The OneRange Vero Admin Guide

Build courses and pathways, run assessments, and design learning experiences your team will actually finish

For administrators, managers, and course owners · 20 min read

Getting started — Roles, navigation, key concepts
Building a course — From blank canvas to published lesson
Interactive modes — Standard training plus the four interactive block types
Reference docs & knowledge — Grounding the AI in your content
Assessments — Skill checks and competency baselines
Pathways — Sequencing courses into journeys
Assignments & notifications — Getting work to learners
Analytics & insights — Measuring impact and ROI
Company settings — Branding, domains, seats, knowledge library
Tips & tricks — Patterns from high-performing admins
Hands-on labs & scenario-based training — Deep dive on experiential training formats

1. Getting started

Roles, navigation, key concepts

Vero is an AI-powered interactive training platform. Admins design courses (short, chat-based lessons), group them into pathways, and assign them to learners. The AI tutor delivers each course as a guided conversation, checking understanding as it goes.

Roles at a glance

Admin	Full access. Manages users, billing, branding, all courses and pathways.
Manager	Builds courses, assigns training, sees team analytics.
Course Manager	Builds and edits courses and pathways but does not manage people.
Employee	Takes assigned training and self-launches available courses.

Where things live

Dashboard — your home view. Admins/managers see team activity; employees see To-Do and Completed.
Courses — the catalog of lessons your company can deliver.
Pathways — ordered (or unordered) groupings of courses.
Assessments — standalone skill checks (4 AI-generated questions per skill).
Employees — invite, manage roles, view individual training history.
Analytics — workforce skill data, ROI, time-to-competency.
Settings → Knowledge — your company's reference library used to ground the AI.

Tip: Bookmark Settings → Knowledge early. Courses become dramatically more accurate the moment you upload a few real reference docs.

2. Building a course

From blank canvas to published lesson

A course is a focused, ~10–30 minute training conversation on a single topic. The flow is intentional: short explanation → concrete example → quick check. Repeat. End with a graded assessment.

The 3-step builder

1. Brief	Name, description, target level, duration, and the 5 skills the course covers.
2. Curriculum	AI generates a topic outline + assessment questions. Edit, regenerate, or accept.
3. Confirm & Publish	Nothing is saved until you click publish — experiment freely.

Filling out the brief well

Name: be specific. "Handling pricing objections in B2B SaaS" beats "Sales 101".
Description: describe the training itself — what the course covers, what the learner will do inside it, and the outcome or goal they should walk away with. Treat it as a mini-syllabus, not a tagline. Aim for 2–3 sentences: (1) what the training is about, (2) what it contains (topics, formats, who it's for), (3) the outcome the learner achieves. Omit trailing periods to match the in-app style.
Target level: New, Intermediate, or Experienced — drives how much scaffolding the AI provides.
Duration: use the stepper. Shorter courses (10–15 min) finish more often than 45-min ones.
Skills: pick exactly 5. The catalog has 10k+ skills, classified Hard or Soft.

Description examples

Weak	"Learn about pricing objections" — vague, no contents, no outcome.
Better	"A 20-minute training session for AEs on handling pricing objections in B2B SaaS deals. Covers discount framing, value re-anchoring, and a role-play with a skeptical procurement lead. By the end, you'll be able to defend list price in a live call without losing rapport."
Outcome-led	"Onboarding for new SDRs covering our discovery framework, objection patterns, and a branching scenario based on a real Acme account. Goal: book a qualified meeting on your first cold call by week two."

Tip: Descriptions also feed the AI. A clear contents-and-outcome description steers curriculum generation toward the right topics on the first pass — saving you regenerate cycles.

Curriculum review

Each topic has an explanation, an example, and check questions.
Click Regenerate on any topic if it feels generic — give it a steering note.
Aim for 4–7 topics. More than 8 usually signals the course should become a pathway.
The assessment is multiple-choice, with rationales for every option.

Tip: Use Preview Mode before publishing — you'll see the same welcome message learners see and can sanity-check tone in 30 seconds.

Editing after publish

You can improve a published course from the Courses page. Be aware that regenerating content shows a confirmation dialog — once accepted, the previous topic content is replaced.

3. Interactive modes

Standard training plus the four interactive block types

Every Vero course runs in Standard training mode by default — a guided chat conversation. On top of that, the AI tutor can emit one of four interactive blocks per topic when the format fits. You choose which interactive types are eligible in the course brief; the AI decides per topic which one (if any) to emit.

Standard training (the normal mode)

ChatGPT-style conversation: short explanation → concrete example → quick check.
Inline A/B/C/D quick-reply buttons for check-ins (always available, not a separate mode).
Bold text for emphasis and underlined questions at the end of each turn.
Used for any topic that does not need a richer interactive block.

The four interactive block types

Branching scenario	Multi-step decision simulation. Learner picks A/B/C/D and sees consequences. Best for ethics, sales discovery, leadership judgment, escalation calls.
Guided hands-on lab	Step-by-step checklist with checkpoints. The AI verifies progress as the learner works through a procedure or tool. Ends when the learner signals completion. Best for SOPs, tools, and operational workflows.
Role-play simulation	AI plays a persona (angry customer, skeptical exec, junior report) and the learner takes the counterpart role. Ends when the conversation reaches a natural close. Best for soft skills — coaching, feedback, objection handling.
Code lab	In-browser Monaco editor with a Deno JS/TS sandbox. Learner writes and runs code against checks. Best for engineering, data, and scripting courses.

How to request them

In the course brief, open Interactivity preferences and toggle the types you want eligible across the course.
After topics generate (or when adding a topic manually in Build Topics), each topic shows interactivity tags you can click to toggle on or off per topic — no regeneration needed.
Click an active tag to open its configuration: add a short description of what the block should cover, and attach supporting documents.
Documents can be (a) uploaded fresh, (b) re-used from the Step 1 reference uploads, or (c) pulled from the company Knowledge Library — all selectable inline without re-uploading.
Drag topics by the grip handle on the left to reorder them before publishing.
Less is more. Pick the 1–2 formats that genuinely fit the skill — the rest stays in Standard training.
Learner-facing chat strips the markers; only the rendered block appears.

Editing interactivity after generation

Deselecting a tag on an AI-generated topic just hides it — toggle it back on anytime without regenerating the course.
Per-topic descriptions and attached documents are preserved with the topic and passed to the AI tutor at runtime.
Use Preview Mode to see exactly how each interactive block will render for learners before publishing.

Other graded elements

End-of-course quiz	Multiple-choice with rationales. 70% to pass. Automatic — not part of the four interactive modes.
Standalone Assessment	4 AI questions per skill, in-memory. Used for baselines and re-certification.
Competency baseline	Optional 3–5 question diagnostic at the start of a course.

Tip: Pair a role-play with a branching scenario for soft-skill courses: the role-play builds the muscle, the branching scenario tests judgment.

4. Reference docs & knowledge

Grounding the AI in your content

The single biggest lever for course quality is grounding the AI in your real content. Vero supports two ways to do this.

Per-course reference docs

Upload PDFs, DOCX, slides, or text in the course brief.
Files are validated and scored — you'll see a quality badge before they're used.
Max 5 skills per course; reference docs can be larger and richer.
The AI quotes and cites these documents during the lesson.

Company Knowledge Library

Settings → Knowledge. Upload once, reuse across many courses.
Use it for: brand voice docs, product specs, SOPs, compliance policies.
Documents are isolated to your company by storage RLS.

Tip: If a course feels generic, the fix is almost always "add one good reference doc." A 5-page internal playbook outperforms a 200-page textbook every time.

What makes a good reference doc

Specific to your way of doing things, not the generic industry version.
Cleanly extractable — searchable PDFs and DOCX work best; scanned images do not.
Under ~30 pages per file. Split monsters into focused chunks.
Strip internal-only acronyms or define them at the top.

5. Assessments

Skill checks and competency baselines

Assessments are standalone skill checks — distinct from the quiz at the end of a course. Use them to baseline a team or verify mastery.

Each assessment covers 1–5 skills. The AI generates 4 questions per skill.
Questions are multiple-choice with rationales.
Optional Competency Baseline: a 3–5 question diagnostic at the start of a course.
Results feed directly into Workforce Analytics, contributing to gap-closure metrics.

When to use which

Pre-training baseline	Standalone Assessment, assigned before a pathway starts.
In-course check	Built-in quiz (automatic) or competency baseline (opt-in).
Re-certification	Reassign the same assessment quarterly to track drift.

6. Pathways

Sequencing courses into journeys

Pathways are ordered or unordered collections of courses. Use them to package onboarding, role-readiness, or certification programs.

Strict order: learner must complete courses in sequence.
Any order: learner picks the order, must complete all to finish the pathway.
Pathways can be shared via secure link (JWT) with non-employees for previews.
Pathway completion shows up as a single milestone in analytics.

Tip: Keep pathways under 6 courses. Long pathways have steep drop-off after course 4.

7. Assignments & notifications

Getting work to learners

Assign from a course, pathway, or assessment page — or from an employee's profile.
Set a due date. Learners get an email at assignment and a 7-day reminder.
Unassign is only allowed while status is Not started.
Self-launch: any published course visible to the employee can be started without an assignment.
Bulk outreach: upload a CSV of external addresses to send demo links.

Tip: Stagger due dates across a team. When everyone has the same Friday deadline, completion quality drops on Thursday afternoon.

8. Analytics & insights

Measuring impact and ROI

The Analytics page rolls up activity into a workforce view. Filter by Hard vs Soft skills and by team.

Key metrics

Skill Uplift	Average proficiency change before vs after training.
Time to Competency	Median time from assignment to passing the related assessment.
Gap Closure	% of identified skill gaps closed in the period.
Training Efficiency	Outcome per hour invested — flags courses that are too long.
Internal Mobility	Promotions and lateral moves attributed to completed training.

Per-employee insights

View any employee's full history: assignments, time spent, completed assessments, login dates.
Each completed assignment gets an AI-generated performance summary (one-time).
Managers see their team only; admins see the whole company.

9. Company settings

Branding, domains, seats, knowledge library

Branding	Upload a logo and pick a brand color (HEX). Applied across the app and emails.
Domains	Verify a primary domain via DNS TXT. Used for SSO and login redirects.
Seats	Self-serve adjust up or down. Cannot drop below active employee count.
Knowledge	Upload company-wide reference documents and connector placeholders.
Roles	Promote/demote users. Site Admins can impersonate for support.
Feedback (NPS)	0–10 scale collected periodically; rolling 90-day score visible to admins.

Authentication

Vero supports Google and Apple SSO out of the box.
First-time Google users without a company go through a short onboarding flow.
Inactive subscriptions auto sign-out standard users (admins remain to fix billing).

10. Tips & tricks

Patterns from high-performing admins

Course design

One job per course. If you need the word "and" in the title, split it.
Lead with the hardest example. Learners trust the AI more when it tackles edge cases first.
Write the assessment first. If you can't write the test, the topic isn't scoped tightly enough.
Use real names. "Handling Acme's renewal call" hits harder than "Handling renewal calls".
Write descriptions like a syllabus. Spell out what the training contains and what success looks like — both your team and the AI use it as a north star.

Working with the AI

Regenerate aggressively. The second pass is almost always better than the first.
Steering notes beat editing. A one-line note in the regenerate prompt rewrites cleanly.
Don't fight the format. If you want a procedure, request a guided lab — don't cram it into prose.

Rollout

Pilot with 5 power users before assigning to the whole team. Their feedback shapes the rest.
Tag pathways by quarter (e.g. "Q1 Onboarding") so analytics stay readable as you grow.
Re-rate courses quarterly via the in-app rating dialog. Low-rated courses surface to you for fixes.

Things to avoid

Long courses with no interaction. Anything over 25 minutes needs a lab, role-play, or branching scenario.
Generic skills ("Communication"). Pick the specific child skill ("Written status updates") so analytics mean something.
Vague descriptions. "Learn about X" tells nobody what's inside or why it matters — use a contents-and-outcome format instead.
Uploading every doc you have. 3 great references beat 30 mediocre ones.
Assigning before previewing. Always run Preview Mode once.

11. Hands-on labs & scenario-based training

Deep dive on experiential training formats

Passive reading rarely changes behavior. Hands-on labs and scenario-based training are the two highest-leverage formats Vero offers — they put the learner in the driver's seat and force a decision or an action before moving on.

When to reach for which format

Guided hands-on lab	Procedural skills with a verifiable end state — running a SQL query, configuring a tool, closing a ticket, performing a safety check.
Code lab	Engineering and data work where the learner writes and runs real code in a multi-stage editor (intro → starter → hints → review).
Branching scenario	Judgment calls with no single right answer — discovery questions, escalation paths, ethics decisions, prioritization.
Role-play simulation	Conversational skills under pressure — angry customer, skeptical exec, underperforming report, partner negotiation.

Tip: Rule of thumb: if the skill is do this, use a lab. If the skill is decide this or say this, use a scenario or role-play.

Integrating into a course

Open Interactivity preferences in the brief and toggle the formats you want.
Use the per-format Advanced note to anchor the AI: persona, tools, the messy detail.
Re-run Generate curriculum after toggling so the topic outline reflects the formats.
In Preview Mode, walk one topic to see the format actually fire before publishing.

Writing great steering notes

Lab note	"Walk through configuring a new SSO connection in our admin panel — primary domain, IdP metadata XML, test login, then disable."
Code lab note	"Three stages: write a function that parses our CSV export, then add error handling for missing columns, then optimize for 1M rows."
Branching note	"Tier-2 escalation from a frustrated enterprise customer threatening churn. 4 steps. Bad choices should escalate further."
Role-play note	"VP of Sales who thinks the new pricing model will tank Q4. Pushy, data-driven, interrupts. Break character after 6 exchanges."

Designing the experience

Hands-on labs: 5–8 steps, each with a single verifiable action; include at least one step that can fail.
Code labs: multi-stage with starter code that compiles, plus test cases for self-verification.
Branching scenarios: 3–5 steps, 2–4 plausible choices each; consequences should compound.
Role-plays: name and emotional state for the persona; cap at 6–10 exchanges before the AI breaks character.

Common pitfalls

Stacking formats. One interactive block per topic.
Vague personas. Name them, give them a beef.
Labs without consequence. If clicking "I did it" never fails, it's a slideshow with extra steps.
Branching scenarios with one obviously-correct path. That's a quiz, not a scenario.
Skipping Preview Mode. Interactive blocks are where AI generation surprises you most.