How AI Helps L&D Teams Create Better Training in 2027

Start with the performance problem, not the AI tool

Before creating a training course or video, L&D teams should identify the specific knowledge or skills employees are missing. For example, if employees repeatedly make mistakes when submitting expense reports, an instructional designer should investigate whether they misunderstand the reimbursement policy, struggle with the software, or lack clear instructions.

Each problem calls for a different solution. Employees unfamiliar with the software might benefit from a demonstration and guided practice. Employees who misunderstand the policy need an explanation of its requirements. If the software itself prevents employees from completing a task, additional training may accomplish very little.

Effective instructional design also requires a clear learning objective. In their 2021 review of workplace instruction, researchers Kurt Kraiger and J. Kevin Ford identify 15 empirically supported instructional principles covering how information should be organized, presented, practiced, and reinforced.

For example, breaking a complicated procedure into manageable steps can help learners follow it. Providing opportunities to practice allows them to apply what they’ve learned, while feedback helps them recognize mistakes and improve. These principles give L&D teams a foundation for deciding what to teach and how to teach it, whether they create the materials themselves or use AI.

Where AI can help in the L&D workflow

AI can reduce the preparation work involved in developing training materials, particularly when instructional designers need to organize large amounts of information.

Consider an L&D team preparing training for a new customer relationship management system. Employees have submitted dozens of survey responses and support requests describing problems with the software. An instructional designer could provide those responses to an approved AI tool and ask it to identify recurring issues, such as difficulty finding customer records or updating sales opportunities.

Instead of manually sorting every response into categories, the designer receives an initial summary to investigate. They can review the original feedback, speak with employees, and determine which problems require training. Once the team identifies the learning objectives, AI can help draft course outlines, instructional scripts, and practice questions based on approved source materials.

Research suggests that AI can reduce some of this preparation work. In an October 2026 randomized trial, the National Foundation for Educational Research studied an AI lesson-planning assistant involving 464 teachers across 108 English primary schools. Teachers assigned to use the assistant spent approximately 24% less time preparing lessons and resources, with no statistically significant difference detected in the assessed quality of their materials.

The study examined schoolteachers rather than corporate L&D professionals, and it didn’t measure student learning. Nevertheless, it provides useful experimental evidence that AI-assisted educational preparation can reduce workload.

Instructional designers still need to direct the work. For example, a designer could provide an AI tool with an approved employee-feedback policy and ask it to draft a training scenario about a manager addressing an employee’s repeated missed deadlines. The designer would then review whether the conversation teaches appropriate management behavior, while a subject-matter expert verifies that its recommendations reflect company policy.

Match AI-generated training videos to the skill being taught

Instructional videos are particularly useful when employees need to observe a procedure or see how someone performs a task. A screen recording can show exactly where to click in a software interface, how to complete a form, or how a sequence of actions should unfold.

Research supports the educational value of video. A 2021 systematic review of 105 randomized studies involving 7,776 higher education students found that replacing existing instruction with video produced modest average learning improvements. Adding video to existing instruction produced larger average benefits. These results suggest that video can be a useful instructional resource, although they don’t establish that it’s preferable for every workplace task.

L&D teams can use Visla’s AI Video Agent to transform existing scripts, documents, presentations, and other materials into editable training videos. If the objective is to demonstrate software, Visla’s Screen Step Recorder can capture an employee performing the process and use AI to organize the recording into individual steps, complete with instructional text, visual annotations, and narration.

The resulting video remains editable. In Visla’s Scene-Based Editor, teams can adjust individual scenes, correct instructions, and replace outdated visuals. For organizations that regularly update their software or operating procedures, this reduces the amount of video production work required to keep training materials current.

Use scenario-based video to teach workplace judgment

Some workplace skills require employees to interpret situations and make decisions rather than follow a fixed procedure. Managers need to handle difficult conversations, customer service representatives must respond to unexpected problems, and sales professionals need to address objections without losing sight of a customer’s concerns.

Scenario-based learning uses realistic situations to help employees understand how knowledge and skills apply in practice. Video can make these situations particularly concrete by showing dialogue, reactions, and the consequences of different behaviors.

Consider a training video in which a manager discusses a missed project deadline with an employee. The manager explains that the delay prevented another team from completing its work. The employee responds that they were waiting for information from another department and didn’t realize they should have raised the problem earlier.

A well-designed scenario can show the manager listening to the explanation, identifying the communication problem, and agreeing on a better process for reporting delays. It could also demonstrate an unproductive response, such as immediately blaming the employee without investigating what happened.

Learners can compare the approaches and consider how they would handle a similar conversation.

A 2017 meta-analysis by Lacerenza and colleagues, covering 335 independent samples, found that leadership training improved learning and other outcomes, with effectiveness varying according to instructional design. Programs incorporating practice, feedback, and other appropriate training methods generally produced stronger results.

For L&D teams, the research supports designing scenarios around meaningful workplace behaviors and providing opportunities to practice them. Video can demonstrate those behaviors, while discussion, role-play, and assessments help learners apply what they’ve observed.

Realistic scenarios need meaningful instructional content

A realistic-looking conversation won’t necessarily teach employees how to handle a difficult situation. The scenario also needs to present the information, constraints, and decisions that make the situation challenging in real life.

A 2023 meta-analysis by Chernikova and colleagues examined 214 studies of simulation-based learning in higher education. The researchers found that simulations with greater functional similarity to real tasks were associated with stronger learning outcomes. Helping learners identify relevant information also contributed to effectiveness, while physical resemblance alone did not show a statistically significant moderating effect.

For our manager-training example, the employee’s explanation should therefore introduce a realistic complication. Perhaps another department repeatedly delivers information late, and the manager must decide how to address both the employee’s communication and the underlying workflow problem.

That creates a more meaningful training situation than a conversation where the employee simply apologizes and promises to improve. Learners have to consider competing responsibilities and decide what an appropriate response requires.

Create AI-generated scenario-based training videos with Visla

Traditional scenario videos often require actors, filming locations, equipment, and editing. Visla’s Scenario-Based Video offers an AI-generated alternative for creating dialogue-driven training videos featuring consistent characters, voices, and environments.

An instructional designer can start by describing the workplace situation, its characters, and the intended learning objective. They can also supply an approved script if they want greater control over the dialogue.

Visla uses that input to prepare the characters, assign voices, establish the environment, and organize the conversation into scenes. The designer can review the dialogue and visual directions, then examine AI-generated storyboard images before choosing which scenes to turn into AI video clips.

After generation, the team can refine the video in Visla’s Scene-Based Editor, adjusting the timing, changing dialogue, and regenerating individual clips when necessary. This allows an instructional designer to work with a subject-matter expert throughout production, rather than discovering problems only after the entire video is finished.

For a more detailed walkthrough, see our guide on how to create realistic scenario-based training videos with AI.

Add practice when learners need to make decisions

Watching a scenario can help employees understand appropriate workplace behavior. When the learning objective involves making decisions, employees should also have opportunities to choose a response and receive feedback.

For example, a manager-training video could pause after the employee explains why the deadline was missed. Learners would then select how the manager should respond. The training could explain the strengths and weaknesses of each choice or show how the conversation unfolds afterward.

A 2024 meta-analysis of interactive educational videos examined 17 studies and found that videos incorporating instructional interactions, such as questions and learning tasks, improved learning on average compared with videos lacking those features.

Visla can create AI-generated video clips representing different conversations, decisions, and outcomes. Teams can export those clips into an e-learning authoring tool such as Articulate Storyline 360 or Adobe Captivate, where they can add clickable choices, branching paths, and assessments.

Our guide to using Visla with scenario-based learning tools explains how these platforms work together.

Protect training quality with human review and accessibility checks

AI-assisted training needs a clear review process. A generated explanation can sound convincing while containing incorrect instructions, and a realistic workplace conversation can still recommend behavior that violates company policy.

Subject-matter experts should review technical accuracy, workplace procedures, and the situations presented in training scenarios. Instructional designers should examine whether the material addresses its learning objectives, provides appropriate examples, and gives employees useful opportunities to practice.

For the manager-feedback scenario, that means checking whether the dialogue reflects company expectations and whether learners can identify what the manager should do when the employee raises a legitimate obstacle.

Accessibility also needs attention throughout production. Training videos should include accurate captions, readable on-screen information, and descriptions of important visual content when necessary. Video players and interactive learning activities must also support accessible navigation. The World Wide Web Consortium’s audio and video accessibility guidance explains these requirements in greater detail.

AI can help generate captions, transcripts, and other supporting materials, but teams should review the output before publishing.

Measure whether AI-assisted training improves performance

AI can make producing training materials easier, but L&D teams still need to determine whether employees actually learn from them.

Production time, video views, and course completion rates provide useful operational information. They don’t tell us whether employees can perform the skills the training was intended to teach.

Assessment should reflect the learning objective. After a software tutorial, employees might complete a representative task without guidance. Following manager-feedback training, participants could conduct an observed role-play in which they’re evaluated on how they explain the problem, listen to the employee, and agree on next steps.

L&D teams should also consider training transfer, meaning whether employees apply their learning in real workplace situations.

A 2026 systematic scoping review by Ardondi and colleagues examined 45 studies of workplace-training evaluation tools. The researchers found considerable variation in how training transfer was assessed. Some tools measured employees’ reported application of new skills, while others focused on conditions associated with successful transfer.

For organizations, this means evaluation should extend beyond asking employees whether they enjoyed a course or feel more confident afterward. Where practical, teams can observe job performance, review relevant work outputs, or ask supervisors to assess specific behaviors over time.

Returning to the manager-feedback example, the most useful evidence would come from whether the manager can handle similar conversations effectively after completing the training. A successful role-play provides evidence of skill development, while subsequent workplace observations can help determine whether that learning carries over into everyday work.