Can AI automate 80% or 60% of your calls?

Srinivas Njay
Srinivas Njay

In the financial ecosystem, with everyone offering nearly the same products and services, there is less room to compete on price; the experience customers have with their financial institutions is what differentiates one from another.

Earlier, individuals with mobile phones could receive text messages containing banking/account information, & this was a commonly used mode of communication. With the rise of more complex transactions, an Interactive Voice Response (IVR) system was developed. An IVR system is a static automated telephony system that registers customer requests by defining, segmenting, and linking them directly to an executive who can answer their questions.

Different ways of incorporating AI in Call Centers

With an increasing need for accessible and efficient customer service, Artificial Intelligence (AI) has become essential to ensure better efficiencies. There are vendors in the market who are offering AI in two different ways for Financial Institutions -

Option 1 – Vendors who are setting up AI within an IVR system

Option 2 - Vendors like interface.ai who are replacing the IVR system with AI

The vendors who set up AI within IVR make promises of 80% call automation, whereas interface.ai makes a promise of 60% call automation.

While the automation levels promised by the former vendor type(Option 1) appear to be higher, it is essential to understand how these metrics are measured to get a clear picture.

Let us compare and contrast the two options of implementing AI offered by vendors and how these vendors are measuring their automation levels and accuracy.

Option 1 - AI within an IVR system

A few AI vendors offer a service that would incorporate AI into a financial institution's IVR system. This integration helps the end customers to have a conversation with the AI when requested explicitly. This enables financial institutions to address a limited number of intents (queries) - usually around 3 or 4 - through AI, via a customer request for it through the IVR menu. In this case, only a few calls are transferred to AI, providing immediate customer service only for a few customers and for limited queries.

Let us understand how this works more thoroughly through an example of Chartway Federal Credit Union, powered by one of the other vendors. Let us refer to the below image (Image 1) –

The following assumptions are considered to explain the customer journey through an example-

a)     The financial institution receives 10k calls every month
b)     Cost to address each call - $10

Image 1 - Customer journey in a setup having AI within an IVR system

Observing the above illustration, when Chartway FCU receives 10k inbound calls to its call center, it automates only 1600 calls which is 16% of the total calls received.

How is the automation % this low when the vendor claims 80% automation?

As seen in image 1, only a small subset of customers explicitly choose to have a conversation with AI. From our experiments, this subset is anywhere between 5%-20% of the calls. Also, the number of intents (queries) that the AI can answer in such a setup is usually limited to 3-4 (answering questions related to Check balance, Transferring funds between accounts, and Checking recent activities). Even the 90% accuracy claimed by the vendors offering such a setup is associated with addressing only these 3 to 4 intents with 90% accuracy.

In conclusion, even though the vendor claims to automate 80% of the calls with 90% accuracy, it turns out that only 16% of all calls are automated at 90% accuracy for just 3-4 intents.

The table below represents the impact of implementing the AI within the IVR as per the above illustration(Image 1).

Table 1 - Impact Generated through AI within an IVR system

Do you think such implementations provide financial institutions with high automation possibilities, improved service levels, enhanced customer experience, and a good return on investment?

A limited combination of AI and IVR would not be sufficient to satisfy most of the customers or financial institutions. All the workflows which are not automated would be transferred to the customer support staff, thereby not significantly enhancing the operational efficiency and service levels.

Option 2 - AI Replacing the IVR system ‌‌

On the other end, with interface.ai, the IVR is replaced by a market-leading AI-powered Intelligent Virtual Assistant (IVA) to enable financial institutions to transform their call center from being cost centers to revenue centers. This is the second way of incorporating Artificial Intelligence in call centers of financial institutions.

When an IVR is replaced with an IVA powered by interface.ai, the IVA handles 100% of the calls received by the financial institution. In this case, financial institutions can experience significantly improved call automation, service levels, and efficiencies. Also, having natural, human-Like communication ensures a better customer experience with technologies such as Neural Voice. With interface.ai's IVA on the call center, financial institutions can automate 60% of the total incoming calls within just a few weeks, even by handling 100+ intents, interface.ai's IVA ensures accuracy of over 90% since day 1 of launch.

Let us look at interface.ai powered University Credit Union (UCU) call center where the IVR is replaced with AI and understand how it is set up with the help of the below image (Image 2)

The following assumptions are considered to explain the customer journey through an example-

a) The financial institution receives 10k calls every month
b) Cost to address each call - $10

Image 2: Customer journey in a setup when AI replaces the IVR system

Unlike the previous AI setup (illustrated in image 1), we understand through the above illustration (Image 2) that when UCU receives 10k inbound calls to its call center, 100% of the calls are being attended by interface.ai's IVA that has replaced the IVR system. interface.ai’s IVA automates 60% of the total calls, i.e. 6000 calls out of 10k calls, is automated with 90% accuracy for more than 100 intents. And this is being achieved from day 1 of the implementation.

The table below represents the impact of AI replacing the IVR system as per the above illustration(Image 2).

Table 2 - Impact Generated through AI replacing the IVR system

This setup of AI replacing the IVR leads to greater efficiencies in the operations of the financial institutions and improved service levels. The IVA also enhances the customer experience by eliminating call abandonment rates, call wait times, and ensuring that the customer service staff get sufficient bandwidth to cater to customers in need and help them achieve financial wellness.‌‌

The table below summarizes the differences between the implementation of AI by different‌‌ vendors.‌‌

Table 3 - AI within an IVR System v/s AI replacing the IVR System

We can see from the above table that when the AI is set up within an IVR system, only 16% of the total calls are automated (unlike the claimed 80%) whereas, with AI replacing the IVR (interface.ai), AI handles all the incoming calls and automates 60% of total calls.

We also observe that when the AI is set up within an IVR system, 90% accuracy is achieved for only 3 to 4 intents whereas when the AI replaces the IVR system, 90% accuracy is achieved for 100+ intents.

In addition to this, with the AI replacing the IVR, the financial institution is guaranteed ROI as interaction with AI is not left to chance. It also enables customers to engage with AI in a natural, human-like fashion, thus guaranteeing a better customer experience.

The impact of the two options on financial institutions

Let us compare the impact the two options may have on a financial institution that receives 10k calls on the call center every month.

Table 4 - Impact on the ROI the two systems may have on a financial institution‌

By looking at the above example, it is evident that even though the setup of having the AI within IVR promises 80% automation, it is the setup where AI replaces IVR that automates a larger number of total incoming calls, thus offering improved efficiencies, better savings and offers vastly better ROI.

In conclusion, financial Institutions need to evaluate AI implementations based on the overall possibilities with a solution. It is essential to consider how many intents will be handled by the AI and how these implementations will impact the overall automation levels, service levels, member experience, and revenue generation opportunities.

interface.ai provides you with the best-in-class experience for your customers through our AI call center, where the AI replaces the IVR system.

Benefits of interface.ai's AI Call Center

  • Automating calls and Saving costs: Automates 60% of your calls with 90% accuracy from day 1 by addressing 100+ work intents.
  • Improving Customer Experience: It provides human-like empathy and gives a solution to complex queries along with providing personalized product recommendations – Helps you to cross-sell and up-sell
  • Improving Employee Productivity: Provides front-line assistance, which helps employees solve customer issues with greater efficiency by providing customer information and relevant policies. ‌‌

You can read more about the impact we have enabled and customer testimonials here - https://interface.ai/case-studies/

Would you like to incorporate a technology that can enhance customer experience, improve automation, enable significant top-line and bottom-line impact, and free up your support team to provide personalized support to members? or would you instead incorporate a limited technology that will hinder your operations & leave you with a lot of opportunity cost and unrealized ROI.

There is no wonder the latter technology is offered at a cheaper price considering the ROI is lower and the technology provides a sub-standard member experience. On the other hand, with a slightly higher cost, you get a technology that ensures significantly higher ROI and enhanced member experiences.

Choosing between interface.ai and another vendor when it comes to an AI setup is purely a choice of win-win or lose-lose.