Introduction
As artificial intelligence (AI) has developed rapidly over recent years, so have the applications of AI in business. One application that has caught on quickly is the customer experience (CX) copilot. These AI-powered assistants are powerful tools for the customer service professional. Let’s explore the possibilities that an AI copilot can provide to your contact center, customers and business.
Basics
So what exactly is a CX copilot? Where did they come from, and what is it that an AI copilot does differently that allows for increased and improved engagement?
AI has revolutionised the way we think about the customer experience. It allows us to easily take simple, rote tasks that agents, supervisors and other CX staff previously had to handle and automate those tasks. And as AI has advanced, so have the things that we can ask it to do.
A CX copilot is an AI-infused assistant for human agents and their supervisors. It can act proactively in real time to surface critical information without being asked for it, provide context and guidance for agents and supervisors, and more. How is this done? Put simply, AI. Let’s dig more into how this all works.
AI has allowed for great strides in what’s possible in providing customer support. Let’s explore how that’s happened.
The underlying components that go into AI are important to understand in order to know what AI copilots are capable of doing. Two of the most important subsets of AI are machine learning and natural language processing (NLP), as they allow an AI bot to grow and adapt over time and to understand what people are asking it.
First, machine learning involves developing algorithms that allow computers to learn from and make predictions based on data. It enables a system to improve performance over time without being explicitly programmed or re-programmed by a person.
NLP, on the other hand, enables computers to understand, interpret and generate human language. It does this by breaking language into smaller pieces, analysing grammar and semantics, and understanding context.
What makes CX copilots worthwhile for your business? What are the benefits that they can provide your customers, agents and business?
In the case of a copilot for agents, for example, it provides real-time customer insights by connecting all the parts of that customer’s journey and telling an agent what they’re most likely to need next. It can also surface information to help solve issues that the customer may not even fully understand.
Automation through AI means less time spent by human agents and supervisors on simple or repetitive tasks and queries. For example, the copilot takes care of things like surfacing necessary information for agents and pulling key insights from transcriptions; and scoring and summarising interactions for supervisors to make assessment easier.
Exploring
Let’s take a deeper look into how CX copilots function in different industry applications.
A customer visits a retail website looking for a specific item, but they’re having trouble finding it. Luckily for them, this business employs an AI copilot for its agents. After a certain amount of time spent looking through various types of product, they contact a customer service agent.
The copilot is able to support the agent in providing a personalised experience. It allows the agent to more quickly help the customer find the specific item they’ve been searching for, but it can do it without offering options that include things the customer’s already purchased in the past. It can also identify options for upselling or cross-selling based on what they’re looking for today, meaning the suggestions will be tailored for the individual customer rather than being random.
The copilot then provides the agent’s supervisor with an automatically generated summary, including insights like the reason for the interaction, the resolution, and customer sentiment and the reasons behind it. This makes the quality assurance process faster and easier, and helps the supervisor provide the proper coaching to the agent.
A patient is looking for information on a healthcare system’s website, but is unable to find what they need. A patient representative steps in after the patient engages them via chat. The agent’s copilot automatically identifies the patient and surfaces relevant information for the agent.
Most importantly, the copilot makes the transition to the agent simple by informing them what the patient’s issue is and where they’ve been looking for information so far so the patient doesn’t have to repeat themselves to the agent. It also surfaces general information on the patient like their name and other relevant personal information — without pulling in anything that would violate privacy rules. This provides the agent enough information to help efficiently and quickly, without doing any harm. Afterwards, it quickly and automatically summarises the chat and suggests a wrap-up code, saving the agent time and effort and allowing them to move on to the next customer who needs help.
One of the most important responsibilities a business has is the privacy of its customers, both on a moral basis as well as in terms of meeting the standards and requirements of law. Here’s how that applies to CX copilots, with the wealth of information that flows through it.
First and foremost, it’s critical to make sure that any CX copilot your business uses meets the requirements of all applicable regulatory standards in the jurisdictions in which you do business (GDPR, CCPA, etc.). Each has slightly different requirements, and you’re responsible for not just the location where you’re based, but anywhere that your customers are located. If you’re operating a multinational business, you’ll have more responsibility here than if you’re running a mom-and-pop retail shop.
In order to ensure data privacy and compliance alike, security is critical. When searching for a technology solution, its security should always be top of mind, and an AI copilot is no exception, particularly considering the volume of data it needs to have available to it in order to properly function. Ensure that the most stringent data protections possible are being utilised to make sure your copilot doesn’t become a security hazard.
Implementing
Once you’ve decided a CX copilot is right for your business, it’s time to integrate it into your overall customer service ecosystem.
First, determine your use case. Who is going to be using your copilot — agents, supervisors or both — and what will it need to be able to do? Once you know what it is you’re looking for, investigate potential vendors that can help implement a CX copilot, or that offer a native copilot as part of their overall tech stack. You’re not going to be doing this alone, after all; the copilot is a sophisticated piece of software and the less you have to ask of your internal IT team, the better.
There are also benefits to a native copilot rather than a standalone copilot. If your current contact centre software doesn’t provide a copilot natively you’ll need to make a switch, but a native solution is superior as the data the copilot needs to function is all being created and managed within the solution — meaning there’s no need for integrations, and context is maintained.
Finally, educate your team on what it can do and how it will help them (not replace them!) in delivering great experiences for your customers and your business. Operational efficiencies and time and costs saved help everyone. It can also help in journey management by helping you to locate points of friction as they appear. And once your copilot is in place and working, don’t just leave it to its own devices — measure how well it’s doing, assess where it could do better and iterate to improve performance.
Depending on who’s using it, there’s a wide range of KPIs you could use to measure the efficacy of your CX copilot. For customer support, there are several primary metrics you could be looking at. Average resolution time (ART) measures how long it took the customer to solve the issue, and first-contact resolution rate (FCR) shows how often a customer’s issue is solved the first time of asking. Together these help you see whether the copilot is making it easier for customers to solve problems.
On the agent side, next issue avoidance measures agents’ steps to prevent further issues proactively, average handle time (AHT) measures how long it takes agents to solve an issue from a customer reaching out to the end of the interaction, and after-call work (ACW) measures the amount of time an agent has to spend finishing up an interaction after the call or chat is over. These show how the copilot is helping agents to do a better job solving issues more quickly.
On the supervisor side, you’ll want to look at quality management administration costs, evaluation time — including time spent on training and onboarding — and, for global businesses, multilingual review time. Time spent reviewing interactions, efficacy of coaching, and long-term improvements in the customer experience from knowledge gained from the copilot are also KPIs that can be tracked.
Now that you know what you’re looking for, let’s find it. Here are a few of the things you should be focused on as you search.
The simplest option is to choose a fully native copilot, making integration a non-issue and scalability easy. But no matter what solution you choose, there are a few important things to keep in mind. The copilot needs to be able to draw data from all parts of your business, so it needs seamless access to that information. Also some providers offer flexible pricing models, meaning that you pay as you use the copilot, and can scale based on your needs without commitments and fees.
The copilot needs to be able to meet all of your needs, so make sure these are top of mind while you’re vetting potential partners. For instance, which languages do you need support for? An enterprise business may need a copilot to function in many languages, depending on who they serve worldwide. A smaller business may only need English and Spanish.
The copilot also needs to be able to function alongside the software you’re already using. This is less of an issue when the copilot is native to your contact centre software, but it still needs to be able to connect to the knowledge base you use, so that data can flow from the base to the copilot and vice versa.
In a lot of cases, your copilot will be ready to go on day one — at least for most of the needs it will meet. For the rest, your vendor should be able to help in this process, but your own inputs will be invaluable.
The KPIs mentioned earlier only really do you any good if you have a starting place to measure from. Before implementing your CX copilot, measure the metrics you’ll use to judge its success. This is your baseline that you’ll compare all future measurements to in order to see whether the copilot is helping.
A copilot runs on a knowledge base that needs to be continuously updated — so if your knowledge isn’t up to date, your copilot isn’t up to date. This doesn’t only apply to the knowledge that it triggers, but also the accuracy of summarisations, suggested wrap-up codes, insights and translations of calls. You have to ensure that you’re consistently monitoring and optimising to improve accuracy.
No matter your business or industry, there are going to be unique terms that your copilot needs to understand that it won’t know on day one. Things like brand or product names, service tiers and acronyms may mean something very different in the context of your business than elsewhere, so you will need to teach these meanings to your copilot. This will improve both transcription and summarisation, making your copilot’s outputs more helpful and easier to understand.
Optimising
Obviously, as with anything else you’re likely to face some challenges as you work to implement your CX copilot. Let’s take a look at some common issues you might run into and how they can be avoided.
Not every business has an unlimited budget for their customer support department, but that doesn’t mean an AI copilot is out of reach. Flexible pricing models are available that allow you to use what you need and pay only for that. And a copilot will save your business money by helping to make your human agents more effective and efficient, and thereby helping you keep the customers you have by reducing service wait times and more. The ROI of a copilot is huge, and you’ll see it quickly.
A copilot is a complex piece of software, and the complexity of effort required to integrate it into your legacy systems might be greater if those systems are older or more complicated. A good vendor will help you to make sure everything fits together properly, but if your tech stack incorporates particularly dated software, particularly if it’s housed on-premises rather than on the cloud, the best answer might be an update there as well. The best move is to use a native copilot that’s built into your contact centre software stack, so integrations aren’t an issue at all.
As we mentioned earlier, education is key here. There’s a lot of talk about AI but not a lot of real knowledge about how it works and how it can help, rather than replace, your existing human agents and staff. Make sure your team knows why you’re implementing a copilot, how it can make their day-to-day lives easier and what the future might look like.
Copilots already feel cutting edge, but considering how new the technology is, there’s plenty of room for growth. Let’s look at what’s next for copilots.
In addition to copilots for agents and supervisors, administrator copilots are coming, as well as potentially copilots allowing the ability to support even more roles within your business. These would provide the ability to construct workflows, and to suggest improvements generated by AI.
When your back- and front-office teams are siloed, it can create customer service issues. A copilot that spans that bridge helps to ensure seamless service. It streamlines workflows and ensures that the customer experience is fully supported and complete.
Customer journey management helps to manage the impact of your copilot — the copilot feeds data into journey management, and the copilot can be optimised by insights from journey management software. This fusion creates a cycle of continuous improvement. Integrating your AI copilot with your customer journey management software helps to understand each step of the customer journey, and optimise and improve the customer experience.
Conclusion
Copilots are a tremendously valuable tool for agents and supervisors, and for businesses as a whole. As businesses look for ways to leverage AI as it grows more and more useful, CX AI copilots are one extremely potent way to do so. And as they evolve, copilots will likely soon become a necessity for customer service professionals, not a luxury.
Frequently asked questions
CX copilots differ from other CX technologies designed to assist employees in part because it’s baked into the user interface for both agents and supervisors. They don’t have to toggle between screens to search for knowledge or surface intelligence about an interaction — it’s all there natively.
The main way a copilot works with other existing systems is by using them as a source of data. A CX copilot draws data from all the existing systems that your business uses in order to build as complete a picture of your customer as possible. That allows it to determine what information is relevant to a given situation, surface it and deliver it to an agent or supervisor. Copilots also create data of their own — for instance, summaries of interactions — that need to be available to your overall system.
Yes, it can. A copilot helps human agents do their jobs more easily, making those interactions faster, more efficient, and more accurate to the needs of the customer, thereby improving the customer experience. It does the same for supervisors, taking rote tasks out of their hands and making the evaluation process easier — and training and coaching more effective.
A CX copilot can help practically any business that wants to upgrade its customer service capabilities. A flexible consumption model can be implemented to use only what your business needs, helping small businesses adopt tools like copilots without massive expenses and usage commitments so they can scale as necessary. In most cases, the cost savings of AI copilots literally pays for itself.
Your copilot vendor should be able to assist with any issues you come across, from installing and getting the copilot up and running, to training the copilot, to troubleshooting and even potentially helping with team adoption and teaching agents and staff how — and why — to use the copilot. This is all easier if your copilot is native to your contact centre software — support will be included.
Ready to see what an agent or supervisor copilot can do for your business? Reach out to Genesys, and see how we can help you take your customer service abilities to the next level.
Thank you for your interest.
We’ll contact you directly to set up a date and time that works with your schedule.
