Knowing the rule is not the same as using it under pressure

A course can explain the approvals policy perfectly and still leave the person on the phone unsure what to say when a customer pushes back. A manager can pass the quiz on the change process and still freeze in the first hard conversation with their team. The gap is practice. People learn to make a call by making it, seeing what follows, and trying again.

Most training never gives them that chance, because practice is the expensive part to build. In an authoring tool a branching scenario is a tree of screens, with every path written and wired by hand. A role-play is harder still. So the practice gets cut to a multiple-choice question, and the course goes out as reading with a test at the end.

Katie builds the practice as part of the training. You brief her on the decision or the conversation in plain words, your experts supply the real situations, and she designs and builds the scenarios for your approval. Inside the finished training she also role-plays with learners herself. If the role is new to you, the guide to the AI instructional designer explains it in full.

The practice Katie builds

The portfolio shows the formats. Each one is built around a situation from the learner's own work, not a generic case.

A branching scenario

A decision at work, with the options in front of the learner. They weigh the options, see the consequences of the one they chose, and try again with feedback. The policy becomes everyday judgement rather than a page to read.

A role-play simulation

A space to rehearse the conversations that matter: the pushback from a customer, the awkward exchange with a colleague, the call nobody wants to make. Realistic prompts and guided reflection turn a difficult exchange into useful practice before it happens for real.

A leadership scenario

A chance to work through the decisions that come with leading a team through change. Learners consider different perspectives and reflect on how their response affects the people around them.

The AI instructor, who role-plays in the moment

Beyond the scenarios built into the course, learners can chat with Katie inside the training. She asks questions, role-plays realistic scenarios and gives instant feedback, grounded only in the material your team approved. The learner gets a conversation partner who stays inside the lines you drew.

How Katie builds a scenario, step by step

She starts from the decision, not the content

Bring a business change, a performance gap or a vague training request. Katie helps define the outcome, the audience and the key behaviours, which for a scenario means the decision people face and where they get it wrong today. She writes the strategy and the training plan for your approval.

Your experts supply the real situations

The situations that make a scenario believable live in your experts' heads: the customer who always asks the same thing, the exception nobody wrote down. Your subject-matter experts provide that knowledge, answer Katie's questions and review the drafts for accuracy. Katie runs the back-and-forth herself.

She proposes the scenarios before she builds

In the new-hire demo on the homepage, Katie's first reply to a request is "I drafted the training outline and two practice scenarios. Take a look, and once approved, I'll start building." You review the scenarios as a short outline, not as a finished tree you would have to unpick.

She builds and revises in chat

From the approved plan she produces the complete training: text, video, practice, simulations and assessments. Your team and your experts give feedback in plain English on Teams, Slack or WhatsApp, and Katie makes the changes right away until it is right.

She tailors the scenarios for each audience

In the IT onboarding demo the request is "Support policy is approved. Tailor a version for IT." Katie's reply: "IT version ready. Core policy stays the same, with IT-specific scenarios and quiz questions set for your approval." The same policy, with practice that fits the team.

A worked example: a decision scenario for frontline teams

This is the frontline training request from Katie's own demo on the homepage, from the first review to the learner's practice.

The course as first built

The L&D manager asked for training on the approvals policy and handed Katie the policy document. Katie defined the need, the manager approved the plan, and the policy training on approval limits arrived in the team's chat marked ready for review.

The review

The manager reads it with the frontline teams in mind and sends one note: "The policy section is too dense for frontline teams. Split it into two parts and add a decision scenario."

What Katie builds

Her reply is "Done. Preview the new version here." The preview card now reads policy essentials, frontline training. The dense section is two short parts, and between them sits a decision scenario built the way the portfolio describes a branching scenario: a situation a frontline employee meets, the options they have, the consequence of each, and feedback that sends them back to try again.

The expert's check

The manager asks the policy owner to read the scenario. The expert replies in the chat that one consequence is too mild for what would really happen. Katie changes it and shows the new version.

Approval and publishing

The team approves the complete course. Katie handles the SCORM export, and the team packages it for the LMS or shares a direct link.

In the training

A frontline learner works through the decision scenario, picks the wrong option, sees what follows, and tries again. Then they ask Katie, inside the training, what they should say if the customer pushes back. She role-plays the exchange and gives feedback grounded only in the approved policy.

A version for IT

When the support policy is approved, the request "Tailor a version for IT" gets IT-specific scenarios and quiz questions, with the core policy unchanged, set for approval.

After launch

Katie analyses learner errors, questions and drop-offs. If learners keep choosing the same wrong option, she recommends a fix for the team to review and publish.

What your team approves, and when

  • The objectives and the structure, before production. For practice, that means the decisions and the scenarios in outline. Nothing is built until you say they are the right ones.
  • The complete training, before publishing. You and your experts review every scenario and every consequence against what would really happen. Katie learns your terminology, business context and standards from each approval.
  • Every later change. A new scenario, a version for another department or a rule change is prepared for review and published only when you decide.

The AI instructor stays inside the lines you drew: its questions, role-plays and feedback are grounded only in the material your team approved. You own the source materials and the finished training, including if you stop working with Katie. Before rollout, Trust Katie starts a security review covering your content, learner data, the AI providers involved and your security requirements.

What the learner gets

A chance to practise before it counts. Learners meet situations drawn from their own work, make the call, see the consequence and try again with feedback. The conversations they dread get rehearsed in a space where getting it wrong costs nothing.

Then they get a partner. Instead of clicking through slides, learners chat with Katie, who asks questions, role-plays realistic scenarios and gives instant feedback. Alt text, captions and audio description are part of every training, and Katie works in more than 20 languages with full right-to-left support, with local teams reviewing each adaptation before rollout.

What it costs in workload

Katie is priced on the workload she takes off your team, never per creator or learner seat. One credit is one cent, and she shows the exact cost before every action.

The rate card prices the work, not the format, so there is no separate line for a role-play scenario. Practice and simulations are part of a training build: one build of up to two chapters is 2,000 credits, and a typical training on the card's own arithmetic is 2,555 credits, including a 14-message design interview at 30 credits a message, five AI edits at 15 credits each and three images at 20 credits each. A change made during review, such as the frontline note above, is made in chat; an edit or regeneration of one item with AI is 15 credits.

On the learner side, a learner completing a chapter, with the AI tutor included, is 25 credits a learner a chapter. The card's worked month counts 500 learners completing a two-chapter training inside the listed plan of $1,000 a month for 100,000 credits. One line to read carefully: the card's software simulation from screenshots or a recording, at 500 credits a simulation, is a system walkthrough, a different job, which the software training from screenshots page covers. The pricing page has the full rate card and an estimator for your own month, and the pilot is eight weeks with one team, on one real training, from brief to publish.

Or bring the situation to a 30-minute scoping call.