It’s that time of year again – when management teams return from the holidays, rejuvenated and re-energized for what 2019 will bring.
And while most teams are developing and executing plans to drive growth for their businesses, top-performing companies are also focused on mitigating risk: the possibility that the company will deliver lower-than-expected revenues or profits, or even losses. In fact, in our recently-concluded survey of customer experience industry leaders, nearly 50% cited predicting and managing risk (including customer churn and reputational risk) as one of their major day-to-day challenges.
Business risk arises from a variety of sources. Well-known examples in the business world come mostly from industries dealing with volatile commodities. Financial institutions with risky investment or trading positions, for example, often have entire organizations dedicated to algorithmic risk assessment and management. Airlines often pursue sophisticated hedging strategies to mitigate changes in oil prices.
But hedge funds and airlines aren’t the only ones that need a strategy in place for managing risk. Businesses across the B2B and B2C spectrum face a variety of daunting challenges – from legal and fraud risk to reputational risk and new competition.
Consider some of the statistics:
- According to research from Experian, online shopping fraud rose 30% from 2016 to 2017 – more than twice the rate of overall ecommerce growth.
- In the United States, 2.2% of GDP is spent on tort litigation, with plaintiffs winning 55% of cases on average.
- Customer churn costs US companies an average of $41 billion a year
So how can businesses without dedicated risk management departments better protect themselves against risk?
1 – Understand your business’s risk landscape
Before rolling up your sleeves and getting into the data, begin by mapping out – and prioritizing – the types of risks that are relevant for your business. Using the following framework can help guide the team through this task:
- Type of risk: First, does the risk represent downside primarily on the demand side (i.e., it leads to a decrease in customer demand for our products or services) or on the supply side (i.e., we incur additional costs or obstacles in bringing our products and services to market)?
- Source of risk: Second, does the risk arise primarily via vulnerabilities in our own operations and systems (e.g., IT), or because of changes in external factors (e.g., regulatory requirements or competitive landscape)?
Here’s an example of what a completed risk matrix might look like for a SaaS marketplace platform:
Mapping out the risk landscape can help the team focus on the highest-priority issues for your business and begin to set the stage for remediation tactics.
2 – Implement an AI-based solution to mine customer interactions
Some management teams contend that business risk is inherently impossible to predict – or that no company could possibly amass enough data on the minutiae of customer interactions and sentiments to do so accurately.
Neither could be further from the truth.
In reality, most companies already have an abundance of data on customer interactions within their most commonly-used interaction channels. The call center, for example, is a treasure trove of exchanges with customers on everything from troubleshooting to fulfillment issues to service delays. Similarly, for many companies, email and chat are major repositories of information on customer issues, satisfaction, and more.
The challenge with all of this data is that it is unstructured – meaning that it consists of human language rather than quantitative data easily processed by a computer. For this reason, the most effective solutions for mining this data rely on artificial intelligence (AI), a wide range of machine capabilities that mimic human capabilities, like perception and cognition. Common applications of AI include the ability to understand human speech and draw complex inferences from a long chain of events.
AI is the ideal tool for parsing these vast quantities of data to draw actionable conclusions and inferences. For example, an AI platform might learn from studying customer interactions that the following phrases – embedded in a negative service context – are predictive of high intent to take legal action: “hoax,” “scam,” “outrageous,” “offensive,” “violate,” “lawyer,” “attorney.” In fact, AI algorithms can go much further, leveraging deep learning to mine contextual signals and infer intent even from far more subtle ultimatums of legal action – like, “I hope we don’t have to get anyone else involved.”
Using an AI platform, begin to tag, classify, and measure the incidence of business risk using the matrix that you previously mapped out for your business.
3 – Mitigate, monitor, and measure
Now it’s time to take action. A robust AI platform should offer the ability to flag individual interactions and create alerts – for example, to notify members of management when new legal risks or competitive mentions arise.
Before initiating outreach, make sure that you have laid out a well-validated mitigation strategy for each quadrant of business risk. For example, a response plan for off-platform fraud might involve issuing the user a preliminary warning, and then suspending the user from the platform if the behavior persists. A response plan for customer dissatisfaction might involve escalating the issue to an experienced customer care or customer success consultant.
As you initiate your outreach plan, measure changes in the incidence and resolution rates of different business risks over time. This will help the team validate the success of the overall strategy – and also test and optimize between different potential responses. (For example, should all legal threats be routed immediately to legal, or should they first be addressed by a senior manager within the customer care organization?)
At the end of the day, any business that is committed to delivering innovative products and services will confront an array of operating risks. And while these risks can’t be altogether eliminated, AI can identify and help mitigate them by intelligently mining the customer data that businesses already have at their fingertips – enabling management teams to focus on delighting their customers.