Data analytics turns data into evidence for decisions; AI adds systems that can generate predictions, classifications, recommendations, or content. In organizations, the work is not just choosing a model: it runs from obtaining and preparing data through analysis and evaluation to deployment, governance, and ongoing monitoring. Adoption figures offer useful snapshots, but they vary by geography, population, and definition.
What data analytics and AI cover
Analytics and AI are connected parts of a broader operating process, not a single product category. A typical learning path or project may include the stages below, although organizations do not all use the same architecture or follow an identical sequence.
Obtain and prepare data
First, an organization identifies and obtains data relevant to a task. ETL—extract, transform, load—describes one way to extract data from sources, transform it into a usable form, and load it into a destination. Data quality, access, and permissions shape what can responsibly be analyzed.
Explore and analyze
Exploratory data analysis examines data to understand its contents, patterns, and limitations before relying on it for decisions. This stage can help clarify whether the available information is fit for the question being asked.
#1 Best Overall
Develop and evaluate models
Machine-learning models can be developed for a defined task, then evaluated to understand how well they perform. Evaluation should relate to the intended use: a model that performs acceptably for one task or context is not thereby validated for every other one.
Deploy and observe
Deployment makes a model or AI-enabled process available for use. The work continues afterward: telemetry and monitoring can help teams observe real-world behavior, investigate changes, and respond to problems. O’Reilly’s catalog for Maxine Attobrah’s Essential Data Analytics, Data Science, and AI: A Practical Guide for a Data-Driven World (Apress, December 2024) lists coverage spanning these topics, including adversaries and abuse.
What adoption figures say—and what they do not
There is no single universal rate for business AI use. The figures below measure different populations and questions; they should not be treated as directly comparable estimates of the same thing.
| Measure | Reported result | How to read it |
|---|---|---|
| U.S. firms using AI in a business function | 18% during the November 2025–January 2026 reference period | The U.S. Census Bureau’s 2026 AI supplement to its Business Trends and Outlook Survey reports a firm-level measure. Source. |
| U.S. employment-weighted AI use | 32% during the November 2025–January 2026 reference period | This is weighted by employment, not a share of firms, in the same Census Bureau paper. Source. |
| Expected U.S. firm use | 22% expected within six months | The Census Bureau paper reports this expectation alongside its November 2025–January 2026 measures; it is not a later observed adoption rate. Source. |
| Reported UK AI uses | Researching information: 28%; summarizing or collecting in-house information, or drafting reports or correspondence: 21% | The UK Business Data Survey 2026 identifies these among the most common reported uses. Its population and wording differ from the Census Bureau’s firm-use measure. Source. |
| AI guidance on access to business data and files | 62% | This share is among UK businesses reporting an AI policy or guidelines, not all UK businesses. Source. |
| AI tools integrated into existing business systems | 21% | This share is among UK businesses using AI, according to the UK survey. It is not directly comparable with the Census Bureau’s U.S. firm-use measure. Source. |
The UK Department for Science, Innovation and Technology cautions that differences in how AI use is defined, and variation across tasks and roles, make overall use difficult to measure consistently. A reported use such as researching information also does not, by itself, establish how deeply AI is embedded in an organization’s systems or decisions.
Recommended Free Tools
How to think about responsible AI governance
The National Institute of Standards and Technology (NIST) describes its AI Risk Management Framework as voluntary guidance for incorporating trustworthiness considerations into the design, development, use, and evaluation of AI systems. Its four functions organize risk work across those stages:
| Function | Role in the framework |
|---|---|
| Govern | Establishes the organizational structures and practices for AI risk management. |
| Map | Builds context about the system, its intended use, and the risks that may arise. |
| Measure | Assesses and analyzes identified risks. |
| Manage | Prioritizes and responds to risks over the system’s life cycle. |
NIST says AI RMF 1.0 is being revised. Its page also identifies the Generative AI Profile, NIST-AI-600-1, released July 26, 2024. The framework is voluntary guidance, not a universal certification or a substitute for applicable legal or organizational requirements. See the NIST AI Risk Management Framework for its current status and materials.
Why governance must continue after deployment
Pre-release evaluation cannot guarantee that an AI system will behave identically across changing real-world conditions. In its March 9, 2026 announcement about a report on monitoring deployed AI systems, NIST points to variability and unpredictability in deployed systems as reasons post-deployment monitoring matters. The report focuses on monitoring categories and challenges; the announcement does not establish one required tool or a universal monitoring frequency. Read NIST’s announcement.
For an organization, follow-up evaluation should be connected to the intended use and the risks identified before launch. Teams need a way to notice when behavior changes or a problem emerges, assess what it means in context, and decide whether to adjust, restrict, or stop use. The precise monitoring approach depends on the system and its use; the cited NIST material does not prescribe one schedule for every deployment.
Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallA practical way to assess an analytics or AI proposal
The following questions are a decision aid, not a vendor ranking or a claim that one platform suits every organization. They are especially useful when the data may include financial or other sensitive information.
- What decision or task is it meant to support? Define the task and intended users before comparing technology. A tool for summarizing documents and a system that informs consequential decisions call for different evaluations.
- What data will it use? Identify the source, sensitivity, access permissions, and whether the data is appropriate for the purpose. Ask who can access business files and how that access is governed.
- How will it fit existing work? Consider whether the system must integrate with current processes or systems, and how people will verify or act on its output.
- How will performance and risk be evaluated? Decide what evidence would count as acceptable performance for the actual task, what risks need attention, and who is accountable for reviewing the results.
- What happens after launch? Establish how real-world behavior will be observed, how issues will be escalated, and who can change or suspend the system if needed.
- What are the operating constraints? Account for deployment environment and the practical costs and limits of running the system, rather than judging only a model demonstration.
These questions help expose what a proposal leaves unspecified. The available evidence does not establish a best analytics platform, AI model, universal return on investment, or definitive map of software categories; those judgments depend on the task, organization, data, and deployment conditions.
Where to learn more
For an introductory resource that spans data analytics, data science, and AI, O’Reilly’s catalog lists Maxine Attobrah’s Essential Data Analytics, Data Science, and AI: A Practical Guide for a Data-Driven World, published by Apress in December 2024. Its catalog description includes obtaining data, ETL, exploratory analysis, model development and evaluation, deployment, telemetry, and adversaries and abuse. It is a learning resource, not a current comparison of platforms. See the publisher catalog listing.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
