Recently, we wrote about Florida Blue’s multi-channel assistant. A Visual IVR system with a voice assistant that helps the health insurance company’s customers apply for replacement Digital ID cards.
The system was built by Kore AI and currently approves over 1000 new cards every day for customers.
So why did Florida Blue select Kore AI for this project? And how should any company select any conversational AI vendor for that matter?
Daniela says Florida Blue went through 5 key stages in their vendor selection process, and we’ll outline them here.
1 Get your team together
First, Florida Blue put together a large internal mobilisation team, including people from technology, digital, business units, contact centre and more. Everyone was represented.
This is often an overlooked part of implementing conversational automation. Some businesses treat automation as a point solution. Something that’s there to solve a specific, isolated problem.
However, with conversational AI, you’ll quickly realise that you need support from all areas of the business. From branding (how should your assistant sound?) to technology (how will it integrate with your business systems?), to customer services (what should this thing do?) to finance (where will the money come from?) and so on.
Florida Blue did the right thing early on by assembling a mobilisation team from across the business.
2 Get a long list of vendors
If you don’t have expertise in conversational AI and you need to find a partner to help, Daniela recommends starting with a long list of potential suppliers.
You can use Gartner’s Magic Quadrant for Enterprise Conversational AI applications, the SPARK Matrix™: Virtual Customer Assistants 2022 report, The Aragon Research Globe™ for Conversational AI, 2022 as places to start.
Florida Blue interviewed 11 suppliers in total. As Daniela says, “all of the vendors will say, out of the box, you can be ready in three weeks.” But Florida Blue found that time can increase significantly when you include the integrations needed to make the solution useful.
3 Make the short list
The long list was shortened during the interview process by asking the right questions and understanding the potential partner’s ability and experience in delivering similar solutions to similar companies.
Florida Blue were asking questions like:
- Who have you worked with in the past?
- Do you have proof and references of previous work?
- What kind of solutions have you deployed i.e. use cases, channels etc?
- Can you show me a working example of this technology in action?
- Can I have access to the platform to kick the tires?
If your provider doesn’t have the relevant experience, if they can’t show you working examples of their technology working in production, if they don’t have client references that show they can deliver, and if you can’t see the technology working, then they shouldn’t make it through to the next stage.
Through doing these interviews, Florida Blue were able to narrow the potential vendors down to just 4. Then, an RFP (Request for Proposals) was conducted with the 4 vendors.
By this point, each of the vendors had been able to understand the needs of the use case, and understand the requirements of the project. Putting together a more detailed proposal at this stage is more feasible.
At the same time, Florida Blue had been educated about the practicalities of implementing its use case, and so could write a more effective RFP.
5 Bake off
The RFP process whittled it down to just two vendors. Then, Florida Blue had each of the two remaining vendors bake-off.
Each vendor built a technology proof of concept to prove that their solution was capable of delivering against the use case.
This isn’t something that all vendors will jive with. Building a proof of concept is a big effort, and to do this before you have a deal signed is costly.
However, as Daniela says: “of course, it’s costly for everybody, but it is, right now, especially with new technology, the ultimate proof”.
Every business is different, and each has unique situations to deal with. You need a partner that can work with what you have and support you, no matter what situation you’re in.
One of the deciding factors for Florida Blue is that the business isn’t 100% cloud. Much of its business systems are running on prem. This means that they had a specific requirement for a vendor that can support this hybrid set-up. Here, there had to be true cooperation and collaboration between both companies to ensure that whatever is built can be implemented on Florida Blue’s systems.
Ultimately, in this instance, Kore AI came up trumps.
The need for education
When we published the top 20 biggest AI adoption challenges, the number one challenge was lack of education among senior leaders and budget holders regarding the value and practicalities of conversational AI.
Raj Koneru, CEO, Kore AI, told us “IT folks have been used to UI based computing all their lives, right? It started with the green terminal, [then] mainframe terminals, to Windows, to web to mobile, and everything on the screen. For the first time, the challenge is to deliver an experience with just language.”
That means one of the primary jobs for AI vendors is to educate customers early on.
Proving value to stakeholders
As soon as you’ve selected a vendor, you’ll need to demonstrate that the project is delivering on the value it promised. Daniela told us it was vital to quickly start convincing the upper management that their investment in the project was bearing fruit.
And it did bear fruit! Once the Visual IVR system was live, Florida Blue asked all customers for a thumbs up/thumbs down review. Over 90% were satisfied.
That was what they needed to keep the project’s momentum going. They had their proof, they communicated it internally, and then they had buy-in to continue evolving the project into other use cases and channels.
According to Raj, that’s often how it goes. “That’s what I hear from a lot of customers. Once they try it, they get some benefit out of it, they see the light at the end of the tunnel.”
For more information on Kore AI, you can view case studies here.
And if you’re serious about conversational automation and don’t know where to start, check out our Buy vs Build Guide to conversational automation.
This article was written by Benjamin McCulloch. Ben is a freelance conversation designer and an expert in audio production. He has a decade of experience crafting natural sounding dialogue: recording, editing and directing voice talent in the studio. Some of his work includes dialogue editing for Philips’ ‘Breathless Choir’ series of commercials, a Cannes Pharma Grand-Prix winner; leading teams in localizing voices for Fortune 100 clients like Microsoft, as well as sound design and music composition for video games and film.