LegitScript analyst enablement · powered by Longhand

Start here.LegitScript analyst onboarding.

Go from your organization invitation to one supervised, evidence-based classification review—with clear stop points before live work.

8 observable steps45–50 minutesmacOS + WindowsProgress saved

The finish line

One supervised classification completed

You are in the right workspace, can route work between queues and My Work, and can show your lead exactly what happened in the first run.

Progress is saved in this browser.

What good onboarding proves

Access is only the beginning. You should be ready to operate safely.

01

Join the LegitScript workspace with the invited identity

02

Recognize the applied Analyst experience before starting work

03

Explain where work comes from and where your judgement is needed

04

Start only the queue and batch your lead approved

05

Review every submitted field against its evidence

06

Recover or escalate with an exact, observable handoff

The goal is not a tour of every feature. It is one complete, supervised loop that your LegitScript lead can verify.

The guide

Eight steps from invitation to an approved first run.

How the work moves

The system carries the work. You own the judgement gate.

Think of onboarding as learning one operating loop—not learning every Longhand feature.

Stage 1

Classification queues

Choose the queue your lead approved and start the next available batch.

Stage 2

My Work

Watch the resulting runs move through Running, Failed, Needs Review, and Done.

Stage 3

Review console

Judge every field against its evidence and record exact corrections.

Stage 4

LegitScript database

Only a reviewed result continues to the governed submission step.

The onboarding stop rule: do not start a batch until your lead names the practice queue, and do not submit a record whose evidence does not support your decision.

Step 1 of 8 · 3 min

Confirm the identity, queue, and supervisor for your first session.

Prepare your invitation

Your invitation is tied to a specific email address, and the first practice batch can become real classification work. A two-minute preflight prevents the two expensive mistakes: joining with the wrong identity or starting the wrong queue.

Do this

1Find the Longhand invitation to the LegitScript organization and keep it open.
2Confirm the exact work email address that received the invitation. You must use that same address when you sign in.
3Ask your LegitScript onboarding lead which queue is approved for your supervised practice run.
4Confirm whether the practice run is allowed to submit to the LegitScript database or must stop at review.

Remember: Do not use a personal email or a different work alias. Longhand accepts the pending organization invitation when the invited identity signs in.

What you leave with

Invitation email ready

Correct sign-in email confirmed

Practice queue named

First-run stop rule confirmed

Step 1 checkpoint

Your preflight is complete

Practice evidence target

You can name the invited work email, the designated practice queue, the person supervising the session, and the approved stopping point for the first run.

Recording this checkpoint saves self-reported lesson practice in this browser. It is not evidence of certified competency; certification requires a separately submitted artifact and qualified human evaluation.

Step 2 of 8 · 5 min

Use the invitation handoff first, with the download page as the fallback.

Install and open Longhand

The invitation link tries to open the installed desktop app before sending you to the download page. Following that path preserves the invited email handoff and avoids treating a normal browser prompt as a failed invitation.

Do this

1Click Open Longhand in the invitation email.
2If your browser asks permission to launch Longhand, choose Allow or Open Longhand.
3
If the app is not installed, use the Download option and install the correct version:
- macOS: install the Apple Silicon version, then open Longhand from Applications.
- Windows: run the Windows installer, then open Longhand from the Start menu.
4Return to the invitation email and click Open Longhand again if the app opened without the invitation context.

Installation paths

Platform	Install	Open from
macOS	Open the downloaded disk image and install Longhand	Applications
Windows	Run the downloaded installer	Start menu

Step 2 of 8 · 5 min

Use the invitation handoff first, with the download page as the fallback.

If you need the download fallback

Do this

2If your browser asks permission to launch Longhand, choose Allow or Open Longhand.
3
If the app is not installed, use the Download option and install the correct version:
- macOS: install the Apple Silicon version, then open Longhand from Applications.
- Windows: run the Windows installer, then open Longhand from the Start menu.

Longhand download page with buttons for Mac and Windows — Use the download page only when the invitation cannot open the installed app.

Step 2 checkpoint

Longhand opens from the invitation

Practice evidence target

Longhand opens as a desktop app on your computer, and clicking the invitation's Open Longhand button brings the app to the foreground.

Recording this checkpoint saves self-reported lesson practice in this browser. It is not evidence of certified competency; certification requires a separately submitted artifact and qualified human evaluation.

Step 3 of 8 · 5 min

Sign in once; Longhand accepts the pending organization invitation for you.

Join with your invited identity

LegitScript has already provisioned the workspace, policies, billing, shared catalog, and analyst experience. Signing in with the invited identity is the step that connects this installation to that organization.

Do this

1Choose Sign in with Longhand when the app asks how you want to set up access.
2Use the exact work email address that received the LegitScript invitation.
3Complete the browser sign-in and allow the browser to return to Longhand.
4Wait for Longhand to land on Home. A LegitScript organization member should skip the general self-serve setup wizard.

Remember: Do not paste a LegitScript, Anthropic, OpenAI, or other organization secret into the app. Organization-provided access is applied through the signed-in workspace.

What each outcome means

What you see	What to do
Home with the LegitScript experience	Continue to the workspace check.
A browser sign-in loop	Stop and capture the exact browser or app message.
The general Connect an AI Provider wizard	Do not add a personal key; stop and contact your onboarding lead.
A different organization	Stop. Confirm the invited email before trying again.

Step 3 of 8 · 5 min

Sign in once; Longhand accepts the pending organization invitation for you.

A successful organization sign-in looks like this

Do this

2Use the exact work email address that received the LegitScript invitation.
3Complete the browser sign-in and allow the browser to return to Longhand.
4Wait for Longhand to land on Home. A LegitScript organization member should skip the general self-serve setup wizard.

Longhand Home screen with the LegitScript organization experience loaded — The LegitScript wordmark and Home screen confirm the organization experience loaded. Example metrics are obscured.

Step 3 checkpoint

The organization sign-in landed

Practice evidence target

The sign-in finishes, Longhand returns to Home, and you are not being asked to configure a personal AI provider as part of a general first-run wizard.

Recording this checkpoint saves self-reported lesson practice in this browser. It is not evidence of certified competency; certification requires a separately submitted artifact and qualified human evaluation.

Step 4 of 8 · 4 min

Make sure the applied experience is LegitScript's before starting work.

Verify the Analyst workspace

A successful sign-in is not the same as being in the right working context. The LegitScript Analyst workspace is intentionally focused: it carries LegitScript branding and puts run review above the classification queues.

Do this

1Confirm you can see the LegitScript wordmark and dark-blue organization accent.
2Open Agents and confirm My Work appears above Classification queues.
3Confirm Projects is not the primary analyst work surface; shared classification work comes from the queues.
4If any of these checks fail, stop before starting work and send your onboarding lead a screenshot.

Workspace acceptance check

Expected	Stop signal
LegitScript wordmark and dark-blue accent	Stock Longhand branding only
Agents navigation	A generic Projects-first workspace
My Work above Classification queues	Either surface missing
No personal organization secret requested	A request for a LegitScript or team API key

Step 4 of 8 · 4 min

Make sure the applied experience is LegitScript's before starting work.

This is the Analyst workspace

Do this

1Confirm you can see the LegitScript wordmark and dark-blue organization accent.
2Open Agents and confirm My Work appears above Classification queues.

Longhand Agents view with My Work above the Classification queues — My Work appears above Classification queues. Live identifiers and counts are obscured.

Step 4 checkpoint

You are in the right workspace

Practice evidence target

You can point to the LegitScript brand, the My Work board, and the Classification queues on the applied Analyst experience.

Recording this checkpoint saves self-reported lesson practice in this browser. It is not evidence of certified competency; certification requires a separately submitted artifact and qualified human evaluation.

Step 5 of 8 · 5 min

Know where work comes from and where your judgement is needed.

Learn the two work surfaces

The screen is designed around one simple distinction: Classification queues are where more work comes from; My Work is what you already owe. Keeping those roles separate prevents duplicate starts and missed reviews.

Do this

1Use Classification queues to see the shared backlog and start the next approved batch.
2Use My Work to follow each run through Running, Failed, Needs Review, and Done.
3Open a Needs Review card to enter the field-by-field review console.
4Open a Failed card to read the failure before choosing a recovery action.

The analyst's operating map

Surface or state	Meaning	Your next action
Classification queues	Shared work available to the team	Start only the queue your lead approved
Running	Longhand is executing the workflow	Let it run; inspect only if it stalls
Needs Review	The run is deliberately waiting on a person	Open the review console
Failed	The run stopped before completion	Read the cause before retry or re-drive
Done	The run reached its terminal result	Verify the supervised outcome, then clear only if appropriate

Step 5 checkpoint

You can route the next action

Practice evidence target

Given any item on screen, you can say whether it is waiting to be started, currently running, waiting on your judgement, failed, or finished—and which surface owns the next action.

Recording this checkpoint saves self-reported lesson practice in this browser. It is not evidence of certified competency; certification requires a separately submitted artifact and qualified human evaluation.

Step 6 of 8 · 8 min

Begin work from the designated queue and let the run launch automatically.

Start one supervised batch

Starting work is intentionally a queue-level action. Longhand safely claims the next available items for you and launches the matching runs, so analysts do not manually coordinate ownership or start each record one at a time.

Do this

1Have your onboarding lead present or explicitly authorize the practice start.
2Locate the designated queue and read its waiting count, batch size, and estimated cost.
3Click the queue's Start next N sites or Start next N products button once. The button's number is the amount of work this click will take.
4Keep Longhand open while the runs appear in My Work and move into Running.
5Do not press a separate Run control for individual queue items; a successful claim starts the work automatically.

Remember: This is the first action that starts real processing and may eventually write to the LegitScript database. Do not perform it unsupervised during onboarding.

What the start action guarantees

You do	Longhand does
Choose the approved queue	Scopes the claim to that queue
Click the queue's Start next N button once	Uses first-claim-wins so teammates do not duplicate work
Keep the app open	Claims the batch and launches the corresponding runs
Watch My Work	Moves each run toward review, failure, or completion

Step 6 of 8 · 8 min

Begin work from the designated queue and let the run launch automatically.

Start from the approved queue

Do this

2Locate the designated queue and read its waiting count, batch size, and estimated cost.
3Click the queue's Start next N sites or Start next N products button once. The button's number is the amount of work this click will take.
4Keep Longhand open while the runs appear in My Work and move into Running.

Longhand classification queue cards with a Start next 10 sites button — The button tells you how many items this click will claim. Your approved queue and batch size may differ.

Step 6 checkpoint

One click became the expected runs

Practice evidence target

One authorized click on the designated queue creates the expected runs in My Work, with no duplicate manual starts and no work taken from a different queue.

Recording this checkpoint saves self-reported lesson practice in this browser. It is not evidence of certified competency; certification requires a separately submitted artifact and qualified human evaluation.

Step 7 of 8 · 12 min

Use evidence to judge every field before the run is allowed to continue.

Review and correct one result

The AI handles the middle of the workflow, but the analyst owns the control point before submission. The console turns that responsibility into a field-by-field decision with rationale, evidence, corrections, and a visible handoff to the next record.

Do this

1Open a card in Needs Review, or use the column's review shortcut to open the next waiting run.
2Read the record identity, rationale, evidence links, and any jurisdiction or network context before judging the fields.
3Mark each field Correct or Incorrect. When a field is incorrect, enter the exact corrected value or the additional detail required.
4Use Mark all correct only after you have actually checked every displayed field.
5Choose Submit & next record or Submit & next run. Longhand merges your corrections, resumes the workflow, and advances the queue.
6If the evidence does not support a decision, stop and escalate instead of guessing.

Remember: Submission is the control point that lets the workflow continue toward the LegitScript database. A fast skim is not the same as a completed review.

Review decisions

Decision	Use it when	Required follow-through
Correct	The evidence supports the proposed value	No edit; continue to the next field
Incorrect	The proposed value is wrong or incomplete	Enter the corrected value or missing detail
Stop and escalate	The evidence is missing, conflicting, or outside your authority	Do not guess or submit; capture the record and question

Step 7 of 8 · 12 min

Use evidence to judge every field before the run is allowed to continue.

This is the human judgement gate

Do this

2Read the record identity, rationale, evidence links, and any jurisdiction or network context before judging the fields.
3Mark each field Correct or Incorrect. When a field is incorrect, enter the exact corrected value or the additional detail required.
5Choose Submit & next record or Submit & next run. Longhand merges your corrections, resumes the workflow, and advances the queue.

Longhand Website Classification Review with evidence and field-level Correct or Incorrect controls — Read the evidence, check every field, then submit. The example record and live values are intentionally obscured.

Step 7 checkpoint

Your judgement is visible

Practice evidence target

You complete one field-by-field review using the evidence, record any necessary correction, and submit only after every displayed field has been deliberately checked.

Recording this checkpoint saves self-reported lesson practice in this browser. It is not evidence of certified competency; certification requires a separately submitted artifact and qualified human evaluation.

Step 8 of 8 · 5 min

Close the loop with an observable result and a clear escalation path.

Verify, recover, and hand off

Onboarding is complete when the system and a human agree on what happened. A Done card is useful evidence; a Failed card with the exact cause is also useful evidence. Hiding the state or retrying blindly teaches the wrong operating habit.

Do this

1Return to My Work and verify that the supervised run reaches Done—or capture the exact state where it stopped.
2If it failed, open the run and read the cause before using Try again, Retry, or Re-drive. Follow your lead's direction for the correct recovery path.
3Do not dismiss an active failure merely to clean the board; dismissal moves it out of the active view without fixing the cause.
4Copy the onboarding check, fill it out, and send it to your LegitScript onboarding lead with any screenshot or exact error text.
5Wait for your lead to approve independent queue work.

Recovery rule

State	Safe response
Done	Record the result and ask for sign-off
Needs Review	Finish or escalate the field-level decision
Failed with a clear transient cause	Retry only with lead approval
Failed because the run no longer owns its item	Use Re-drive when the app offers it
Unknown or repeated failure	Stop; send the exact state, run, and error

Step 8 of 8 · 5 min

Close the loop with an observable result and a clear escalation path.

Failure evidence is a stop signal, not cleanup

Do this

1Return to My Work and verify that the supervised run reaches Done—or capture the exact state where it stopped.
2If it failed, open the run and read the cause before using Try again, Retry, or Re-drive. Follow your lead's direction for the correct recovery path.
3Do not dismiss an active failure merely to clean the board; dismissal moves it out of the active view without fixing the cause.

Longhand failed workflow step showing the cause, evidence, and recovery context — Open the failed step and read the cause before any retry or re-drive.

Step 8 of 8 · Run it

Send one complete onboarding check

Do this

4Copy the onboarding check, fill it out, and send it to your LegitScript onboarding lead with any screenshot or exact error text.
5Wait for your lead to approve independent queue work.

Copyable onboarding handoff

LEGITSCRIPT ANALYST ONBOARDING CHECK

Analyst:
Invited work email:
Operating system:
Onboarding lead:

ACCESS CHECK
LegitScript workspace visible: Yes / No
LegitScript branding visible: Yes / No
My Work visible: Yes / No
Classification queues visible: Yes / No
Personal API key requested: Yes / No

SUPERVISED PRACTICE
Designated practice queue:
Practice run ID or name:
Result: Done / Needs Review / Failed
Review submitted: Yes / No

FOLLOW-UP
Issue or question:
Screenshot or exact error text:
Approved for independent queue work by:

Practice evidence target

Your onboarding lead receives a completed check with the workspace evidence, practice run, final state, review status, and any exact error—and explicitly approves or assigns the next step.

Step 8 checkpoint

Your lead can verify the outcome

Practice evidence target

Your onboarding lead receives a completed check with the workspace evidence, practice run, final state, review status, and any exact error—and explicitly approves or assigns the next step.

Recording this checkpoint saves self-reported lesson practice in this browser. It is not evidence of certified competency; certification requires a separately submitted artifact and qualified human evaluation.

You have completed the supervised loop

Now make the operating habit repeatable.

Keep the same discipline on every batch: confirm scope, let the system run, use evidence at the judgement gate, and leave a clear record when something stops.

Understand the model

Longhand AI Overview

Learn the gears, skills, value model, boundaries, and accountable review pattern behind the workflow.

Open training

Find the next lesson

Longhand AI Roadmap

Use the broader training map to close the next capability gap after onboarding.

Open training