Mon 24 Aug 2026

Whimsical

Stop Explaining Complex Safety Processes Only with Paragraphs

A safety procedure may be technically correct and still be difficult to follow.

Long paragraphs can hide decision points, responsibilities and stop-work conditions. This week, let’s explore Whimsical, an AI feature inside Whimsical that can turn a written prompt into an editable flowchart, mind map, sequence diagram or other visual workspace content.

Whimsical is not dedicated HSE software. However, we can use it to prepare the first visual draft of a safety process and then improve it through follow-up instructions. The generated objects remain editable on the board.

We can turn an HSE procedure into an editable visual workflow, without drawing every box and connector manually.

How can we use it?

Step 1: Define one process clearly

Start with a focused topic such as:
Permit-to-work approval
Incident reporting and escalation
Confined-space entry
LOTO verification
Emergency response
Contractor onboarding
Audit and CAPA closure

Step 2: Open Whimsical Agent

Create a board, open the Agent icon in the top-right corner and describe the required diagram. Whimsical recommends specifying the type of visual required, such as a flowchart, rather than simply asking it to “make a diagram.”

Step 3: Generate a simple first draft

Ask for the major stages, responsible roles and decision points first. Avoid placing an entire 40-page procedure into one heroic prompt and expecting wisdom to emerge fully formatted.

Step 4: Improve it through follow-up prompts

We can ask the Agent to:
Add missing approval stages
Insert decision diamonds
Show rejection or stop-work paths
Rearrange the sequence
Separate responsibilities by role
Add colour coding
Simplify overcrowded sections

Whimsical Agent can modify existing board content directly through the same chat interface.

Step 5: Verify it against approved documents

A competent HSE professional must compare every step with the approved procedure, legal requirement and site-specific control.

Step 6: Share or export the approved visual

Boards can be shared for review and exported as PNG or PDF. PNG exports on the Free plan include a “Made with Whimsical” watermark.

Practical HSE applications
1. Permit-to-work workflow

We can map:
Job request → Risk assessment → Isolation → Gas testing → Permit approval → Work execution → Suspension or revalidation → Closure

The diagram can clearly show who performs each action and what happens when a required control is incomplete.

2. Incident escalation flowchart

We can create a decision tree covering:
Immediate response → Medical assistance → Area control → Notification level → Investigation requirement → Regulatory reporting → CAPA → Verification and closure

This can help supervisors understand escalation routes, but the approved emergency and reporting procedure must remain the governing document.

3. Audit and CAPA closure process

We can visualise:
Finding raised → Risk classification → Owner assigned → Target date → Evidence submitted → HSE verification → Accepted or returned → Closed

This makes it easier to explain why uploading a photograph is not automatically the same as verifying that an action is effective. A distinction apparently requiring both diagrams and persistence.

Trial prompt:
Create an editable, left-to-right confined-space entry flowchart for an industrial facility. Include job-scope confirmation, risk assessment, permit approval, energy isolation and verification, atmospheric testing by an authorised person, ventilation, attendant confirmation, rescue readiness, entrant checks, entry authorisation, atmospheric monitoring, emergency evacuation, permit suspension or revalidation, and final closure.

Add decision diamonds for incomplete isolation, atmospheric readings outside approved limits, inadequate rescue readiness and changing conditions during entry. Do not invent exposure limits or company requirements. Mark any site-specific value as “Insert approved site requirement.”

After generation, we should ask:
Review this flowchart for missing decision points, unclear responsibilities and situations requiring work stoppage. Do not add technical requirements unless they are provided in the approved procedure.

There is one important pricing-page inconsistency. Whimsical’s live pricing FAQ currently mentions 20 AI credits per workspace, while its Agent guide states that each user receives 10 free Agent credits. Users should therefore check the actual credit balance displayed inside their workspace before starting a demonstration.

Free and paid plans
Free plan: US$0

The Free plan currently includes:
Unlimited private files
Three collaborative or team boards
Ten guest seats
Seven-day version history
A limited allowance for trying Whimsical AI

Pro plan
US$12 per editor monthly, or
US$120 per editor annually, equivalent to US$10 per month

It adds unlimited shared boards, private teams, 50 guests and 90-day version history.

Business plan
US$24 per editor monthly, or
US$240 per editor annually, equivalent to US$20 per month

It adds controls such as SAML SSO, SCIM, workspace restrictions, centralised access management and priority support.

The current Agent guide also lists additional AI credits at US$25 for 250 credits. Credit consumption depends on the complexity of the request, and Whimsical describes the Agent as a feature still under development, so limits and behaviour may change.

Data and safety precautions
Whimsical states that Claude by Anthropic is used for diagram generation and that Anthropic does not preserve submitted Agent data or use it for AI training. Whimsical’s AI terms also state that customer data is not used to train the machine-learning models powering its AI features.
Whimsical also states that it is SOC 2 Type II compliant and provides GDPR and CCPA-related compliance support.

Even so, we should:
✅ Begin with generic or approved procedures
✅ Remove employee names and personal information
✅ Avoid confidential plant layouts and security vulnerabilities
✅ Keep emergency contacts, limits and responsibilities site-specific
✅ Verify the diagram with process owners and workers
✅ Apply normal document-control approval before use

❌ Do not treat an AI-generated flowchart as an approved procedure
❌ Do not use it for live emergency decisions without formal validation
❌ Do not upload client-controlled information without permission

A thought for our HSE fraternity
> A procedure protects people only when its logic is visible, understood and followed.

> Let’s use AI to make complex safety processes easier to discuss, review and improve.
> Not to replace professional judgement, but to help everyone see where the next step leads, especially when the safest decision is to stop.

Note:
All previous AI Friday capsules are uploaded in our website https://jabharathsefoundation.org/artificial-intelligence-ai-tools-in-hse/

J&A BHF / GHF Annual Keval Shanti & Dr Pravin Patel - Mita Patel Students Scholarship Program 2026 | Last Date 31 Aug 2026 | Click Here