The AI Customer Service Blueprint

The easiest way to think about integrating AI customer support tools and apps into your business is to think of it like building a house.  You need a good foundation of support content to help your customers succeed, a main floor for boosting customer retention, and an upper floor of customer expansion. This AI Customer Service Blueprint breaks down how to integrate AI for customer support into three main levels. First, we’ll lay down the foundation, then set up the main floor, and finish with building out the upper floors.

This blueprint is the first exploration of how the next generation of AI tools can be integrated into your existing customer service interactions. It includes:

• A practical blueprint anyone can follow to future-proof customer support teams with AI 
• Interviews and quotes from leaders in AI, customer service, and AI customer support
• Recommendations for AI tools with clear instructions on how to use them
• Action-oriented takeaways and tips you can implement today 

The Foundation: How to start using AI in your customer service

Build your house right: AI in customer service starts from the ground up 

You wouldn’t build a house on a swamp, would you? You’re not Shrek, so the answer is no. 

You need a solid foundation to set up a successful building. The same is true when setting up your customer experience strategy. You need to establish a strong foundation for adoption; then, you can guide potential customers seamlessly to the next phase in the customer journey. 

To do so, you need to integrate AI into your strategy. You likely already know this, or you wouldn’t be reading this blog post. But if you’re still on the fence about AI in customer service: know that without AI, the average number of support tickets an agent can handle is 21 per day. With an AI-powered messaging tool, that number jumps to 46. Mind = blown.    

Integrating AI into your customer support foundation will result in better adoption rates. In this chapter, we’ll show you how you do so through three main methods: knowledge, self-service, and proactive support. 

The Foundation will cover:

• Generative AI and knowledge bases
• Using AI to build self-service chatbots
• Using AI to offer proactive support

Let generative AI do the heavy lifting for your knowledge base

Knowledge is power, and AI is all-knowing (sort of)

Want an efficient customer experience? Then you need to lay the groundwork for a customer knowledge base, much like your house's concrete layout. 

A knowledge base is something — or someone — customers can pull information from. People want to know about your company, products, or services, and these curiosities must be answered before they buy in. 

A knowledge base can help your support team answer questions quickly and efficiently, or it can exist in a self-serve manner, where your customers quickly get the answers they need to adopt your product or service. 

Now, AI tools can automate this process and provide accurate data — when set up correctly. You can think of these AI tools as the construction workers doing the heavy lifting, like pouring the concrete of your knowledge base. 

The takeaway: You need a knowledge base. AI can help you create one.

Here’s how to use AI to build a solid knowledge base for your customers 

3 steps to make it happen: 

1. Get a generative AI tool to create a knowledge base
2. Get a human to check over the output for accuracy 
3. Integrate AI-enabled search capabilities so your customers can surface relevant information — faster

“AI absolutely can be useful to solve easy, simple questions. Agents when they are trying to answer a customer problem. They’re on the spot. They’ve got to find an answer, and sometimes they have to go dig for it. They have to find relevant content that the internal team gives them. And sometimes those articles are outdated. Sometimes there’s just a gazillion different resources for them. And meanwhile, the customer’s on the line waiting for an answer. That’s where AI may be able to empower the agent to service the customer faster.”
- Stacy Sherman, Founder, Doing CX Right

What is generative AI? 

Generative AI is a type of machine learning involving the creation of systems that can produce new and original content. It’s the tech behind software like ChatGPT, DALL-E, and MidJourney. Gen AI can converse with humans naturally; it’s come a long way from the clunky, rules-based first-gen bots we all know and hate. 

How to use a generative AI tool to create your knowledge base 

5 steps to make it happen: 

1. Grab a generative AI writer, such as Writesonic, or a dedicated knowledge base tool like Document360. 
2. Give your writer as many detailed and specific instructions as you can. Include any relevant, existing material here, such as your FAQ pages. Provide brand voice prompts to your tool. 
3. Generate your knowledge base. 
4. You (the human in charge, remember?) need to edit the knowledge base, looking for any irregularities, errors, or missing pieces of information.
5. Save your knowledge base by uploading it to the relevant area of your site or Document360 workspace. Treat yourself for doing a good job (you deserve it). 

Pro tip: Other generative AI software can pull updated information from your web pages as a knowledge base. Keep the information in your help center, FAQ page, or other company pages up-to-date.

What is AI-powered search?

AI-powered or AI-enabled search learns from user data to provide accurate and relevant search experiences. To deliver refined answers, the AI will pull from the user’s data, such as their search history or past purchases, and use deep learning to understand search intent. 

Why is it important for AI in customer support?

Would you buy a house with an incredibly confusing layout? Nope. It’s why architects don’t get inspo from corn mazes. 

88% of consumers are unlikely to re-visit your site after a bad experience. Today, customers expect to be able to find what they want quickly. Delivering fast and accurate results to website visitors keeps them engaged.

How to integrate AI-enabled search capabilities

You have your knowledge base created and loaded, right? Then, find a reputable AI-powered customer service solution. Brands like Elasticsearch, Solr, and Sphinx have remained fan favorites for the past decade.

A brief breakdown of each tool: 

Use AI to build self-service chatbots for simple questions and requests

The robot workers happily do your bidding (better than before) 

Want to know the secret ingredient to success? Put people above all else.

That’s it. Put people before profits. Prioritize the user experience over trendy design elements. Put people in your dead-center focus, and you’ll organically build a thriving business because people will buy into it. People buy into themselves.

When you design your customer service strategy with people at the center of it, you’ll naturally gravitate toward efficiencies to improve the experience, like automating your communication through chatbots. Chatbots are like little robot workers putting up support beams for your house’s foundation. 

It’s much easier to integrate a chatbot than ever before. These bots are trained to handle your customer support requests at scale. They can crawl your knowledge base for relevant information while using machine learning and natural language processing to have human-like conversations with your customers.

Don’t get it twisted: These bots will not replace your staff. They lack a personal touch and the ability to understand situational issues. Rather, bots free up your team to do strategic, meaningful work. They take care of the monotony while you build relationships with your customers, plan for the future, and troubleshoot complexities.

The takeaway: Self-service chatbots can automate simple customer support requests, freeing up your staff to do strategic work. Consumers prefer chatbots for low-complexity tasks but humans for high-complexity tasks. Your customers want a seamless AI customer service experience.

Here’s how to build self-service chatbots for simple questions and requests

Find a chatbot with positive reviews and the capabilities you need, such as a generative AI chatbot’s ability to prompt email copy or a conversational AI’s ability to hold an organic conversation with your customer. Netomi, for example, is a fan favorite for its ability to use generative and conversational AI.

Install your chatbot, then test its conversational abilities. Be sure to optimize for success continuously. 

3 tips for success: 

• Take a look at your user journeys to find areas of friction. For example, a common customer complaint may be long waits on support tickets. Install a chatbot or AI-powered virtual assistant to automate support. 
• Be sure to program your chatbot to uphold your brand guidelines during conversations. 
• Use both generative AI and conversational AI chatbots to support your needs.

“AI can assist agents in handling complex inquiries by providing real-time information and recommendations. AI-powered virtual assistants can help alleviate long wait times and provide more consistent and personalized service.”
- Brad Cleveland, Author, Contact Center Management on Fast Forward

A Netomi chatbot takes off with WestJet

Here’s an IRL example of a chatbot working for the man 

WestJet, master of the skies, wanted to provide their customers with more communication tools. They brought Netomi on board to provide conversational AI for customer service through a chatbot named Juliet. 

WestJet’s co-pilot, Juliet, took off on Facebook Messenger, WhatsApp, and Google Assistant. The little-bot-that-could helped travelers with simple tasks like pre-booking, booking, day-of-travel, and post-trip support.

‍

After integration, Juliet resolved roughly 700 highly repeatable queries! She also knew her limits and escalated complex issues to human reps. Juliet automates 87% of all WestJet support tickets without human help (up from 24% when she first started, the clever bot). It’s one of the uber-successful examples of AI in customer service.

Pro tip: WestJet programmed Juliet to specify “ask me a simple question” in her chat. This sets the customer’s expectations from the get-go. 

The WestJet-Netomi case study is a prime example of both chatbot limitations and capabilities with machine learning. Find yourself a chatbot that can learn to solve and seamlessly escalate complex issues to your customer support team before your customers get frustrated.

“Here’s the mistake that companies make: Regardless of the current state of how great generative AI is, they relied on self-service to replace customer support. Human to human interactions. Big mistake. I don’t care how good the machine is. In its current form, and in the near future, it will not be able to eliminate every concern and question."
- Shep Hyken, Award-Winning Author & Keynote Speaker

Use AI to offer proactive support during customer onboarding

Want to do a magic trick? Solve problems before they happen with generative AI 

Solve a problem before it becomes a problem, and you’ll have happy customers. 

Today, customer experience rules all. We expect products to be designed with users in mind and presented with simple software that just works. We subconsciously believe that customer service should solve our problems before we even know they exist. 

We crave personalized solutions, especially during the onboarding period when we have to spend effort learning about a product. 

Today, generative AI tools can craft personalized emails based on a specific buyer persona or a use case. Giving people context before they ask for it will raise your production adoption rates and turn your customers into brand ambassadors. 

The takeaway: Use generative AI to proactively solve problems for customers to develop brand loyalty. 

What is proactive support?

Providing proactive support means being one step ahead by solving issues before they appear. Anticipate potential problems and proactively suggest solutions that will save time and resources. 

Benefits include fewer support tickets, improved satisfaction levels among existing customers, and increased loyalty towards the company.

How to use AI to offer proactive support during customer onboarding

There are tons of ways you can use AI to evolve your customer service from a reactive task to a proactive driving force. 

Here’s one example of better onboarding using Magical’s AI Assist in six easy steps: 

1. Identify a trigger point for when you want to offer proactive support (for example, when a customer has signed up for your software but doesn't use it for three days).
2. Download Magical.
3. Use AI Assist to draft multiple outreach email templates based on different personas who might hit this trigger point (for example, recruiters, salespeople, and marketers might use your software).
4. Save the best templates in Magical as shortcuts with an easy-to-remember trigger, such as -stuckrecruiter.
5. Share these shortcuts with your team, so they can use them with any customer who hits the trigger point.
6. Before you send the email, Magical will instantly personalize the template with the customer's information (like their name, account number, company name, or persona). All your team has to do is type in -stuckrecruiter and give it a once-over before hitting send.

Boom, you suddenly have a real-life person offering customer support instead of an automated email. And the best part? It takes your team less than 30 seconds to fire off a message like this. 

Pro tip: Offer this support over your in-app chat or as a LinkedIn message to avoid your message landing in the no man’s land of inbox spam. The cool thing about Magical is that these shortcuts will work anywhere — well, maybe not on Tinder, so you’ll have to use your templated pickup lines elsewhere.

“The problem with current efforts is that we so often wait until the process breaks down and just hope customers contact us instead of bolting for our competition. AI can help us proactively understand where customers are in their journey and reach out at those key journey points before customers become aggravated.”
- Jeremy Watkin, Director of Customer Experience and Support at NumberBarn

The Main Floor: How to use AI to improve customer success

Build bridges, not walls: AI's role in retention and constructing solid customer relationships 

A company without a proper retention plan is like a house without a main floor: You’re going to step inside the front door and fall right through. 

Retention is the lifeblood of your company, and right now, there are more eyes than ever on your performance regarding happy customers. Keeping existing customers loyal and satisfied is a key factor in business growth and profitability. 

AI-based customer support makes retention easier and more efficient than ever before. 

In this section, we will cover: 

• Documentation and customer data
• Emails and support tickets
• AI for coaching
• Predictive retention

AI can help you improve your documentation & customer data 

AI for customer data: The plumbing and windows of your main floor 

The better you know someone, the easier it is to keep them around. Happy marriages and customer-company relationships alike. 

Accurate data allows you to target customers with relevant offers, provide personalized customer service, and get specific with your marketing efforts. Data quality allows you accurately segment customers and to use their preferences and desires to retain them. 

You can use AI to improve your documentation of customer calls. Think of these AI-powered tools as the plumbing of your main floor: It lets information flow through your house.

Now, note-taking for customer data can be automated using AI tools. These tools improve the accuracy of data you have on each customer; instead of their personal data being translated through your lens, it's created by an unbiased system. It’s like the windows and doorways of your main floor, giving you through sight lines and visibility into your environment. 

Plus, you can use AI to improve and expand on your common customer note templates. 

You can pass their full history off to whoever is next in the interaction chain, giving your colleagues (or your future self) the data breadcrumbs needed to understand the customer best.

The takeaway: Use documentation AI to get to know your customers accurately and thoroughly without wasting time and effort. 

Three ways you can use AI to better your customer data 

“I think there is an interesting opportunity for us to effectively scale the delivery of knowledge to our customers and to our agents. I think there’s an opportunity for us to scale the accuracy of getting information out to people. And I think there’s also an opportunity for us to scale the way we collect data and leave it as bread crumbs for whoever’s next in the interaction chain so they have a better source of truth on the customer.”
- Justin Robbins, Evangelist, 8x8

Mission Lane waves a magic wand and saves 25,000+ hours 

A customer support team discovers the magic of AI

Mission Lane’s customer support team was sinking under monotonous customer demands. A financial technology company offering tools and support for millions of Americans, Mission Lane had an endless queue of support calls and emails coming in on any given day.

Emely Leal, a Bilingual Account Coordinator, was struggling with the sheer volume of it. Luckily, she not only discovered how Magical could save her thousands of hours but also, that many of her teammates were already using the app. They pooled their popular text expansion shortcuts using the Magical team Mission Lane had set up — making the absolute most out of their shortcuts. 

Now, Emily uses Magical to take notes on calls so she doesn’t have to split her focus, type out support emails in seconds, and for responses to common questions or queries. Magical’s spread throughout Mission Lane, with dozens of colleagues sharing prompts. Their productivity has skyrocketed, and they have the timeframe stats to prove it. 

Besides winning bragging rights for being the most productive department, the Mission Lane customer support team has saved a collective 25,000+ hours of work by using Magical. 

AI to improve emails and support ticket workflows 

Get support tickets handled faster than you can say, “I hate support tickets!”

AI can sift through your mountains of emails and tickets faster than a cheetah on caffeine, saving you time and boredom. You can use it to deliver personalized, accurate responses because it learns from past conversations (unlike your ex). 

Now, AI provides intelligent ticket sorting, such as by sentiment analysis (address the Karens first or last, your choice, boss). It can also provide predictive text, saved templates, and personalized responses. 

The responses are instant, consistent, and within your brand voice, so you don't have to worry about your support team having a bad day or forgetting to take their meds.

Plus, with AI, you can say goodbye to human errors like typos, miscommunications, or accidentally copying, pasting, and sending a love letter as your support ticket response.

The takeaway: AI saves you time and effort by automating emails and supporting ticket workflows. 

2 AI tools and how you can use them to improve your emails and support ticket workflows

Magical

Say goodbye to hours spent crafting the perfect response to every single email. This magical tool quickly generates personalized responses that you can save as templates to use again and again.

You can use Magical AI Assist to generate email copy templates, or pull from these prime customer service responses, then save them as shortcuts through the Magical app. For example, a bug in your system may be causing an onslaught of the same email over and over again. Yikes. 

With Magical, you only need to type out and save one reply. Tag it with a specific trigger, like -nastybug, then simply type your trigger into your email and watch Magical fill in the rest. Plus, the tool’s Variables feature lets you customize each email with the help of tags like “First Name.”

Ultimate AI 

The ultimate ticket-handling virtual agent, Ultimate AI, automatically tags and routes requests. This bot knows precisely what your customers are after, so every ticket gets slotted into the correct department’s inbox. Ultimate uses “intelligent ticket triage” to label, merge, and pre-fill inquiries for you. 

“AI can investigate prior interactions and the current interaction in real-time, help the agent better understand background and context, and assist them in tailoring services that improve the customer experience.”
- Brad Cleveland, Author, Contact Center Management on Fast Forward

Use AI to defuse Karens with live phone call coaching

Serve rude customers a human-AI customer support sandwich 

AI-powered coaching or translation tools are the expert contractors of your home build. These guys point you in the right direction and can stop you from making a costly mistake. 

With AI, you can provide real-time feedback to agents during customer calls, helping them to improve their performance and provide better customer experiences. These aren’t just simple AI customer support and assistance tools; but rather investments into your human team. AI can coach, train, support, and help mentor your agents in real-time. 

These tools are helpful for both your customers and your agents’ learnings and can positively impact your brand reputation. By calming angry, upset customers and turning bad situations into good, you can inspire brand loyalty.

All of that being true, you won’t want to rely entirely on AI for customer service coaching. Be sure to give your agents the support and training they need to battle the mega-Karens. 

The takeaway: AI call coaching for your customer support staff helps retain customers but also provides professional development for your staff.

About AI for live coaching 

When using live coaching AI for customer support, you want to be sure you’re paying close attention to the data it produces. Powerful AI customer care tools can pull out insights from conversations, give you a predicted Net Promoter Score, and measure the customer experience of calls in real-time. 

The customer data you can acquire is invaluable. It can influence business decisions and provide actionable insights. Not something to sleep on.

As far as choosing software goes, Cogito is a fan favorite. We love it because not only does it measure customer sentiments, but it also gives leaders a look into team happiness-stress levels. Cogito dashboards show the entire team’s activity, can track employee fatigue, and alert supervisors to early signs of burnout. 

How to use AI for live coaching 

1. Data collection: Collect relevant data to train the AI system. This could include previous coaching sessions, performance metrics, or any other data that can provide insights into the coachee's progress and challenges.
2. Develop the AI model: Build or customize an AI model based on the selected tools and data. This may involve training the model using machine learning techniques and programming algorithms to understand and analyze the coaching data.
3. Integrate AI into the coaching process: Integrate the AI model into the live coaching session. This can be done in a few different ways, like using chatbots, voice assistants, or real-time feedback systems. The AI system can analyze the coachee's responses, performance, or behaviors and provide immediate feedback or suggestions to enhance the coaching process.
4. Monitor and refine: Continuously monitor the performance of the AI system and gather feedback from coachees and coaches. This feedback will help identify areas of improvement and refine the AI model to serve the coaching process better.
5. Don’t forget the human touch: AI has limitations. You’ll want to be sure you have a human agent involved in the coaching process. Your human coaches should still actively engage with coachees, getting feedback on the sessions and ensuring the AI gives the right advice. AI is best used as a supportive tool, not a replacement. 

“How do you think the customer would rate this call based on just a transcript? The computer comes back and says, they’re going to give you a net promoter score of nine out of 10. Guess what? They were right. Like almost every freaking time. Think about how once again as a training tool, you can learn what’s working and what’s not working. AI can instantly coach, train, support, and help mentor your agents.”
- Shep Hyken, Award-Winning Author & Keynote Speaker

AI and customer service live coaching and accessibility 

One of the most interesting (at least to us digital nerds) facets of using AI live coaching in customer service is how it makes the industry more accessible. 

Live coaching can act as a how-to guide for neurodivergent individuals who have difficulty understanding emotionally-charged situations or misjudge social norms. Or those who find it stressful to know how to react to emotional cues in real-time. It’s one of the most accessible benefits of AI in customer service.

Clear instructions on what to say and how to respond to customers in an elevated emotional state can be a welcome relief for neurodivergent folks. Neuroinclusion benefits both the employee and the employer; these individuals can be celebrated for the diverse skills and abilities they bring to the company instead of being passed over due to an unsupportive system. 

“The machine is listening to the call and then saying to the agent, “The customer is upset. Slow down. You’re talking too fast. Give them a chance to vent. You’re interrupting them.” It’s coaching while you’re on the call, in real-time.”
- Shep Hyken, Award-Winning Author & Keynote Speaker

Using AI for live translating and localized accents 

It’s no secret people are drawn to what they feel is “normal.” We gravitate to what we know because when something’s familiar, it feels safe. 

Software companies like Sanas have capitalized on those proclivities with AI that gives your speech a localized accent. So, for folks in call centers overseas, they can transform their local accent into the caller's local accent with no down-time. 

Machine Translation (MT) isn’t new technology — we’re pretty sure Google Translate has been around since the dinosaurs. But recent AI-related advances (like Neural Machine Translation) have made it much more user-friendly. 

Meta, for example, built a voice-translation machine for translating oral-specific languages for the company project Universal Speech Translator. You can see it (here or below) for the spoken language of Hokkien. 

Best practices for integrating human-AI customer support 

When it comes to frustrated customers, you can never be too prepared. You’ll want to set your team up with human and AI-powered support. Speaking of which, here are 8 human-to-human tips for dealing with Karens. 

AI has its limitations. A Magical survey reported 72% of customer support professionals say they don’t think AI can provide adequate customer service on its own. Your jobs are safe… for now. Just joking. No one’s threatening your job — we pinky-promise!

The more support structures you can put in place for your customer service team, the happier your customers will be. Developing standard responses is one way you can provide support. If your customers are angry and your AI coach isn’t delivering, you need a backup plan. Managers can give standardized answers that are available at their agents’ fingertips for these situations.

And no, you don’t have to provide those answers manually. Using Magical, you can create customer support templates and drop them into chats or your agents’ scripts with just a few keystrokes. You can let your entire team access these resources through the Magical app.

“In terms of impact on agents, AI is not meant to replace human agents. AI-powered tools can help reduce the stress of pulling the right information together, enabling them to focus on customers and making the work more enjoyable.”
- Brad Cleveland, Author, Contact Center Management on Fast Forward

Use AI for predictive retention (AKA AI fortune-telling)

It’s not creepy — it’s just knowing everything about your patterns to predict your future actions 

Predictive analytics, sentiment analysis, and deep learning AI-powered tools are like the rebar, siding, and support beams you build into your home structure. They are the scaffolding that prepares your house for future storms. 

With predictive analytic tools, you can use AI to identify patterns and proactively address common outcomes to those patterns. For example, if a customer hasn’t activated an account in a certain time frame, the likelihood of them churning may be high. You can use predictive analytics to tell you the likelihood, then send the customer an activation incentive.

This data will help you anticipate customer needs, prevent attrition, and increase retention. 

Don’t sleep on using AI for proactive support; it’s a trend showing no sign of slowing down, and your competitors all know it. 

The takeaway: Predictive analytics tools pick up on patterns, then give you actionable insights to increase retention.

3 ways to use predictive AI for retention 

1. Stop churn & reward loyalty 

Predictive analytics tools like Pecan can turn your data into actionable insights. With a behind-the-scenes algorithm, it can predict customer behaviors, such as churn, loyalty, and lifetime value. Then, you can focus on stopping churn by sending personalized emails (oh hey, Magical!), booking a call (don’t forget your live AI call coach), or reaching out within your app or platform. 

Churn is one thing, but you’ll also want to reward loyal customers. Find the ones that Pecan predicts to have a high lifetime value score and offer them referral codes, discounts, or upgrades. Be sure to ask these folks for a review or shout-out on social.

2. Create new products or services that meet customer needs 

You can also use the data from your predictive analysis tools for proactive customer support. Is there a product that consistently has lower satisfaction ratings? Then it’s time to innovate. Pull data from your customer support calls to see the common complaints and address them in an update. 

Pro tip: Ask your customers with high loyalty rates to Beta-test your product pre-release. You’ll gain valuable insight, plus you’ll be acknowledging and cementing those customers’ loyalty. 

3. Tailor marketing campaigns to customer segments with personalized marketing 

Targeting customers with personalized offers increases the likelihood of retention and brand loyalty. To do this, you can use predictive analytics to identify customer trends or preferences. Then, use that data to craft a hyper-targeted marketing campaign tailored to a specific customer persona. 

Here’s what this might look like: Let’s say you sell outdoor gear. AI tells you a cohort of your customer base buys gear they can use in water, like water-proof flashlights or dry bags. These are loyal customers, likely to buy more gear. 

Create a campaign showcasing people on fishing trips, kayak adventures, or canoe rides. Then, target your audience with a personalized offer, like, “Tag us in a photo of you using your gear from CampGearForever on your next outdoor adventure and get a free water-proof radio!” 

Predictive AI has told you who these people are and that they will likely be repeat customers. You can encourage retainer rates with a campaign that rewards them for using your product. Plus, you’ll be collecting a swath of user-generated content you can use on your social channels!

“If you track what a bunch of these customers are doing, you can start to predict what sales are going to be. What products people are going to need and buy. And if you train your support people the right way, they can engage the customer. Earn the customer’s trust. And ethically, cross-sell or upsell… AI isn’t just a great support tool. It’s a great sales tool.”
- Shep Hyken, Award-Winning Author & Keynote Speaker

The Upper Floor: How to use AI to raise the roof on expansion

Building customer service skyscrapers: Revolutionizing customer expansion with AI-scaffolding

Pretend you’re a wildly rich king in the 18th century who wants to meet God. What are you going to do? Build your castle as high into the sky as you can, obviously. This chapter is all about expanding the upper floor of your customer service house sky-high. Kiss the ring, and let’s get building, peasants! 

AI won’t become your best salesperson, but it will help your best — and worst — salespeople reach new heights. Consider AI for expansion in regard to the customer experience. You can use AI-powered tools to better support your customers by understanding where they’re at in their journeys and then ethically upselling and cross-selling. 

Even beyond direct communication, AI’s capabilities stretch into other areas of your business closely linked to customer support, like marketing. AI-powered marketing applications, for example, are basically the power tools of your next home build. 

Within the Upper Floor you will find: 

• Researching customers
• Upselling opportunities and outreach

Using AI to research your customers to know when the time is right for expansion

Use AI to turn your customer journey into a hallway free of obstacles 

We briefly discussed using AI for customer data in the previous chapter on retention, but retention and expansion are closely connected. One could say it’s as easy to get from one to the other as climbing a staircase between floors. See what we did there? And yes, we’re still on the home metaphor.

Your customer data can tell you where a customer is in the customer journey and whether or not it’s the right time to schedule a call or reach out to see if they need more products or services. 

Customer journey management can be taken a step further with AI and into the world of customer journey orchestration, making it more about where the customer is from moment to moment. 

The takeaway: AI can give you insight into when the time is right for expansion. 

What is customer journey orchestration?

Customer journey orchestration is an advanced form of customer journey management. Customers are unique; customer journey orchestration considers customers as individuals (vs segments). It’s different from a customer journey map, which is a static asset that does not see engagement. Traditional customer journeys can’t adapt on the fly to evolving situations. 

With orchestration, agents use data to accurately predict a customer’s next step, often in real-time. Then, they can ensure a customer’s experience with the company is tailored to that customer. Journey orchestration is all about the holistic customer experience. Sound like work? It’s easier with AI. 

“AI will enable organizations to better understand their customers’ existing emotions and to build interventions to activate positive emotions such as confidence and trust while minimizing loyalty-destroying emotions such as frustration and annoyance. An emerging capability is that of journey analytics and orchestration, identifying blockers in the journey and identifying next-best steps for individual customers. Adding the identification of emotional markers will make this capability even stronger.”
- Jim Tincher, Author, How Hard Is It To Be Your Customer?

How to use AI for customer journey orchestration 

Here’s an example of journey orchestration IRL: A customer buys a cellphone from your company. Unfortunately, the customer has received a phone with a defective charging port and has opened a support ticket. 

In the meantime, your marketing team issues a sequence of onboarding emails to that customer, including recommendations for a lightning-fast charger. The customer is understandably frustrated by these emails. 

With journey orchestration, all teams have clarity into the customer’s experience. Once the ticket is raised, the onboarding sequence of emails is paused. Instead, the ticket is resolved, the customer gets their product, and the emails that follow are tailored to the customer’s experience. Maybe you send them a free lightning charger with a “We’re so happy you can use this now!” note attached. 

The sequence and solution drive trust and maximize your customer’s lifetime value; they’ll remember how intuitively you addressed the defective product and the personalized level of service that followed. Retention leads to expansion. 

AI-powered customer journey orchestration tools 

AI tools like Medallia Experience Orchestration (MXO), a customer journey orchestration platform, can help. MXO allows you to analyze customer journeys at scale, then instill tailored orchestration at every touchpoint — online and off. 

You’ll be able to automate the analysis of customer journeys and build hyper-specific audiences. You can also automate the analysis of data sets and predictive models from customer insights based on real-time journey behavior. 

It’s important to remember that all AI customer service tools are support for — not replacement of — human teams. Automate as much as you can in your journey orchestration, but be sure to keep a hand in the process. 

MXO x Lego: Building customer journey orchestrations brick by byte 

Lego’s among the many companies using AI for customer service. They use the Medallia platform to closely scrutinize sentiment in the customer experience. Lego says the tools allow them to “react quickly with intelligent action.” 

MXO focuses on moment-to-moment customer experiences, which helps conglomerates like Lego to have a one-on-one relationship with their consumers. We love it when a massive company like Lego can reach individual customers where they’re at, listen to them, and adjust. Customer journey orchestrations put the customer experience firmly in the driver’s seat. 

Installing a Medallia-powered customer insights program allowed Lego to bump up their Net Promoter Scores and, as a result, expand revenue both in brick-and-mortar stores and online. 

“You can’t have a cookie-cutter approach to customer experience ever. You can’t have a cookie-cutter approach to agent success ever. It has to be personalized and customized to the needs of both.”
- Stacy Sherman, Founder, Doing CX Right

Using AI to trigger upsell opportunities and outreach makes for a better customer experience in the sales funnel

Upselling and cross-selling your products or services to existing customers will drive cost-efficient revenue growth. It’s less expensive and time-consuming to sell to customers who have already bought into your brand. 

Predictive analysis AI can now understand when a customer is naturally interested in purchasing add-on services, products, or accessories. The machines can tell you when a customer is ready to make the leap to premium-grade items, too.

Timing is everything with sales. When your customer is in the right place in their journey to be upsold, and you can do it without them having to ask, you’ll make the customer sales experience intuitive and smooth. 

The takeaway: Predictive AI can empower your sales team to know when a customer is ready to be upsold, providing a better customer experience in the sales process.

How to use predictive analytics to know when a customer wants to be sold to 

Predictive modeling is like having a crystal ball that tells you when your customer is ready. Some AI models will give you data reports of buying history and patterns that can inform you of when a customer is likely to be ready. And some models will give you actionable insights pulled from the customer data, like “Peter is likely to purchase a premium subscription if you proposition him within the next 48 hours.” 

Predictive modeling works best to recognize complex patterns in customer data. Download a platform like Pecan to have it do the heavy lifting for you. You simply need to feed your historical and customer data into the machine. Some platforms (like Pecan) will pull public sociodemographic information in for additional predictive power. 

Once your data has been processed, the platform can spit out a predictive model for you, including customer upsell likelihood. Then, you can adjust your selling strategy to match what the AI thinks is best. 

Pro tip: Personalized, customized cross-sell offers have a much higher likelihood of success and generally better your customer experience. 

With AI’s capabilities, you can offer customers deals or promotions that are likely to get them hooked. It's like shooting fish in a barrel, except instead of fish, it's customers. And instead of a barrel, it's your sales funnel. It’s like shooting customers in your sales funnel, which should not be taken literally.

“What’s really cool is that the AI could not only answer the question the customer had but would also identify the type of customer. So the agent would be told, why don’t you spend an extra moment and tell the customer about this so they don’t have to call back (because they’re going to)? And not only that, the AI was able to say this customer is just like 2,000 other customers. Not only do we know a question, but we also know what product they’re going to buy next. So why not try to sell it to them today.”
- Shep Hyken, Award-Winning Author & Keynote Speaker

Your Construction Strategy: 3 ways to add AI to your support team 

Action without strategy is a good way for your house to collapse 

You’ve got your house blueprint all mapped out. Your next step is to start construction! Be sure your toolbox is stacked (see the next section for a summary of the tools mentioned throughout this blog post). Then, begin building with this construction strategy in place. 

Ready to break ground? Let’s go! 

1. Get buy-in from your team and train them to use AI

The key to getting buy-in from your team is to emphasize that AI is not here to replace them — it's here to support them. You can tell your team this until you’re blue in the face, but the best way to get them to believe you is to show them. 

Make sure you know how to use the tools yourself, then give your team a demonstration. Show them not only how to use the tools but how the tools will benefit them. Clearly demonstrate how the tools can take on their most hated, repetitive tasks, like answering FAQs, responding to support tickets, or processing refunds. 

Once your team sees how AI can lighten their workload, they'll be more likely to embrace the technology.

Make sure your team is comfortable using the tools

Of course, training is also crucial to success. Make sure your team is comfortable with the new tools and understands how to use them effectively by asking for feedback regularly until adoption is widespread. Provide plenty of training opportunities, and be patient if your team needs extra support during the transition period. Learning new things can be tough! 

“Do not think for a moment that [AI] replaces the live agent. This supports the agent taking care of many issues that that the agent does over and over again, that they no longer have to do so they can support customers who have bigger, harder, more difficult, complicated issues.”
- Shep Hyken, Award-Winning Author & Keynote Speaker

Action items: Download your tools and get comfortable using them in the same way your team will. Schedule a team-wide demonstration where you clearly outline the benefits the tool brings. Have open-office hours dedicated to supporting the tool’s adoption and request feedback from your team. 

2. Make sure to set or adjust the right OKRs

Objectives and key results (OKRs) are crucial to measuring the success of your AI implementation. Make sure you're monitoring the right metrics, such as tickets handled without representative intervention, customer satisfaction/NPS score, and customer retention/churn. 

By setting clear goals and objectives, everyone on the team will know what they're working towards and can make adjustments as needed.

Start with small goals you can easily achieve 

Remember to start with small, incremental steps. Implementing AI in stages allows for testing and adjustments before rolling out the technology to the entire support team.

“I have the mindset to walk before you run. You need to implement AI in small incremental steps to ensure it’s effective and well-received by your team and your customers. Make sure you define clear goals and objectives, and everyone is marching to the same song.”
- Stacy Sherman, Founder, Doing CX Right

Action item: Before you implement AI tools, set your OKRs. Knowing your industry benchmarks can help you to set SMART goals (Specific, Measurable, Achievable, Relevant, and Time-Bound). Then, share your defined goals with your team so everyone is pulling in the same direction. Measure your progress and use your data to adjust your business decisions as you gain clarity. 

3. Master your customer service prompts

When it comes to customer support AI, the quality of the prompts makes all the difference. It’s so important, in fact, that it’s earned its own title: Prompt engineering.

Prompt engineering is the process of creating high-quality prompts to guide AI tools, especially language models. Prompt engineers masterfully manipulate AI with strategic input. Your output will only be as good as your AI tool and your prompt. Here are a few you can steal: 

Generate a Customer Sentiment Analysis 

Grab a segment of your brand’s customer reviews to check how people perceive your products or services. This can give you valuable insight into areas your brand excels and opportunities for improvement. 

We chose ProductHunt and copy-and-pasted the top customer reviews into Jasper.ai. We used the prompt, “Give me a customer sentiment analysis on the tool Magical AI. Read through the customer reviews, look for patterns, then give me the average sentiment levels and provide actionable insights I can take to better the tool.”

We included six reviews, one of which mentioned they hoped the pricing would not rise (from free). The output was this: 

Creating a Knowledge Base 

Using generative AI to create a knowledge base is a pro-time-saving move. We used Jasper again and typed in this prompt: “Create a knowledge base for the tool Magical AI. The knowledge base should answer frequently asked questions, give useful information on how to use Magical, and include everything a customer would need to navigate a product and information to help them self-serve if they run into any issues while using Magical.”

We also copied and pasted as much information as possible on the tool. The voice was a little off, so we recommend experimenting with the voice and tone prompts until you find the right ‘tude. Here’s the output: 

Formulating a Customer Feedback Survey 

Customer feedback surveys are great for understanding the sentiment behind your NPS. You’re going to want to send these out to gather data and feedback on your product and search for opportunities for optimization. 

Here’s the prompt: “Create a customer feedback survey designed to elicit honest feedback. The survey will be for an AI-powered tool called Magical. Magical promises to automate your most hated tasks, such as writing messages from scratch, updating forms, and filling in customer data. We want to know if there are areas we can improve and which areas we are excelling in. The survey should look for ways to optimize Magical and potentially use feedback in marketing campaigns.”

Here’s the output: 

Action item: Steal the prompts above and edit them for your personal use. 

10 AI tools you need in your customer support toolbox

There are tons of customer support AI tools out there for you to peruse, but the ten we’ve mentioned in this blog post are listed out here. Download these tools today for a fully kitted-out toolbox for your customer support home build! 

1. Cogito 

Cogito helps you sound smarter on phone calls with AI-based live coaching.

2. Magical AI

Magical takes care of all the boring work stuff you hate doing — from automated form filling to hassle-free messaging.

3. Sanas AI

Say goodbye to phone call struggles with Sanas AI, which uses AI to change your accent to your caller’s local one, allowing them to understand you better. 

4. Netomi 

Netomi's chatbot is like having a personal assistant handle customer queries.

5. Ultimate AI

Ultimate AI's ticketing automation works so well, it's like having a superhero on your team.

6. Elasticsearch, Solr, & Sphinx 

Elasticsearch, Solr, and Sphinx are search engines powered by AI, making it easier for your customers to find what they need.

7. Writesonic & Jasper

With Writesonic or Jasper, thanks to their generative AI writing abilities, you'll never have to worry about writer's block again.

8. Document360 

Document360 makes it simple to create, save, and publish knowledge bases.

9. Pecan

Predictive analytics software Pecan's algorithm helps you predict and understand customer behaviors for a stellar customer experience.

10. Medallia Experience Orchestration (MXO) 

MXO is the conductor of your customer journey, ensuring a seamless experience from start to finish.

Start using AI for customer service today and demolish the competition

We've covered a lot of ground, but the question remains: Are you ready to take the leap and implement these tools? Because, let's be honest, who doesn't want happier customers and higher profits? And if you're still feeling hesitant, just remember: Even the most advanced AI system can't replace good old-fashioned human empathy. 

So, give these tools a try today and see where they take you. Who knows, maybe the robots will take over the world after all. Just kidding (or am I?).

By this point, you’re all set to build a customer service mansion. You’ve got the skills to implement AI-powered customer support tools from the ground up, but if you have questions, Magical is only a phone call away. 

Get out your comically large scissors, and let’s cut the ribbon on this new build!

About the Author:
Colleen Christison is a copywriter, copyeditor, and branding strategist. She's spent her agency and freelance career writing for brands like Hootsuite, Semrush, the University of Pennsylvania, and the Mark Anthony Group.