3 Ways to Leverage AI for Qualitative Customer Data

AI-powered text and sentiment analysis tools allow companies to leverage qualitative customer data and feedback for key insights.

qualitative customer data

Illustration by Ana Duje.

Whether you splurge on a new phone or settle into a new job, you’ll probably feel compelled to tell someone all about it. In conversation or texting with friends, you’ll use casual language driven by emotion and riddled with colloquialisms—that language won’t change much on social media or when crafting a devastating online review. In the realms of Customer Relationship Management (CRM), Customer Experience Management (CEM) and even workplace management, the solution for harnessing such unruly feedback is to transform it into unstructured data.

Directly and informally derived from customers and employees, this qualitative data is invaluable, even if won’t fit in a predefined model of organization or marketing and doesn’t immediately make for a pretty pie chart. Found in emails, support tickets, call transcripts, comments, reviews, social media posts as well as open-ended survey responses, unstructured data appears chaotic at first glance, but with new forms of AI data analysis it can be tamed to help solve business problems, make organizational changes or tweak a service or product.

1. Use text analytics to make sense of feedback

Historically, unstructured data has been difficult to analyze and relegated to second-tier status as supporting quotes for more easily analyzed quantitative data. We’ve been missing out on a goldmine of direct, authentic feedback. What customers think about products and services has been online for years on review sites and other channels, and customers have been calling complaint lines for far longer than that—the difference now is that we have much better ways to analyze what people are saying, identify what matters most to them and respond quickly.

AI text analytics alongside other CX tools can provide the broad brushstrokes of customer and employee satisfaction, which can be used to inform further quantitative research. Yet as AI tools have become more powerful, the gritty details of feedback have also emerged. That level of analysis groups similar reviews by the thousands and links feedback to date, location and other customer demographics, massively boosting the value of unstructured data for marketing, customer research and other business purposes.

First, data is gathered or “mined”—AI software analyzes natural language text (such as the kind found in unsolicited written feedback online) to identify key concepts, themes and patterns. An AI text analytics tool like Keatext then makes sense of those concepts, themes and patterns by rapidly sifting through the text, cross-mapping information and adding valuable, unbiased context to the commentary. The more data that AI tools analyze over time, the more accurate the tools become. This results in new levels of insight that decision-makers can use to solve business problems.

2. Go online to find the most useful unstructured data

While solicited feedback, such as customer commentary gathered from surveys, purchase feedback and after-service feedback, has its value and typically provides you with what you’re looking for, unsolicited feedback offers deeper context and more surprises. Often, this kind of unstructured data provides you with exactly what you didn’t know you were looking for.

The best place to find all kinds of unstructured data is on the web. Companies like Keatext offer a service that uses software scripts that “scrape” this data from sites such as Amazon, Yelp, and TripAdvisor for customer comments, and sites like Indeed and Glassdoor for current and former employee feedback. Identifying which sites feature the most relevant data for the clients’ needs and scraping several offers a more dynamic data set that can provide a bigger feedback picture on your company.

3. Embrace negative commentary and let complaints solve problems

While some companies conduct surveys with both customers and employees, the responses to survey questions typically aren’t as genuine as online commentary. There’s something about making a public comment, anonymous or not, that prompts people to say what they mean, grammar mistakes and all. Most often that’s because they have something negative to say about a product, a service or a company—they not only want to be heard, but want something to be done about the problem. That feedback is gold for companies big and small, helping them pinpoint previously unknown issues and work to determine their impact, whether that be on product sales or employee retention.

Organizations look for AI solutions for two major reasons: to fix issues they already have and to fix issues they don’t even know about. So if a data analysis returns with 1,000 praises but 7,000 problems, we consider that a windfall. Patterns emerge from those 7,000 complaints, letting us rank problems as well as map correlations to customer information, dates, locations and any other relevant data. This wealth of knowledge can lead the way to making informed changes to products, services or your company.

Keatext decorative accent

Analyze your qualitative customer data with Keatext

Explore our platform

Related Stories

What is Text Analysis?

Text analysis is a machine learning technique used to analyze and surface insights from unstructured text data such as customer feedback.

Keatext decorative accent

Subscribe to our blog