ITTrainingNow

Explainable AI (XAI) Fundamentals

Learn to interpret, question, and communicate explanations of AI-supported predictions.

Course overview

This one-day instructor-led course introduces explainable AI to people who use, manage, or review AI-supported decisions. Learn the difference between understanding a model as a whole and explaining an individual prediction. Using a fictional service-request prioritization example, participants interpret instructor-provided explanations, examine their limitations, and practice communicating findings to different audiences. No coding or prior AI experience is required.

1 dayCourse length
6 modulesGuided exercises and practice
Entry levelBusiness and technical teams
Instructor-ledCase studies and workshops

Course coverage

Explore what AI explanations can show, what they cannot prove, and how to use them responsibly.

  • 01 AI predictions, explainability, and stakeholder needs
  • 02 Interpretable models and global versus local explanations
  • 03 Feature importance and individual feature attributions
  • 04 Counterfactual examples and their practical limitations
  • 05 Explanation quality, uncertainty, bias, and human review
  • 06 Plain-language communication and escalation

Participants will learn to

  • Explain why different stakeholders may need different explanations.
  • Distinguish a model-wide summary from an explanation of one prediction.
  • Interpret simple feature importance and attribution examples.
  • Describe counterfactuals and question whether proposed changes are realistic.
  • Recognize that explanations do not establish causation, accuracy, or fairness.
  • Identify uncertainty and questions for model owners or specialists.
  • Communicate an explanation and its limits in plain language.
  • Plan human review and a route for challenge or escalation.
Practical outcome: Prepare a plain-language explanation brief and review checklist for a fictional prediction, including evidence, limitations, and questions for a model owner.

One-day course outline

Morning - XAI foundations and types of explanation: Introduce AI predictions, explainability, and why users, affected people, managers, and technical teams may ask different questions. Contrast an inherently interpretable model, such as a small decision tree, with post-hoc explanations of a more complex model. Distinguish global descriptions of model behaviour from local explanations of a single prediction. Interpret an instructor-provided feature-importance chart and a local feature-attribution example for a fictional service-request prioritization model. Discuss why rankings or attributions can vary with data, method, and reference point and do not by themselves prove that a feature caused an outcome. Workshop: Compare the global and local examples, identify what each can and cannot answer, and record questions about the data and explanation.

Afternoon - Counterfactuals, limitations, and communication: Explore a counterfactual example showing how changed inputs could alter a model prediction, and question whether those changes are feasible or appropriate. Discuss explanation stability, missing context, correlated inputs, data quality, privacy, and the risk of overconfidence. A plausible explanation does not establish that a prediction is correct, fair, or suitable for a decision. Contrast model explanations with a generative AI assistant's natural-language justification, which is not necessarily a faithful account of its internal process; verify factual claims against reliable sources. Workshop: Review a fictional prediction and supplied explanation, identify limitations and possible impacts, draft an audience-appropriate explanation brief, and specify human review, challenge, and escalation steps.

Training scope: Exercises use fictional or instructor-provided information and prepared examples; no model building is required. Do not enter real personal, confidential, or proprietary data into an AI tool without authorization and approved safeguards. This introductory course does not provide legal advice or certification, or guarantee fairness, accuracy, or regulatory compliance. Consult appropriate legal, privacy, accessibility, and technical specialists for real-world decisions.

Format and audience

For: Business users, managers, analysts, project and product teams, and technical, risk, privacy, or compliance professionals who want an introduction to explainable AI.

Prerequisites: Basic computer and workplace experience. No coding, statistics, or prior AI knowledge is required.

Workshop setup: A computer with a modern browser and an application for viewing and editing documents or worksheets. The instructor supplies fictional scenarios, charts, explanations, and templates; no AI tool account is required.

How it works: One day of instructor-led explanations, discussions, case studies, and guided workshops. Contact us to discuss dates and virtual or onsite delivery.

Explore the course

Request a quote for Explainable AI (XAI) Fundamentals and discuss your team's learning goals.

Request a Quote

Back to AI & ML training

Responsible AI Fundamentals · AI Governance Fundamentals