Rewriting the Agile Manifesto for AI in Customer Experience

“Agile” might seem like a quaint topic in 2023. After all, the Agile Manifesto that jump-started the agile adoption movement for those who build, develop and use technology was written in 2001. But that manifesto is as relevant today as it was when it was first released.

For those in the customer experience space, agile isn’t an option — it’s an imperative. If our customer experience technology doesn’t satisfy fickle consumers, they’ll take their business elsewhere. In the “State of the Customer Experience” report from Genesys, 31% of consumers reported they’ll stop doing business with a company after just one bad interaction. And they won’t look back.

CX leaders must adapt to customers’ changing preferences, channel choices and expectations — and they must anticipate what the customer will want next and build for that future while meeting short-term business objectives.

Agile development means having short bursts of innovation and experimentation — and being willing to fail quickly — all to enable long-term success. Being agile means adapting to changing situations quickly without being defined by the rigid rules of traditional development practices. Some of the most agile organizations — the disruptors — can innovate faster because they’ve embraced artificial intelligence (AI).

AI is one way to free your organization from rigid rules that were created for a bygone age where customer loyalty was assumed, not earned.

AI Enables Conversational Agility

IVRs were one of the first forms of automation that contact centers embraced. Those IVRs enabled anytime, anywhere service access that agents couldn’t provide without a significant cost. IVRs were programmed using rule-based flows that grew in complexity over time, became more difficult to maintain and were virtually impossible to change.

Anytime, anywhere service wasn’t simple. Customers had to work hard to remember what the options were as they made their selections from lengthy, confusing menus. Many only chose an IVR as a last resort.

Consider some of these statistics:

• 75% of customers become angry when they can’t speak to a live agent
• 30% will hang up if they have to deal with a complicated service model
• Customers prefer natural language IVR systems by a margin of 66%

Natural language offers an alternative that enables anytime, anywhere access while freeing customers from having to press 1 for Yes or 2 for No. Most people would prefer to talk and be understood in their own words versus being forced to listen to and follow instructions on which number or combination of numbers will get them the answer they need.

Beyond the actual conversation, customers are no longer tied to the phone. Having the ability to engage through voice or digital is as important as the actual conversation. Conversational AI makes it possible to accurately transform speech to text (SST), understand and formulate a response through natural language understanding (NLU) and natural language processing (NLP), and then respond digitally or via text to speech (TTS). This form of AI is the technology behind IVAs, also known as voicebots, chatbot or just bots.

Interactive virtual agents (IVAs) are an evolution of the IVR. IVAs were designed to help callers solve their problems — without the help of a human agent. Hence the name: They’re like a human agent, but virtual.

IVAs interact with a caller using natural language. They’re more agile than a traditional IVR, which is built using rules with each possible path fully designed. Modifying those rules can be difficult and many IVRs become bloated and difficult to manage over time.

IVAs are trained to respond to natural language. They are trained; not programmed.

Virtual Assistant Programming Vs. Training

While an IVR is programmed, an IVA is trained. Agility extends beyond the experience to the build process itself. IVRs are programmed with rigid paths based on selections; IVAs are trained using conversational data. The IVA can be retrained with new data to help it understand, process and respond to customers using language.

A virtual agent might present a few menu options to get a user started, but any good IVA is trained on all of the most common questions — into the hundreds. You don’t need a menu when you can just ask.

IVAs are trained to understand common questions, walk a user through a process or flow, and then give them a final answer or route to a human agent, when necessary. But the key part is that callers now have free reign to ask whatever question is on their mind.

This allows the self-service solution to handle a much larger corpus of knowledge, and anything it doesn’t know is also captured, to be reviewed by the IVA builder and, when it makes sense, added as new content. So not only can the IVA handle a larger range of options, collect information and solve problems with a caller, it can also capture information along the way to be reviewed as analytics to better understand the voice of your customer and improve over time.

Knowledge Automation Is Fundamental to Agility

“Intelligence is the ability to adapt to change.”

This is one of the more famous quotes for the agile practitioner, and most often attributed to Stephen Hawkins. One way in which intelligence for customer experience is delivered is through AI-enabled knowledge management.

IVAs draw on knowledge to answer questions. Knowledge can be delivered through rigid rules — it’s not uncommon to use keywords or phrases even in this age of AI. However, an AI-enabled knowledge base uses semantic search. It is more responsive and, by definition, more agile.

AI-enabled knowledge with natural language can enable an organization to create more agile, more responsive, and more intelligent self-service and agent-led experiences. For example:

1. Smart search on the web that supports type-ahead and enables customers to see what questions are the most frequently asked
2. Interactive, visual information delivered through a digital IVA
3. Helpful guidance delivered through a voice IVA
4. Instant, contextual information delivered to an omnichannel agent talking to a customer

This type of agility requires knowledge that can support multiple uses, or variations. A single piece of knowledge might need to appear in multiple ways to the various users of that article. Intelligence needs to be adaptable in order to be agile.

Beyond conversations, AI can replace rigid rules that govern how interactions flow. Intelligence for interactions is already there. Every successful interaction is an opportunity for learning. With AI, routing strategies can be discovered.

Rule Discovery Is the Key to AI Agility

Agile means being able to read a situation quickly and react accordingly. With AI, rules are discovered rather than imposed. For example, predictive routing is based on rules that AI had discovered based on existing conversations.

Similarly, IVAs are trained based on conversations that had already happened. The key to agility is to use data to guide decisions. As the amount of data that is captured grows and becomes more unwieldy, traditional analytics and reporting processes take too long to find an answer.

AI can work with large stores of data and:

1. Organize data into segments and factors
2. Identify common and uncommon (outlier) patterns
3. Create a model based on a desired pattern or outcome
4. Predict an outcome (number, word, sentence)

Once rules are discovered, they can be applied. Some ways these rules are applied are through:

1. Bots: Bots are trained using conversational data to respond to questions and perform tasks. Rules are made up of utterances (what people say) and intent (what people mean).
2. Predictive engagement: Predictive engagement predicts outcomes based on behavior history to suggest the next likely action (an offer) or personalize a conversation
3. Predictive routing: Predictive routing identifies and applies routing rules that optimize a KPI

Staying Agile in Customer Experience

Maintaining agility is sometimes harder than achieving it. It requires vigilance, data, analytics and a growth mindset from all those involved in the process. For customer experience, agility means a willingness to experiment with new ways of connecting to customers, new ways of engaging employees and fully embracing AI as a partner — not just an accelerator.

Read more about what Genesys AI can do for you in this guide on creating better AI-driven experiences with a human touch.