By adopting AI-based support technology, B2B customer support teams can take a proactive stance to head off escalations and manage them at scale.

Today’s consumers aren’t settling for mediocre support—they expect high-quality, fast customer service.

According to a PwC survey, 80% of customers say that a positive experience includes “knowledgeable help and friendly service.” And in a 2020 Zendesk survey, 72.5% of consumers said that a speedy resolution matters the most when they have an issue to resolve with a company.

Support leaders can’t ignore these standards, especially in B2B where retention is so important. While there are a number of ways to meet these expectations, minimizing ticket escalations is at the top of our list. By adopting AI-based support technology, B2B customer support teams can take a proactive stance to head off escalations and manage them at scale.

What is the B2B support ticket escalation process?

Most support organizations have a tiered system for ticket escalation, but the B2B process can vary depending on who is requesting the escalation.

Customer initiated escalation

In this type of escalation, a customer expresses their dissatisfaction directly to the business, and the case is escalated. Perhaps the customer is displeased with the support interaction or the product itself and wants a person at a higher level within the company to resolve their complaint. Essentially, this escalation is the B2B equivalent of “I want to speak to your manager.”

Here’s an example: you run support for Company A, a data analytics software company, and Company B—an international financial services group—is one of your most valuable clients. Their contract, which is worth upward of $300,000 annually, is due for renewal soon.

An IT manager at Company B has been working with your support team to resolve an issue: your software keeps crashing during a routine system update. Your team directs the IT manager to a knowledge base article, but they still aren’t able to fix the problem. The manager becomes annoyed and sends an angry email to ask for the case to be escalated to a higher-level support representative at Company A.

Employee initiated escalation

In an employee-initiated escalation, the frontline support agent recognizes they don’t have the expertise or resources needed to resolve a case on their own. Perhaps they need to pull in product experts or obtain more information from the customer success team. So, they escalate the case to a higher-level support representative or another team.

Let’s take the above scenario. After some initial troubleshooting and screen-sharing, a frontline agent at Company A realizes that they can’t resolve Company B’s software issues without more help. The agent escalates the case to a Level 2 Support Engineer to do a root cause analysis (RCA) to pinpoint the issue and the best approach for resolving it.

Executive initiated escalation

The third level of B2B escalation is the one that CEOs, CROs, VPs, and pretty much everyone dreads. In our industry, it’s sometimes referred to as the “$100,000 escalation,” “executive escalation,” or even the “nuclear escalation.”

This situation happens when the first two types of escalations either fail or don’t get responded to in an effective way. The customer resorts to having their CEO or another high-level executive personally intervene. And the results can be catastrophic.

Let’s circle back to the previous scenario. Imagine that days or even weeks have passed with Customer B’s software update issue going unresolved. Multiple emails have been exchanged between teams at Company A and B, yet the case has languished—lost in the weeds of a support team’s daily firefighting.

It’s 4:30 pm on a Friday, and the CEO of Company A gets an irate call from the CEO at Company B. The CEO at Company B—a high-value customer—is extremely displeased. That original, unresolved escalation? It’s turned into a critical software failure that has completely disrupted Company B’s business and cost them thousands of dollars.

Immediately, Company A’s CEO jumps into crisis mode, pulling in high-level executives from customer success, product, sales, etc. The beleaguered support team—already overstressed and overworked—gets blamed for the crisis. Folks work through the weekend to resolve the issue, with the CEO flying to Company B’s headquarters to do damage control.

But the damage has been done. Customer trust has eroded, and Company B ultimately churns. Company B switches to a competitor, and in the process of doing so, their CEO speaks negatively about Company A. Word spreads. Low morale at Company A leads to agent burnout and turnover, so the organization loses revenue. It’s called the “nuclear escalation” for good reason.

How to stop support ticket escalations with AI signal-based extraction

When businesses integrate an AI-powered platform like BoostTalent SX™ with their ticketing system, their support teams can extract valuable insights from unstructured customer data (we’ll cover what that means shortly). Teams can apply what they learn from this information to provide a more proactive support experience, reducing escalations and improving customer satisfaction.

Use customer sentiment analysis

At the heart of our technology is customer sentiment analysis. BoostTalent’s SX platform continually extracts unstructured customer data to create two scores: a customer sentiment score and an attention score, as well as a ‘likely to be escalated’ prediction. At a glance, a manager can look at these numbers, quickly understand how a customer is trending, and take action based on how urgent their issue is.

• Sentiment score (the higher, the better): 70 indicates a “C average,” and as you go upward toward 100, the customer is happier.
• Attention score (inverse to sentiment score, starts at 30): as you go upward toward 100, the case is more urgent and requires more immediate attention.
• Likely to be escalated prediction: predicts and recommends pre-emptive corrective actions by identifying the factors that are driving the case towards escalation.

By “unstructured data,” we mean all text-based communication associated with a case, including case notes, emails, and voice call and chat transcripts. BoostTalent SX doesn’t look at an email in isolation but rather reads that email and puts it in context with everything else that’s ever happened between a customer and a vendor.

Choose a support experience platform that fits your needs

BoostTalent SX’s domain of knowledge is B2B customer support for highly technical products. The platform’s AI and NLP technology is trained with over 60 million B2B customer support interactions.

Along with being purpose-built for technical B2B support, Support Logic SX is also fine-tuned to your company’s unique goals. And the tool gets more precise over time. Users accept or reject escalation predictions, and the platform’s Machine Learning (ML) technology uses those inputs to quickly adapt and get more precise in its case analysis and escalation predictions.

Act on predictive alerts to reduce customer escalations

By continuously ingesting and analyzing case data, BoostTalent SX is able to send predictive alerts when customer sentiment is trending toward an escalation. Support teams are alerted when there’s a problem, so they can address it before it becomes a full-on crisis. Managers and agents can identify factors that are driving the case toward escalation and take corrective actions.

How might this play out? Let’s circle back to our previous scenario of the nuclear escalation on a Friday at 4:30. Using BoostTalent SX’s predictive insights, that escalation never happens. Here’s what happens instead.

Days earlier, a support manager at Company A received an alert from BoostTalent that this case is likely to escalate. BoostTalent also flags it as a high-value customer whose contract is up for renewal next month. Everyone immediately understands the urgency of this case.

The support manager confers with the lead agent on the case, and they analyze the data in BoostTalent’s dashboard. They work with the product team in Slack to figure out the solution: the client is using version X of the product, and they need to upgrade first to version Y, in order for version Z (the upgrade giving them all the trouble) to work properly. The support team reaches back out to the customer to schedule an assisted upgrade in the near future.

Armed with this information, the support team authors a knowledge base article for other customers that may encounter a similar issue in the future. The piece directs these customers to the support portal, where they can schedule a version upgrade.

This article saves time for both customers and support agents. Company A’s support team receives positive feedback from both the customer and fellow team members. It’s a win-win all around.