Four Opportunities for AI to Innovate Across the Customer Journey

CX faces two powerful but often contradictory forces: Implement AI To Cut Costs and Reduce Customer Rage and Frustration. Difficulty in escalating to a human is the top cause of customer rage, and it has bubbled over into violence in several venues. On the other hand, if AI is applied intelligently in some new areas, it can enhance CX, loyalty, and positive word of mouth. CCMC’s recent CPG study shows that well-crafted digital responses can produce up to 42% delighted consumers.

Four trends are roiling the marketplace

• Inflation and covid have led to consumer perception that value has declined
• Cost cutting, touchless interaction, and AI have all led to more frustration and rage – reaching a human is the top cause of customer Rage
• Pressure to implement AI has become overwhelming due to both cost savings and “everyone is doing it – keeping up with the Jones.”
• Digital channels have surpassed telephone and in-person interactions – chat, email and self-service now account for ¾ of all interactions

Four Opportunities for AI to Innovate Beyond the Basics

Four areas have been ignored where AI can simultaneously enhance efficiency and satisfaction/loyalty. These areas are problem prevention, enhanced response, intentional delight, and escalation.

1. Prevent Problems

Help the customer get better value from the product.

1. Sales Closing – After the customer has purchased, during the closing process, AI stresses the product limitations – as in “get the most from your product by knowing the details”, e.g., warranty, in a low-key manner while showing how you help the customer achieve their goals. The message also stresses the value of downloading the app and onboarding.
2. Customer Onboarding – Onboarding is most successful when tailored to the customer based on customer competence and expectations. Successful onboarding can reduce problems by 40% and customer top-box ratings by 30%. AI can help
   • Assess customer channel preference, competence, and expectations. For example: “Have you used this type of product in the past?”
   • Incentivize partaking of available education – If you watch this three-minute video, we’ll give you a discount
   • Education via multiple preferred via highly visible channels – e.g., email welcome letter, video on executing top three transactions
   • Stress accessibility: “We can only solve problems we know about!”
3. Product Startup – Tour the website, load the app, execute first three transactions – AI tailors the startup process to the level of competence, preferred channels, and customer objectives
4. Product Use – For most products, especially software and financial service products, AI is perfect for monitoring startup and any glitches as well as providing Just In Time (JIT) suggestions.
5. Encounter Issue – Encourage customer to ask for help – AI looks for signs of trouble or opportunities for delight and reminds customers, “We can only solve problems and questions we know about.”
   • AI looks for failed transaction
   • If AI determines the customer is happy, it can suggest the next step or a value add. Also, AI can provide interactions causing delight with education and cross-selling. Things going smoothly is not a time to relax – go further and create delight with extra information!

2. Enhance Response

Go beyond the basics to include rationales and provide value adds/delighters such as intelligent cross-sell, education, enthusiasm, and humor.

6. Intake – The intake process facilitates preliminary analysis via AI Type-ahead (see example graphic of AARP AI-driven Help Page below) and asking questions/analyzing when the customer requests assistance via:
   • Visit website, search FAQs
   • Chat/email
   • Telephone using speech recognition and text analysis
   • Google, YouTube

   

7. Assign to Three Levels of Response Processes, Simple, Grey, or Complex – AI can interpret and classify the communication based on content and customer data.
8. Response production – Key components of any response include the rationale for the answer and anticipation of the customer’s next question. Also, AI can determine whether it is appropriate to inject a delight action like additional education, enthusiasm, or empathy into the response.
   • Simple – AI can provide basic answers with rationale and/or add delight components humor and enthusiasm
   • Grey – AI can ask more questions to clarify the situation and provide the best answer based on flexible solution spaces with the caveat of escalation, or offer escalation if probability of success is judged low
   • Complex – AI can analyze the question and escalate to the proper expert group with all available information to avoid customers having to recount history.
9. AI can analyze the result of the customer Receipt of Answer
   • Happy
   • Less than happy – offer escalation

3. Create Intentional Delight

Delight can be delivered to consumers and business customers. The “extras” can be empathy, enthusiasm, education on a golden nugget on how to get more from a product, or even intelligent cross-selling. Customers who are intelligently cross-sold are willing to pay more and spread powerful word of mouth. AI can select or suggest the appropriate delight action.

10. Customers Receive answer – AI identifies unhappiness or “just” satisfied – consider how to move up to delight
    • Happy – possibility for delight? Do we know or do they just leave? Apply one or more of 10 possible delight strategies based on relevance to customer (see my CustomerThink article on creating delight).
    • “Tell us your outcome so we don’t have to bother you with a survey.” They’ll rather respond now rather than get another survey.
    • If not happy, abandon, keep searching, or attempt to escalate
11. Customer abandon – Get feedback from customers before they leave: “Please tell us if you found what you needed and then we won’t have to bother you with a survey.”
    • AI keeps score of abandons – An abandon causes 10-20% dissatisfaction even if the customer moves on to another channel.
    • Motivate to escalate, For example, both TELUS and Xfinity stress they can only solve problems if reported and they will connect you with a human. At TELUS, even with the executive team!
    • Use AI to identify probable dissatisfaction in speech or text and proactively offer escalation channel
12. AI can identify when customers attempt to escalate by asking questions such as “Executive!” or” Supervisor!” or more than three attempts to find an answer. Other actions that facilitate escalation are:
    • Customer looks for alternative channels – which should be very visible
    • Proactively acknowledge “you don’t seem to be finding what you need.”

At this point, the customer escalation process begins, resulting in either delight or Rage.

4. Facilitate Escalation

AI can contribute in six ways to the escalation process:

• Identify unhappiness via text, voice, transaction failure, or search behaviors.
• Acknowledge unhappiness
• Offer escalation, possibly with one more AI try before human
• Assure expedited, accurate escalation – 90 seconds or less – Geico has short queues for escalated calls, knowing the customer already has a short fuse
• Assure all data is sent to the human handling the escalation.
• Identify reason for escalation, which is usually different than a basic issue, for future analysis. For example, a lack of empowerment or not able to access information or a supervisor on a particular issue could cause escalation.

Recommendations

Nothing is being recommended below that you should not already be providing to your front line and escalated CSRs. Most escalations are due to flaws and weaknesses in basic customer service and customer expectation setting. If you perfect the basic customer onboarding and service systems, you’d have many fewer legal, regulatory, mal-practice, insurance claims and buybacks

1. Audit your processes and make sure the KMS includes rationales and guidance for delighters. Build flexible solution spaces into your KMS for the grey issues – provide 3-6 approaches to a difficult issue depending upon circumstances. Your front-line employees are your best consultants in this area.
2. Assure that your intake process evaluates all questions and assigns them as simple, grey, and complex – to drive the support/response process. Help the customer via type ahead search rather than FAQs or keyword search.
3. Empower your front line and where possible, AI, to have the flexibility and the ability to delight. Digital text can delight using simple words of empathy and enthusiasm. Even experiment with humor — 90% works and you learn from the flops.
4. Measure the outcome of every grey and complex issue as well as abandons. Few companies record and report failed searches – every failed search is a phone call waiting to happen. This outcome data should be fed to the service and marketing continuous improvement activities. Marketing should support because delighted customers are your cheapest way to acquire new customers and raging customers can destroy a brand.