[Part 7] How Top Banks Leverage IVAs to Enable Self-Service: Three Common Questions

By Richard Simons, CEO of Creative Virtual USA

This is the final post in our seven-part blog series – How Top Banks Leverage Intelligent Virtual Assistants to Enable Self-Service.


During our recent webinar on how banks leverage intelligent virtual agents, three frequently asked questions surfaced. I thought I’d answer them for you here.

Question: What call deflection rate can I expect after implementing an intelligent virtual agent?

It really depends on the types of questions that people are asking. If somebody asks, “How do I check my credit limit?,” we can just push them to a “how-to” page or better yet, give them an answer to their question such as, “Your credit limit is $3,000.” When we do this, we see call deflection rates in the 80% range. However, if someone asks a question like, “How do I download transactions into QuickBooks? I’m getting error 4561,” we might see satisfaction in the 20s because this question might best be solved with human help. Financial institutions can see an average call deflection rate of 30% with some seeing rates up to 50%.

Question: Our mobile efforts are siloed and our IT department is struggling to keep up with customer demands. How can intelligent virtual assistants provide support on mobile devices?

There are many different options available for mobile. For instance, we can build separate interfaces and templates for mobile, tablets, and for the web or use responsive design. We can also customize and shorten answers for the small screen. No matter what the option or type of implementation, we deploy the same system using the same code regardless of the user’s device. Using our knowledge management system, we have a lot of flexibility making us completely device agnostic so there’s no work for you to do to make our systems available across all mobile environments.

Question: What is the implementation process in terms of internal resources and time to market for an intelligent virtual agent?

Before anything starts, we project the ROI and then typically with our deployments, start with a pilot. Our ideal scenario is where we can do an assessment. If the client has live chat, then that’s a very good place for us to start. Otherwise, we use call center transcripts. With one client, we started with the intelligent virtual agent being deployed on the contact us page handling specific issues that related to people traveling overseas. We built the content from live chat transcripts, and deployed a three month pilot. We were then able to confirm the ROI and move into a full deployment.

In financial services, beyond compiling the answers to typical customer questions, the hardest part is getting approval for those answers. That really takes as long as it takes, but typically we would get a pilot out in two or three months. Once we’ve run a pilot, we often roll straight into a full deployment which typically takes three to six months for a financial institution.


  1. If you haven’t already, watch the 30 minute webinar to see a live demo of how top banks are leveraging intelligent virtual agents.
  2. You can read all the past blog posts in this series here: Part 1Part 2Part 3, Part 4Part 5 and Part 6
  3. If you’re ready to talk about a running a pilot to determine if an intelligent virtual agent is right for your business, contact us here