Artificial intelligence and chatbot tools generate data. A lot of data. Coping with the massive flood of data from these tools is challenging, but offers insights that are otherwise hard to obtain.
The digital economy of 2018 is fueled by data, and the challenge of turning raw data into facts, trends, and projections. This shift is a boon to external vendors who offer dashboards with multi-channel capabilities that slice and dice all sorts of data so marketing, sales and social media teams can make use of them. However, it’s a bit more challenging for those teams.
These teams, whether they work at businesses, government agencies, or non-profits, are all building their own analytic arsenals from in-house data they accumulate through chatbots, social media, website interactions, and other sources. These huge collections of data can be used for acquiring new customers, planning organization strategy, and much more.
Erick Harlow of Forensic IT, a data analytics firm, recommends a three-part plan to making sense of user data from disparate sources.
“No matter how many devices or marketing vehicles are involved, we always ask three questions,” Harlow says. “1. What do we know? Such as, sales are down. 2. What do we need to know? Have people stopped buying this product all together, or just our product? 3. How do we analyze and fix the problem? It is important to know which data will help in answering the question and which can be discarded.”
Data Warehouses, Data Lakes And Data Reservoirs
This is where terminology kicks in. Depending on what kind of data an organization is saving, they are building up either data warehouses, data lakes or data reservoirs. Knowing how your organization is accumulating data is a crucial part of building in-house strategy.
Data warehouses are huge, organized stores of information from multiple sources which are used for long-term analytics using historic data.
Data lakes are a bit different. While data warehouses are largely organized and collated into a common format, data lakes consist of huge stores of unorganized data accumulated by an organization. They’re a bit harder to work with for business intelligence and analytics purposes.
Then there are data reservoirs, which are the backbone of large-scale machine learning and big data projects. Popularized by research firm Gartner, data reservoirs are architectures to make sense of large stores of structured and unstructured data.
Customer visits to websites, Alexa skills, mobile apps, chatbots, and downloadable software all generate a ton of data—but this data is often formatted in very different, hard-to-combine ways. It’s crucial to have an organizational strategy for making sense of those different data formats.
AI, Bots & The Data Crush
In the late 2010s, brands are as likely to interact with customers through Alexa skills, chatbots, and mobile apps as they are on the phone or in person. Chatbots and apps generate lots of data, but making sense of that data can sometimes be difficult.
Audrey Wu of Convrg, a Los Angeles-based chatbot design firm, says that it’s crucial for companies to understand how chatbots work and what kind of data they generate. For instance, chatbots generate different data for organizations depending on whether they operate via SMS text message based, Facebook Messenger, or Kik Messenger.
“Think about how to make a flow to keep users engaged,” Wu suggests. Interactions users have with chatbots generate logs which can also be mixed with additional data. For instance, Facebook Messenger bot interactions were formerly linked to users’ Facebook IDs. Kik’s platform, meanwhile, tends to attract a young demographic which is already familiar with chatbot etiquette.
Convrg uses an analytics dashboard called Dashbot to gain insight from their chatbots; there are also other dashboards on the market such as Google’s Chatbase.
Alexa skills also generate huge amounts of data for brands to utilize. Although Alexa analytics are still in their infancy, Amazon has been rolling out new features for skill/app developers to better understand how users interact with their Echo or Tap speaker. Brands can gain insights through Alexa’s API into metrics—like how often their skills are used, how many people use them, and how successful users’ interactions with the skills are.
Big Data, Bigger Data, Biggest Data
There’s no shortage of analytics platforms, business intelligence solutions, and data visualization tools for every conceivable need. The digital analytics industry includes everything from mass-market players like Google Analytics and Tableau to niche tools, and of course, there’s no one-size-fits-all solution.
Figuring out the best pathway for making sense of data from chatbots, apps, and smart home skills ultimately depends on your organization’s needs. It also means keeping careful track of platforms—in tech terms, both Facebook Messenger and Amazon Alexa are in their infancy as customer interaction platforms.
A successful data strategy depends on understanding how your customers are interacting with you, your organization’s goals, and what your budget and timeframe is for finding insights. But the good news is that chatbots and AIs are more than just an easy way to reach out for customers—they’re a valuable analytics tool as well.