The adoption of artificial intelligence across the business world has been swift and decisive. Over three-quarters of organizations use AI in at least one business function, up from roughly half in 2023.
One emerging and promising function is artificial intelligence analytics. From meteorology to medicine, AI transforms massive datasets into actionable insights that drive smarter decisions and better outcomes. Savvy companies are capitalizing on the power and flexibility of AI business analytics, including Shopify.
“AI analytics unlocks the ability to work with large language models (LLMs) and ask them business questions,” explains Nathan Davis, product marketing lead at Shopify.
This article explores the ins and outs of AI data analysis, the types of data analytics AI models can enhance, and tips for elevating your business with AI analytics.
What is artificial intelligence analytics?
Artificial intelligence analytics uses AI technologies to analyze data and generate insights. These technologies include machine learning (where systems learn from vast amounts of data), natural language processing (how computers interpret language and speech), and neural networks (models inspired by the human brain).
Unlike traditional data analytics, which relies heavily on manual coding and human interpretation, AI analytics tools can process large amounts of data at scale and deliver nuanced analysis with little to no human input. It also enables you to pose questions about the data and obtain insights in natural language.
“LLMs are really good at writing code and SQL—the primary language for basic analytics, dashboarding, and custom reports,” says Nathan, which is how AI analytics removes technical barriers to understanding data. “With traditional analytics, you’re constrained to building dashboards in graphical user interfaces, or you need to write SQL code or Python to get your output,” Nathan says. “With AI analytics, you can simply ask your question and get the answer.”
5 types of artificial intelligence analytics
- Descriptive analytics
- Diagnostic analytics
- Predictive analytics
- Prescriptive analytics
- Cognitive analytics
Here are the five distinct types of traditional analytics that artificial intelligence can enhance:
1. Descriptive analytics
Descriptive analytics examines historical data to summarize past performance and identify trends, producing outputs such as dashboards, reports, and data visualizations. In a business context, it answers, “What happened?” by providing clear snapshots of past business outcomes, from website traffic patterns to customer purchase histories.
For example, beauty brand L’Oréal partnered with a data firm to review its ecommerce market share and brand insights using descriptive analytics. The analysis enabled the company’s consumer products division to better understand the competitive landscape and interpret trends in multiple channels (like social media, customer feedback, and purchasing behavior). This helped it refine its customer segmentation strategies faster and more efficiently than would have been possible with a traditional, human-centered approach.
2. Diagnostic analytics
While descriptive analytics show you what happened, diagnostic analytics look at why. It reveals the root causes of business outcomes by examining relationships among data points and the conditions that shaped them.
For example, the global marketplace Etsy reported a gross merchandise sales decline of 3.7% year over year in Q1 of 2024 (this “what” was identified by descriptive analytics). Etsy’s diagnostic teams used AI diagnostic analytics to examine market data and discovered decreased consumer discretionary product spending due to a challenging macroeconomic environment—the “why.” Company executives leveraged these insights to focus resources on core customer segments. In the process, they successfully increased active buyers by 1.9% that quarter.
3. Predictive analytics
Predictive analytics uses machine learning models and statistical algorithms to analyze both historical data and current trends to forecast outcomes.
For example, Stitch Fix, an online personal styling and subscription service, relies on predictive analytics to match inventory with customer preferences. Its machine learning models use a range of customer data points, including product return rates, fit preferences, and style trends, to forecast the items a customer is most likely to keep from their next subscription box. This also helps predict the inventory items likely to become popular in future seasons. Predictive analytics has reportedly enabled Stitch Fix to use AI personalization in 75% of its subscription boxes, leading to a reported 30% reduction in returns.
4. Prescriptive analytics
One of the most advanced types of data analytics is prescriptive analytics, which goes beyond predictions to recommend specific actions. It deploys optimization algorithms and statistical modeling to evaluate thousands of potential scenarios (drawn from historical performance data, real-time operational data, external market conditions, and predictive forecasts) and identify the optimal path forward. AI enhances prescriptive analytics by continuously learning from outcomes and automatically refining recommendations. This allows it to adapt to shifting conditions faster than rule-based systems.
Airlines use prescriptive analytics to adjust ticket pricing in real time. These systems recommend exact prices for specific routes at specific times, considering competitor rates, weather forecasts, booking patterns, and historical profitability to maximize revenue.
5. Cognitive analytics
Cognitive analytics combines natural language processing and deep learning to interpret unstructured data. This is information that doesn’t fit into traditional rows and columns, like images, audio, and video. You might get this type of data from sources like customer reviews, social media posts, emails, and voice recordings (from customer service calls, for example). While other analytics types primarily rely on structured data like sales numbers and timestamps, cognitive analytics specializes in extracting meaning from messier sources. It can deliver comprehensive sentiment analysis and a deeper understanding of your business’s qualitative information.
For example, British Airways partnered with a contact center technology firm to use cognitive analytics across its customer service channels. The system uses AI to analyze massive volumes of unstructured communication data, like customer reviews, emails, and more than 35,000 call transcripts annually. Based on the insights, British Airways implemented optimized routing strategies like directing customers to self-service options for routine queries, allowing contact center agents to focus on more complex issues. This helped reduce its contact center workload by 22% and deliver resolutions 60% faster.
How to use artificial intelligence analytics effectively
- Build integrated systems
- Ask the right questions
- Put AI analytics tools in the hands of non-technical teams
- Keep learning
- Choose the right AI tools for your business
Successfully implementing AI-based analytics solutions requires more than just the latest tools. Shopify’s Nathan Davis shares creative strategies for maximizing value from your analytics investments:
Build integrated systems
The power of AI analytics multiplies when it can access and synthesize information across your entire business. Siloed systems limit the visibility and breadth of data that AI and humans can collect for analysis. Integrated platforms, on the other hand, enable comprehensive analysis that reveals cross-functional insights. This helps you connect customer behavior to operational efficiency, eliminate redundant workflows, and uncover hidden patterns across departments and ecommerce trends.
“We’re fortunate at Shopify to have access to systems that are completely interconnected,” says Nathan. He notes that Shopify’s success in using AI analytics comes from enabling non-technical staff to harness the more technical tools and platforms often reserved for engineers at many other companies.
In a typical workflow, Nathan often starts with Cursor, an AI coding platform that integrates multiple data sources via Model Context Protocol (MCP) servers. This architecture gives him access to a single interface, allowing him to query customer feedback systems, the data warehouse, and operational tools.
“The fact that Shopify defaults to open systems allows us to build potent tools and do really powerful analytics—whether you’re a data scientist or a marketer, like I am,” Nathan says.
Ask the right questions
The most sophisticated AI analytics tools are only as valuable as the questions you ask them. Before getting lost in technical capabilities, start with clear business goals and specific information needs. For Nathan, this often means asking the right questions to identify the ideal audience for a particular product launch. He might start by prompting Cursor to “summarize customer conversations and feedback related to Sidekick,” Shopify’s AI-powered ecommerce assistant. He can then combine that qualitative data with the more quantitative data from other systems.
For a recent product launch, Nathan needed to identify larger businesses with specific technical requirements and that worked with customer success managers. “Historically, I would have been very reliant on the data science team, but I was able to bring data from different tables to find merchants with a certain amount of apps installed, a certain number of employees, and so on,” he says.
Knowing exactly which SQL query to create made the task simple. This self-service approach frees data science teams to focus on complex analytical challenges while empowering marketing, product, and operations teams to move faster on routine queries.
Put AI analytics tools in the hands of non-technical teams
AI analytics works when it spreads beyond data science teams to reach marketers, product managers, operations staff, and other business functions. This democratization can change how decisions get made, helping teams make better choices, faster.
“With AI analytics, I’ve become much more independent and am making better decisions,” says Nathan. “Of course, everything we do is still vetted by our data science team, but it’s such an unlock to be able to validate hypotheses on my own.”
The effects are contagious, says Nathan, an early AI adopter at Shopify. He has helped the product marketing team to build its infrastructure in a way that makes it simple for them to work with data and other AI tools. Now, every marketer is incorporating AI analytics into their day-to-day workflow, unlocking previously unexplored opportunities and market segments.
Keep learning
Technical barriers to working with data have traditionally limited the non-technical professional’s ability to independently extract insights from large, complex data sets. Writing SQL queries, debugging code, and navigating complex data structures required specialized skills that can keep non-technical team members dependent on data scientists and engineers, who are often busy with their own projects.
AI tools are changing this dynamic by teaching non-technical professionals to accomplish coding tasks without mastery of coding language while they assist, ultimately democratizing the ability to code. Beyond participating, some non-technical team members, like marketers, have learned to build new tools.
“At Shopify, AI has allowed non-technical team members to build really cool tools around analytics,” Nathan says. When he was trying to write SQL, he’d encounter highly technical minor errors, which were difficult to understand. He says, “ChatGPT unlocked the ability to put in my SQL query with the error and get the right code back—more importantly, it would always explain what it was doing and why.” He describes it as “an ever-patient mentor that teaches you the skill of doing more technical tasks.”
Choose the right AI tools for your business
“The best advice I can give to businesses getting into AI analytics is to work with Sidekick,” says Nathan.
While many AI analytics tools exist, Shopify store owners have access to Shopify Magic and Sidekick, bringing many of the AI capabilities Nathan and his team use directly into your admin interface, where you already manage your business. Rather than requiring separate analytics platforms or coding skills, Sidekick lets you ask questions in natural language and receive insights through custom reports and visualizations.
Beyond iterating on the queries, customer reports, and customer segmentation it provides, Nathan advises taking the time to understand how it works. “Essentially, you’re having a conversation with something that can teach you how to do these more technical tasks yourself in the future,” he says. “While Sidekick excels at completing analysis tasks, you can also be more independent and go a step further and do more complex tasks yourself.”
Artificial intelligence analytics FAQ
What is artificial intelligence analytics?
Artificial intelligence analytics is the application of AI technologies—including machine learning, natural language processing, and deep learning—to automatically analyze data, identify patterns, and generate insights from structured and unstructured data at scale with minimal human intervention.
What are the four types of analytics?
The four main types of data analysis are descriptive analytics (what happened), diagnostic analytics (why it happened), predictive analytics (what will likely happen), and prescriptive analytics (what actions to take).
What is the best AI tool for analytics?
The best AI analytics tool depends on your specific needs. Ecommerce store owners benefit from Shopify Magic and Sidekick for retail-focused insights, while broader business intelligence needs might require platforms like Microsoft Power BI, Tableau, or Google Cloud’s AI analytics services.
