Templates That Waste Time¶

Uncomfortable Reading

The Dirty Secret About Templates¶

Templates are training wheels. They exist because:

1. Some people genuinely don't know where to start
2. Organizations want consistency
3. Someone once forgot something important
4. A consultant had to deliver something tangible

But templates become harmful when:

• The template becomes the goal, not the thinking it was supposed to prompt
• Filling boxes replaces actual analysis
• Length is mistaken for quality
• The template doesn't fit the situation but gets used anyway

Template-by-Template Honesty¶

AI Use Case Identification Template¶

When it helps: - You've never evaluated an AI opportunity before - Multiple people need to assess the same opportunity consistently - You need to communicate the opportunity to people unfamiliar with it

When it wastes time: - You've done this dozens of times and know what matters - The opportunity is obviously good or obviously bad - You're using it to justify a decision already made - The use case is identical to one you've already documented

Skip it if: You can explain why this is or isn't a good AI use case in 2 paragraphs.

Minimum viable version: One page: Problem, Proposed AI approach, Why AI (vs. alternatives), Key risks, Go/no-go recommendation.

AI Business Case Template¶

When it helps: - Seeking significant funding (>$500k) - Need to compare options rigorously - Multiple stakeholders need convincing with detailed evidence - Finance team requires specific format

When it wastes time: - Small internal project with existing budget - Obvious ROI that everyone already agrees on - Pilot/experiment phase where full business case is premature - Going through the motions because "we need a business case"

Skip it if: Total investment is <$100k and you have verbal approval from budget holder.

Minimum viable version: 2 pages: Problem statement, Proposed solution, Costs (rough), Benefits (rough), Risks, Recommendation.

The trap: 40-page business cases for $50k projects. The document costs more than the pilot.

AI Risk Register Template¶

When it helps: - Complex project with many risk types - Multiple risk owners who need a shared view - Regulatory requirement for risk documentation - Need to track risk status over time

When it wastes time: - Simple project where risks are obvious and few - Risks are identified but no one acts on them - Register is created at start and never updated - Same generic risks copied from project to project

Skip it if: You can list the actual risks and mitigations on one page.

Minimum viable version: Simple list: Risk | Likelihood | Impact | What we're doing about it | Who owns it.

The trap: 50-item risk registers where most items are "will continue to monitor."

Stakeholder Engagement Plan Template¶

When it helps: - Large project with diverse stakeholders - Political complexity requiring careful sequencing - Need to coordinate multiple team members doing engagement - History of stakeholder issues on similar projects

When it wastes time: - Small project with 2-3 obvious stakeholders - Stakeholders are already engaged and supportive - Plan is produced but engagement doesn't happen - Used to create appearance of engagement without substance

Skip it if: You can name every stakeholder, their interest, and your approach in a 10-minute conversation.

Minimum viable version: List of stakeholders, what they need to know, how you'll tell them.

The trap: Elaborate engagement plans that sit in SharePoint while stakeholders are actually ignored.

Data Quality Assessment Template¶

When it helps: - First time working with a new dataset - Data comes from source with unknown quality - Quality issues have caused problems before - Need to make go/no-go decision about data

When it wastes time: - Well-understood dataset used many times before - Quick exploratory analysis just to understand the data - Quality assessment more thorough than the analysis that follows - Assessment documented but not acted upon

Skip it if: A data scientist can tell you if the data is usable in 2 hours of exploration.

Minimum viable version: Sample the data, check for obvious issues, make a call.

The trap: Two-week data quality assessment before a one-week pilot.

Model Card Template¶

When it helps: - Model going into production - Model will be used by people who didn't build it - Regulatory or policy requirement - Need to track multiple models over time

When it wastes time: - Experimental model that will be discarded - Model used only by the team that built it - Extensive documentation for low-stakes use case - Template completed but never referenced

Skip it if: Model is experimental, internal, and will likely be rebuilt.

Minimum viable version: What it does, what it doesn't do, known limitations, when to trust it/not trust it.

The trap: Full model cards for every experimental notebook.

Security Assessment Template¶

When it helps: - System handles sensitive data - External parties have access - Connects to critical infrastructure - Required for accreditation

When it wastes time: - Internal tool with no sensitive data - Proof of concept with synthetic data - Identical architecture to previously assessed system - Assessment done but findings not addressed

Skip it if: No sensitive data, no external access, isolated from critical systems.

Minimum viable version: Data classification, access controls, done.

The trap: Full IRAP assessment for an internal analytics dashboard with public data.

The Meta-Problem: Template Proliferation¶

You now have templates for templates. Every time something goes wrong, someone creates a new template to prevent it. But:

• New templates don't replace old ones - they stack
• Each template makes sense in isolation
• Collectively, they're impossible to complete properly
• So people fill them out badly, which defeats the purpose

The result: Template fatigue → checkbox compliance → worse outcomes than having no template

How to Tell If a Template Is Wasting Time¶

Red Flags¶

1. You're copying from the previous project's document

2. If the answer is always the same, the question isn't useful

3. You're inventing content to fill sections

4. "N/A" or made-up content means the section isn't relevant

5. No one will read what you're writing

6. If there's no audience, there's no purpose

7. You're spending longer on documentation than implementation

8. Documentation should scale with project size and risk

9. You already know the answer but have to "show your work"

10. That's bureaucracy, not governance

11. The template assumes a linear process but your project isn't linear

12. Square peg, round hole

Green Flags¶

1. Filling out the template changes your thinking

2. That's the template working

3. Someone else will genuinely use this document

4. Real audience = real purpose

5. The template catches something you would have missed

6. Value added

7. You're adapting the template to fit your situation

8. Shows you're thinking, not just filling boxes

How to Use Templates Without Wasting Time¶

Strategy 1: Start with the Minimum¶

• Open the template
• Delete everything that doesn't obviously apply
• Fill out what remains
• Ask "what's missing?" - add only what's needed

Strategy 2: Interview Format¶

Instead of filling out boxes: - Have someone interview you using the template as a question guide - Record or note the conversation - Write up only what matters - Takes 30 minutes instead of 4 hours

Strategy 3: Differentiate by Risk/Size¶

Create three tracks: - Small/Low-risk: 1-page version - Medium: Short form, maybe 5 pages max - Large/High-risk: Full template

Most projects should be in the first two categories.

Strategy 4: Living Documents, Not Snapshots¶

Instead of one big document at the start: - Start with a skeleton - Fill in as you learn - Update when things change - Never "finish" until project is done

Strategy 5: Question the Template¶

Before using any template, ask: - What decision will this inform? - Who will read this? - What's the minimum version that achieves that purpose? - Do I actually need a template, or just a conversation?

The Templates This Toolkit Should Have But Doesn't¶

Honest gap analysis:

The Hardest Advice¶

The best use of templates is knowing when not to use them.

If you have the judgment to know what matters, you don't need a template to tell you. The template exists for people who don't have that judgment yet, or for situations where consistency across many people matters.

If neither of those applies, put down the template and do the work.

Permission Granted¶

You hereby have permission to:

• Delete sections of templates that don't apply
• Use 1-page versions instead of full templates
• Skip templates entirely when you can justify it
• Adapt templates to fit your situation
• Decline templates that are clearly bureaucratic waste
• Push back on requests for unnecessary documentation

The goal is good decisions and good outcomes, not filled-out templates.

Never forget which one matters.

"A template should be a starting point for thinking, not a substitute for it."