT for QA
SmartScore results are much more human-like.
Depending on the question being asked, the scale of possible outcomes, and the quality of guidelines it has to work in, it is important to understand that SmartScore behaves like a human whereby the outcome of SmartScore can vary each time as it generates a score based on the whole conversation being evaluated rather than the detection of certain words said by either party in a specific sequence.
For example, when testing “Did the agent welcome the customer appropriately” with a scale of 1-10, SmartScore may mark the outcome as a 7, or when asked again, may score the line item as an 8.
In comparison, Auto-QA produces the same response every time, unless the keywords and phrases it analyzes are changed. As a result, Auto-QA is more suitable for heavily regulated environments than SmartScore.
SmartScore is easy to set up
SmartScore requires significantly less set-up (just adding some prompts to your line items) than Auto-QA, which requires specialist skills and a deep understanding of how to build topics.
SmartScore is commercially flexible
SmartScore is much more commercially viable, low-cost usage-based pricing means it only costs money when the feature is used, reducing risk for those unsure about the investment.
SmartScore needs human oversight
In comparison, SmartScore is currently more manual than other automation products, an evaluator should oversee the results before publishing, this makes it a perfect steppingstone for Auto-QA.
We understand that not all businesses are prepared to switch to a completely automated QA process, either due to budget constraints, timing or fear of the unknown. This is why we have introduced SmartScore as a solution that can help you modernize your QA processes while maintaining the optimal balance between automation and human input. With SmartScore, you have the flexibility to choose which aspects of your QA program to automate and when, which will speed up evaluations and enhance coaching, ultimately making SmartScore an asset to your team.
And the best part is, we have made it affordable and accessible to everyone, so you can reap the benefits of this revolutionary tool without breaking the bank. We guarantee that SmartScore will be the star employee of the year for your business!
4 top tips to help prepare for SmartScore
As simple as it is to get going with this new feature, it’s important to note that SmartScore requires proper instruction and training to function effectively, just like any new team member.
- Ensure your line items on your scorecard are clear and concise. If your line items are subjective, expect some subjectivity in the response. Similarly, if you don’t have very clear guidelines, SmartScore will do its best with the information it does have but may struggle with line items where guidelines are not very detailed or absolute.
- It’s important to remember that SmartScore requires human oversight and approval. Just as you wouldn’t let a new QA person start publishing evaluations to the front line without checking, the same goes for SmartScore.
- Unlike Auto-QA, which scores line items the same way every time, SmartScore is more human-like. When evaluating empathy, for instance, it may present a slightly different result each time it is asked. Therefore, it’s crucial to communicate internally with your evaluators how they should use SmartScore.
- In preparation for SmartScore’s launch in April, we recommend having a full-fetch integration set up and clear guidelines on what the evaluator uses in their marking criteria for each outcome. Take the time to review your existing line items and consider how you can improve them to ensure you get the most out of SmartScore.
Overall, SmartScore is a fantastic tool that can revolutionize your quality assurance process. But, like any new tool, it requires preparation, instruction, and training. With the right approach, SmartScore can take your business to the next level.
Ready to get on board?
EvaluAgent’s latest offering already has many existing customers excited. If you haven’t joined the wait list yet, there still time to sign up and secure your 250 free credits before April.
Sign up now
Not yet an EvaluAgent customer?
Don’t worry we’re happy to get you set up on a free trial of our leading QA & Improvement platform where you can also test out our latest feature SmartScore.
Book a Demo Today