The cloud, combined with conversational artificial intelligence (AI), is dramatically expanding the capabilities of the modern-day contact center. These solutions are the twin pillars of contact center success, allowing them to serve more customers faster and more effectively.
The two technologies go hand in hand for creating the flexible, flawless customer experience (CX) that companies everywhere are striving for. The market for cloud contact center solutions is expected to reach $11.74 billion by 2028, and 80% of those companies that migrate to the cloud plan to use AI and machine learning technologies to further improve customer experience in the cloud.[1] In fact, in most cases, investing in a cloud contact center includes access to the platform’s proprietary AI solutions.
Layer upon layer of shifting complexity
In many ways, moving to the cloud simplifies the systems, hardware, networks, and databases contact centers are built on. But it also opens the door for new business-driven complexities around processes and integrations between the APIs, channels, and platforms that are now working together to support the contact center.
The shift to the cloud coincides with a pivot away from VOIP and traditional telephony and a move toward real-time communication where customers open a web browser and go through a website or use voice or video directly from the browser to engage support. With this evolution, digital will become more prominent than voice and will add additional volume for the contact center to accommodate. Customers will also use multiple channels concurrently. It is difficult to transition from one medium to another while providing a seamless customer experience, adding more complexity and volume.
AI adds yet another layer of complexity. In the contact center, AI powers chatbots and voicebots, but it also helps personalize your experience, such as by customizing prompts or changing the order of menu items. It also drives reporting and analytics to enable better understanding of customer feedback and intent, as well as providing real-time agent assistance. All of these things need to be integrated, and the integration and resulting customer experience need to be tested and monitored.
As the cloud and AI expand contact centers’ capacity for customer service, they also introduce a testing burden of such a scale and scope that no company can keep up without automation. With manual testing, it is impossible to keep up with the complexity.
What it takes to execute flawless CX
When it comes to the customer journey, it has always been critical to test every possible path to ensure that the customer can navigate without anything getting in their way.
For instance, even in a simple IVR setup, a customer will call in and receive several options at each menu. Any given customer might select any number of combinations and permutations as they journey through the IVR, creating numerous potential pathways through the system — each of which needs testing. In practice, it’s best to test every potential pathway from end to end. That’s the only way to ensure you can find defects and resolve them before they impact customers. Keep in mind, too, that traditionally a human tester (or a team of testers) manually tests their way through every identified journey to make sure everything works as expected. It’s no small task to keep it all straight and make sure everything is tested thoroughly and regularly.
The true cost of conversation: creating the impossible manual task
When considering traditional IVRs and basic chatbots, customer journeys through these systems are linear with one step naturally following the next, making testing a significant but manageable task. But what happens when you introduce the complexities we discussed above?
Let’s first consider what conversational AI adds to your IVR or chatbot solutions. The real value of this technology is its ability to create natural language and conversational flows within your customer journeys. Unlike the legacy self-service solutions, which were confined to a highly limited and predefined set of customer inputs and responses, today’s chatbots and conversational IVRs have the natural language processing (NLP) capabilities to navigate much more complex conversations.
That increasingly complex, natural language flow quickly multiplies the potential pathways for your customers. Let’s consider the example of an IVR for a bank.
Traditional IVR without NLP capabilities
Traditional IVRs follow linear call flows, and each needs to be tested. For example, when you call into your bank, you can check your balance, transfer funds, hear information about the bank, or speak to a representative. Each of these options represents a different call flow, and each needs to be tested with all the potential combinations of responses a user might input at each step. With all these possible combinations, the number of potential unique call flows grows to be quite large. In our experience, a standard IVR can easily have 1,000 call flows. And each of these call flows needs to be tested – if you are manually testing, that is a big task, but it’s probably still possible.
Conversational AI-based IVRs
This task becomes exponentially more complicated when you introduce conversational AI, where the potential ways a user can respond to each step grows exponentially. The reason being that instead of fixed options of responses – “press/say 1 for checking,” “press/say 2 for savings,” etc. – you now need to understand all the possible ways someone could respond to the prompt. A conversational AI system would typical have 60-100 different ways to respond (we call these utterances) to any prompt.
So now, take that same IVR where you have 1000 call flows:
- 1000 call flows
- 10 steps per flow
- 60 utterances per step
You now have 600,000 call flows to test. That task is impossible to do manually.
This scenario isn’t far-fetched. In fact, it’s conservative. When you consider that many bots are programmed to speak multiple languages and that they often must “disambiguate” or ask clarifying questions, it’s easy to see how those test cases can balloon even further.
Making the impossible possible through automated testing
This ultimately leaves contact center executives with two options. You can, of course, forego adopting these advanced conversational AI technologies and stick to a more manageable manual testing task. In reality, however, this isn’t a viable choice. Research from Aberdeen shows that companies that deploy AI solutions have 2.5 times higher customer satisfaction rates and generate 2.4 times greater increases in annual revenue.[2] Adopting AI is the only way to remain competitive in a changing market.
Instead of avoiding AI adoption, the better option is to adapt your testing processes to fit the new technology. Cloud-based, AI-driven contact centers need to expand the scope of their testing exponentially, and that requires automated testing solutions that can handle the full range of the customer journey through their IVR and chatbot systems.
Cyara Botium is the only solution on the market that can truly make the impossible task possible by covering every pathway. To see the power of AI testing AI for yourself, check out our on-demand demo today.
[1] SkyQuest Technology Consulting. “Global Contact Center as a Service (CCaaS) Market to Hit Sales of 11.74 billion by 2028.”
[2] Aberdeen. “Contact Center & CX Trends 2019.”
