Join the Learning and Development Community

ADDIE is a five-phase process for designing training: Analysis, Design, Development, Implementation, Evaluation.

• Analysis: link the training to a business outcome, study the department and job, build a task inventory, choose which tasks people actually need to learn, set performance measures for those tasks, choose the setting (classroom, on-the-job, self-study, blended), and estimate cost against benefit.
• Design: write terminal and enabling objectives, list the learning steps, build performance tests to show mastery, check what learners must already know before starting, and sequence the objectives.
• Development: choose activities and delivery methods, review existing material so you don't reinvent it, build the courseware, put it together as a complete programme, and validate that it meets its goals.
• Implementation: create a management plan, then deliver the training.
• Evaluation: review every phase, check in the workplace whether learners can actually perform the trained tasks, and revise the programme.

The core point of the article is that evaluation isn't just at the end — it runs through every phase, and the real test is on-the-job performance, not just course completion.

Based on SmarterX's 2026 State of AI for Business report, surveying more than 2,100 professionals across industries and company sizes.

When asked to pick up to three AI topics they most wanted more training on, people chose:

• Integrating AI into existing workflows — 58%
• Using AI agents — 51%
• Building no-code assistants — 45%
• Prompting — only 15%

The article's point is that this isn't beginners asking for an introduction. These are people who already use AI and are now asking for intermediate to advanced, operational training. Prompting was the centrepiece of most corporate AI training, but it's barely on the list now. AI agents are also the number one AI trend these professionals say they're following most closely, which lines up with what they want to learn.

One caution the article itself notes: the respondents already engage with SmarterX, so they're proactive AI users, not a random sample of all workers.

Seven trends, with a practical warning running through all of them: don't adopt everything at once.

• AI-powered personalised learning: platforms recommend modules based on role, assessment results, and progress, and help draft quizzes and content — but human review is still needed for accuracy and data protection.
• Skills-based learning: focus on specific competencies, not job titles, and connect development to career progression and internal mobility.
• Microlearning: short lessons (five-minute videos, quick quizzes) for onboarding, software updates, and compliance — but they still need structure, or you end up with a disconnected library.
• Learning in the flow of work: guidance inside the tools people already use, at the moment they need it, rather than a separate long course.
• Simulations: practice decisions in realistic situations without real consequences — safety, customer conversations, equipment, leadership — with feedback and a chance to try again.
• Custom eLearning: generic libraries don't cover a company's own processes, products, or compliance needs, so tailored courses and LMS customisation matter.
• Learning analytics: move beyond completion rates to assessment results, retention, engagement, workplace performance, and time to competency.

The article's recommended practice steps are: identify the problem first, choose the format that fits it, check technology, accessibility, reporting, and security, pilot with a small group, then measure against metrics defined in advance.