By asking a simple question at the end of each Virtual Agent interaction - Was this solution helpful? yes/no? We capture the end user 'feedback' while showing the value of the Virtual Agent.
Of course, we take into account the total number of interactions along with cost per incident, etc. Sometimes to keep ourselves grounded, we get closer to the interaction (closer than the automatic metrics) and conduct a random session analysis. By taking a group of sample interactions and truly reading the customer intent, we can not only prove that automatic reporting is 'conservative' but we also get to understand what 'Joe' was really thinking when he got a solution he wasn't happy with.
With that, I will share with you a sample ROI - Computed using Resolution Rates.
And because it's Customer Feedback Friday, keep in mind that ROI is computed by asking the customer whether or not a solution was helpful.
In just a few months after going live:
71% resolution rate
24% unresolved (29% were out of scope)
5% partially resolved
Enjoy the holiday weekend.