Kevin Wu is the CEO of Leaping AI. Leaping AI automates complex call centers with self-improving voice AI agents.
Generative AI (GenAI) is omnipresent in the current news cycle. Contact centers are often cited as prime targets for GenAI disruption, due to the manual and repetitive nature of the work often associated with them. It is easy to imagine that implementing a voicebot or voice AI agents could dramatically reduce costs.
As CEO of Leaping AI, a voice AI agent company that helps companies automate complex call center operations, I speak with many different companies that are interested in streamlining their call centers. These companies often see call center work as a repetitive activity type that can potentially be automated with AI.
In the following article, I’ll highlight scenarios in which a voicebot either makes or does not make economic sense. Furthermore, I’ll outline the rationale underpinning that designation.
Where A Voicebot Makes Sense
The following are a few use cases where I think a voicebot would be a great fit:
1. A Large Call Center: Say your business has a large call center with 20 employees or more. In this case, contact center costs, by my estimate, can range upwards of $1 million annually. Typically, as far as I’ve seen in the industry, these businesses receive at least 200,000 calls a year. Introducing a voicebot could significantly streamline costs here. For instance, if 50% of calls can be automated with AI, that means that a significant number of employees are now freed up to focus on revenue-generating activities, such as upselling. These larger cost savings could justify the upfront investment costs associated with setting up a voicebot
2. 24/7 Support Requirements: Some businesses require 24/7 support. Take healthcare, for example. Patients might demand the ability to always reach their provider in case of problems with their medication or device. Having a 24/7 virtual assistant available could allow healthcare companies to offer easy overnight support, while allowing overnight staff to only handle the truly critical emergency calls.
3. Sales Enablement: Some businesses are already using voice AI agents to aid in sales—think lead qualification. These use cases are often located within the sales and marketing teams, instead of customer support teams. In one instance, a company that Leaping AI worked with was able to quadruple the numbers of leads qualified over the phone after adopting an AI-based phone solution.
Where A Voicebot Might Not Make Sense
On the other hand, sometimes voicebots may not make sense for every business. Here are a few use cases where I wouldn’t recommend implementing a voicebot:
1. Smaller Call Center: If your business has a smaller call center of less than 10 employees or more, the cost savings associated with a voicebot might not exceed the costs required to implement the voicebot. Since AI cannot handle all calls, humans are still needed as a fallback, and if the call center is small, reducing headcount could be counterproductive, as it might lead to some understaffed days.
2. Complex Use Cases: If each customer conversation your business deals with is unique (think individualized and comples), a voicebot may not be worth it. It might be too difficult and time-intensive to train the voicebot to be effective across the board.
The Economic Rationale
For any company, the costs associated with a voicebot can be calculated by combining your implementation costs, your base cost for running the voicebot and the time spent maintaining the voicebot. GenAI is not perfect. It is sometimes unreliable. In my experience with some of our customers at Leaping AI, we’ve found that only 50%-70% of use cases and conversations can be automated—at most. That means some human supervision and fallback is necessary.
For a smaller call center, the absolute savings might be limited (you might need two to three fewer employees at the call center), especially since humans still need to be present at all times to act as a fallback if the customer prefers to speak to a human. Contrast that with the five-figure or six-figure investments necessary to set up and maintain a voicebot, and you end up with a negative business case.
Voicebot Best Practices
If you do decide to implement a voicebot, there are a few best practices that I’ve found effective working with clients on their implementations, including the following:
1. Be realistic about whether introducing a voicebot makes economic sense for your organization—and about the potentials and limitations of AI.
2. Rally different parts of the organization to support the implementation of a voicebot. I’ve found this can help you maximize your chances of success.
3. Constantly evaluate the KPIs you agreed on with your vendor. Share and discuss them with your vendor on a monthly basis.
4. Do not treat the voicebot only as a cost-reduction exercise; instead, look at it holistically. It’s important to recognize that it frees up valuable employee time that can be spent on more revenue-generating activities.
Conclusion
Introducing a voicebot might make sense for you if your business has a large call center (or requires 24/7 support) and would like to achieve cost savings at scale.
That said, implementing a voicebot is quite the undertaking. It’s important to stay well-informed about this technology and the potential challenges that come with it. If you keep my best practices in mind, though, you could be well on your way to maximizing your chances of success with your voicebot solution.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
