How to evaluate your AI Virtual Assistant Vendor

If you are in the banking and financial industry, you have already noticed that virtual assistants are everywhere. Virtual assistant is one of the 7 biggest trends that are currently shaping the financial industry. According to a Gartner report, 85% of all interactions between a financial institution and the customer will happen over an AI interface, by 2023, taking precedence over all other 2-way communications like mail, phone, and text. Big banks have already adopted it and are seeing benefits in cost-saving, improved revenues, and stellar customer satisfaction. So if you haven’t already implemented a virtual assistant, chances are that you are in the midst of evaluating one right now.

But the question remains, how does a financial institution find the right technology partner is this overly cluttered world of solutions providers. Given that these implementations take time and use up a considerable chunk of your IT investment, you have to be sure about the decision. Do you talk to a big tech company, do you engage a small niche solutions provider, or do you try and find something in-between and hope that the solution works.

Here are some of the questions you can ask your vendors and the potential response you should look for

Who are your customers?

You want to partner with someone who has worked with other banking clients before this and has worked on implementing solutions that match your scale or larger than your scale. Check the logos on the website and the pitch slides. Ask them about their implementation experience with other customers.

Ideally, it’s best to look for vendors who have a deep knowledge of the domain. There are a lot of virtual assistant providers right now that offer the solution, but they might not have an in-depth understanding of the BFSI industry. It would be painful to explain the nuances of the business and then force-fitting that domain knowledge into their solution. The chances of that getting botched are so much higher.

What problems have you solved for your clients?

Ask if the vendor has worked on problems similar to the ones that you are trying to solve. There are several vendors that are working on niche problems that may not have immediate business value. You want to invest in a solution that solves your current problem, saves operating expenditure, or helps automate a mundane task, or acquire more customers.

Break-down your problem into smaller chunks and try to solve for the easy parts first. Look for a vendor that can do all the chunks well, but start off with the easiest problem and see if the solution works for you. This will help conserve time and resources if the partner did not turn out to be a great match.

Also Read: How to humanize Debt Collection with an AI Banking Chatbot

Is the solution Omnichannel?

As banks are adopting and adapting to meet customer expectations, the user experience has moved beyond just web and mobile to smart speakers, connected appliances, voice assistants, social media and messaging apps. Engage a vendor that can help you give customers the complete omnichannel experience that doesn’t lose context.

What’s the underlying technology used in your platform

Check the under the hood to see what makes the solution special. What drives the platform or the solution. How does AI, ML, and NLP work together? But don’t get lost in the jargon. Rather you might want to understand where in the spectrum does the solution lie. Is it closer to a chatbot or is it closer to a virtual assistant? Is it context-aware or do they forget the context if there is a break in the conversation? Does it understand human sentiment and emotion, or is it robotic? An analysis of these questions will tell you which solution works for you and if the vendor has the capability to solve the problem at hand.

How long does it take for the virtual assistant to go live?

A really big question that often bears heavily on the minds of CIOs while making a decision to invest in technology is how long will the implementation process take. AI requires data to be trained, over a period of time to get the right information in front of your customers. Get an idea about this process and understand how long the process will take from start to go live. You don’t want to be working with a vendor that takes 3-6 months to get a small part of the solution off the ground.

What will be the accuracy of the responses, and how long will it take to reach that level of accuracy.

Most of the top technology players in the industry are able to reach an 80-85% accuracy in a span of 4-6 months. While evaluating the vendor, look for a similar or better statistic.

How long before we see an ROI for implementing the virtual assistant?

This is an important consideration. There is a substantial investment that goes into adopting new technologies and CIOs should ask when they will be able to reach the ROI milestone. The ROI is dependent on several factors like the current support costs, the number of resources engaged in current support activities, and the support window. Alternatively, they can also check to see what was the ROI of their previous customers.

Also Read: 5 Steps to Measure the ROI of an AI Bot (ROAI)

How much of the workforce will be required to manage the solution?

Hiring extra workforce to manage the solution will add to the overall operating expenditure and impact ROI. Look for a vendor that can support or manage the solution without significantly overburdening you with extra costs.

However, have a contingency plan in place to make changes at your end if required in case of an emergency. While that would be difficult to complete without a programmer, a low-code program might be able to strike the right balance.

What happens if the assistant can’t answer the question?

There is constant tweaking required for updating the knowledge base of the virtual assistant so that it can remain current on all the products and services offered by banks. It is thus possible that there will be questions that fall outside the purview of their knowledge base, in such a scenario a seamless transition to a human agent should be done so that customer experience is not hampered.

Will it integrate with my current legacy system and other internal systems?

There has been a lot of discussion about moving away from legacy systems to different types of core banking solutions and CRM solutions. While this movement will take a few years for most traditional banks and credit unions, it is important that there is compatibility between all types of internal systems that are used.

Another thing that needs to be evaluated is how is the overall banking system, is it too clunky with a myriad of different systems? Are new applications just joined together by links and APIs that diminishes the customer experience? Look for a vendor that can ensure a smooth user experience that takes the customer journey forward rather than frustrates them.

How do you ensure data security and regulatory compliance?

As a financial institute, it is of paramount importance that the financial data of customers are secure. The system needs to be robust such that customer data is masked and in most cases not stored so that there is no misuse of the data. Robust security and compliance should be in place that is compliant with laws and regulations of the country.

In conclusion

As a financial institution you are looking for a partner rather than a vendor, who can be a part of your digital transformation process, rather than walk away as soon as the implementation is done. AI solutions require continuous improvement and tweaking. Look for a partner who can understand your problems and helps find a unique solution with you together and be a part of your AI journey.

Also Read: AI in banking industry: The future is here