Artificial intelligence technologies a boon for customer insights
AI can help companies shrink the gap between customer data and actionable insight by feeding intelligence into CRM, marketing automation and other operational tools. Here's how.
You've probably interacted with a number of artificial intelligence technologies today, whether you've asked your Amazon Echo for the weather report or used a virtual agent chatbot on a retailer's website. With every interaction, you leave breadcrumbs of data, which create trails that companies can use to gain insights into your preferences, buying habits and much more.
This new reality means companies have access to a gold mine of customer data, but they are only using about 12% of it for insights, explained Forrester Research vice president Srividya Sridharan, during a recent AI Trends webinar called "Drive Innovation in Customer Insights with 12 Artificial Intelligence Technologies."
"Time and time again, we hear from our clients, 'we are drowning in data and starving for insight,'" said Sridharan, who specializes in customer data research.
She added that companies have made tremendous strides in using data in daily decision-making, but there are only pockets of excellence. Customer insight teams may be creating effective personalized marketing, but the use of data and insight has to happen at the enterprise level to be scalable.
"There's still a long way to go to truly using insights at scale across the enterprise," Sridharan said. "We still see gaps between insights and actions."
How AI provides customer insights
Artificial intelligence technologies can help companies shrink the gap between data and insights by automatically feeding intelligence into CRM, marketing automation and other operational tools. They can also embed tested insights to improve customer engagement, empower employees so they can make decisions and create closed loop systems, according to Sridharan.
Time and time again, we hear from our clients, 'we are drowning in data and starving for insight.'
Srividya SridharanVP, research director, Forrester Research
Customers may have taken complicated paths to your product; artificial intelligence technologies can process the massive amounts of data that informed their actions, in real time. Tools such as facial scanning, text analytics, machine learning and natural language generation are used to let companies see what their customers see, identify the right messaging and convey it in real time.
AI can bring insights to the forefront automatically. Data you weren't looking for can be used to pre-empt customer behavior rather than taking a reactive approach.
AI tools like machine learning can help you use unstructured data from text or call center conversations, for example. "There is a lot of potential in terms of customer insight that you are leaving at the table [that can help you] understand context and meaning behind the behaviors your customers exhibit on your various channels," Sridharan said.
AI to help you sense, think, act and learn
Forrester Research identified artificial intelligence technologies that help companies sense, think and act. These commercially available and open source tools pull customer data, and they will be an important part of a marketer's stack -- if they aren't already.
Sense tools offer image and video analysis using facial recognition technology, speech analytics and text analytics. The AI tools in this category use unstructured data to uncover customer sentiment, emotion analysis, tone and context behind customer behavior to provide a full picture of what a customer is doing and to predict what they'll do next.
A number of large companies in retail and e-commerce, including Michael Kors, Walmart and Amazon, already use artificial intelligence technologies to create better product recommendations, help consumers find what they're looking for and improve efficiency internally, said Heath Terry, managing director at Goldman Sachs, during his presentation at AI World in Boston last month.
The second area in which AI can be used is to augment thinking. Machine learning and deep learning platforms are really the core of AI, Sridharan said in the webinar. Companies use deep learning for image recognition to identify people or places, for instance. The technology that enables machines to interact with humans uses natural language generation, which serves to help companies act by bringing insights to the surface so users can act quickly.
The final category of technologies bucketed by Forrester is software bundles that help companies take action. They include AI-enhanced analytics tools, intelligent research software, recommendation platforms, pre-trained tools designed for specific vertical markets and conversational service platforms.
Many companies already employ conversational agents, or chatbots, as these are relatively simple to integrate using APIs available from cloud providers such as Conversica, IBM, Microsoft and Oracle.
In general, chatbots can take an utterance, classify its intent and map it to a task or inquiry -- but conversational agents should be able to do more than that, said Rob High, IBM Fellow and CTO of IBM Watson, during a presentation at AI World.
"[They] should provide perspective that you didn't intend," High said. "We are taught, 'don't expect too much.' But we need to understand a problem well enough to really help. A conversational agent should be able to immerse itself in a conversation and be able to adjust and provide information autonomously."
How to get started with AI
You'll need access to large amounts of well-curated, non-siloed data for these AI tools to work, so it's important to address any data collection challenges first. If you have data issues, AI will magnify those problems, Forrester's Sridharan said in the webinar.
She suggested researching the artificial intelligence technologies listed above and prioritizing which ones can deliver the most value to your organization based on your customer challenges. If your customer service or marketing efforts aren't reaching the right audience at the right time, you'll want to choose technologies that address those areas.
To further the point, companies can automate, optimize and predict anywhere that there is spoken content, but it's best applicable where there is a clear ROI. In customer service call centers, you need to know how likely a customer is to churn or to buy a product, and what percent of customers never buy, so you don't waste time following up on those leads, said Walter Bachtiger, CEO of VoiceBase, a provider of APIs for speech recognition technology, during a panel discussion at AI World.
"We need to look at calls as information," Bachtiger said. "Voice is a rich source of information, and companies that use it tend to be the winners. When you spend the time [and] do the job right, you create something that really adds value."
He also explained that it's important to understand how that voice data can be used.
"Many customers come to us because they want data on customer sentiment," Bachtiger said. "They want it to predict churn and provide better customer service. What they really want is not sentiment, because there isn't actually a correlation between sentiment and customer churn. If a customer is from New York, and you are from the South, you may think they sound cranky, and your perception of sentiment may be incorrect."
Companies can start with AI by infusing one type of the technology into an existing application. For example, speech analytics APIs can be integrated with call center operations to learn more about customers, as Bachtiger explained.
Start thinking about what your first closed-loop insight system will target, be it HR analytics, marketing or predictive pricing; it should be embedded in software in order to scale, according to Sridharan.
