Learning Plan 1: Eliminate Bottlenecks in 3 Steps

How to Unblock Yourself (Without Becoming an Expert)

The Shift

Waiting for other teams is now optional

The Opportunity

Do baseline work in adjacent disciplines yourself

Time Investment

2-4 hours per new capability

Who This Is For

Anyone with bottlenecks

Welcome to Your Learning Plan

You just discovered you're spending days waiting for tasks that take minutes.

This happens to everyone:

Designers wait 3 days for data queries
PMs wait a week for prototypes
Researchers wait for dashboard access
Engineers wait for design specs

Every time you wait, you're trading days for a task that AI can help you do in minutes.

This learning plan will show you exactly how to eliminate your biggest bottlenecks. You don't need to become an expert in everything. You need just enough capability to unblock yourself.

By the end of this plan, you'll:

Know your top 3 bottlenecks (and their time cost)
Have eliminated at least 1 bottleneck using AI
Created a template so your whole team can self-serve
Saved 5-10 hours per week by eliminating dependencies

Let's get started.

Why This Matters

Every time you wait for another team, you're making a trade:

3 days of waiting for a 10-minute task

A designer waiting for analytics to run a SQL query. A PM waiting for engineering to build a prototype. A researcher waiting for data access to pull metrics.

Before AI, this made sense. If you weren't a specialist, you literally couldn't do the work. The technical barrier was real.

Now? AI eliminated the technical barrier.

You can write SQL with Claude's help. You can prototype with Cursor. You can build dashboards with AI assistance. Not at expert level. At good enough to unblock yourself level.

The barrier isn't technical anymore. It's psychological:

"I don't know how to do that"
"That's not my job"
"I'm afraid I'll do it wrong"

Shipping something imperfect immediately shows more professionalism than waiting 3 days for a 5-minute task.

The professionals thriving right now aren't the ones protecting their specialty. They're the ones eliminating their own blockers.

The Three-Step Method

Step 1: Acknowledge - Map Your Dependencies

Identify the 3 tasks you wait for most often.

You can't eliminate bottlenecks if you don't know what they are. Most people underestimate how much time they spend waiting because it's distributed across small delays. This step makes the cost visible.

1. Review your last 2 weeks of work
Look at your calendar, Slack messages, emails, and project timelines.
2. Find every instance of "waiting for..."
Specifically look for:
- "Blocked on [team]"
- "Asked [person] for this on Monday, still waiting"
- "Can't proceed until [dependency] finishes"
- Calendar invites titled "sync on [thing you're waiting for]"
3. Rank by frequency and delay time
Make a simple list:
What you're waiting for How often Average delay Estimated hours lost per month
Data queries 2-3x/week 2-3 days ~30 hours
Prototypes 1x/week 5 days ~20 hours
Design mockups 1x/month 1 week ~8 hours
4. Pick the top 3 bottlenecks
Focus on high-frequency, high-delay tasks. These are your targets.

What you're waiting for	How often	Average delay	Estimated hours lost per month
Data queries	2-3x/week	2-3 days	~30 hours
Prototypes	1x/week	5 days	~20 hours
Design mockups	1x/month	1 week	~8 hours

Example: Sarah, a product designer, mapped her bottlenecks:

Data queries (3x/week, 3-day delay) = 36 hours/month lost
Engineering feasibility checks (1x/week, 2-day delay) = 8 hours/month lost
Copy/content (2x/month, 1-week delay) = 16 hours/month lost

Total time lost to waiting: 60 hours per month (1.5 weeks of work)

You'll have a clear list of 3 specific blockers with estimated time costs.

Step 2: Become - Learn Just Enough, Just in Time

Use AI to do baseline work in that adjacent discipline.

You don't need to become an expert. You need to become good enough to unblock yourself on simple, repetitive tasks. AI is your coach.

1. Pick ONE blocker from your list
Don't try to solve all 3 at once. Start with the one costing you the most time.
2. Use Claude/ChatGPT to learn the basics (30 minutes)
Ask AI to explain the fundamentals:
Example prompts:
"I'm a designer who needs to run basic SQL queries. Explain the fundamentals I need to know to query a database for user behavior data."
"I'm a PM who wants to prototype web interfaces. What's the simplest way to build a clickable prototype with AI help?"
"I'm a researcher who needs to create data visualizations. What tools can I use with AI assistance to build dashboards?"
3. Attempt the actual task with AI assistance (1-2 hours)
Pick a real, low-stakes example from your work. Use AI to guide you through it.
Example: Designer learning SQL
- Real task: "I need to see how many users clicked the signup button last week"
- Claude prompt: "I need to see how many users clicked X button last week. Here's our database schema: [paste schema]. Write me the SQL query."
- Claude writes the query. You run it. You get the answer in 10 minutes.
4. Aim for "good enough to unblock," not "expert-level"
Your goal: 80% quality in 20% the time
- You're not replacing the specialist
- You're unblocking yourself on basics so they can focus on complex work
- The specialist can review and refine if needed

Example: Step-by-step walkthrough

Sarah (Designer) eliminates data query bottleneck:

Day 1: Learn basics (30 min)

Prompt to Claude: "Explain SQL SELECT, WHERE, and COUNT for someone who's never written a query"
Claude explains concepts with examples
Sarah understands: queries filter data and return results

Day 1: Try real task (1 hour)

Sarah needs: "How many users viewed the checkout page last week?"
Finds database schema (asks engineering for read access, gets it in 10 min)
Prompt to Claude: "I need to count how many users viewed /checkout between May 1-7. Here's the schema: [paste]. Write the query."

Claude writes:

SELECT COUNT(DISTINCT user_id)
FROM page_views
WHERE page_url = '/checkout'
AND viewed_at BETWEEN '2025-05-01' AND '2025-05-07'

Sarah runs it in the database tool. Gets answer: 3,247 users
Verifies with analytics team (matches their dashboard). Confidence boost.

Day 2: Try another query (30 min)

Needs: "How many of those users clicked 'Complete Purchase'?"
Prompt: "Modify this query to also show how many clicked the Complete Purchase button"
Claude modifies query
Sarah runs it. Gets answer.
She's now unblocked on basic data questions

You completed a task you would have delegated, at 80% quality, in 20% the time.

Step 3: Support - Document the Pattern

Create a template so you (and others) can repeat this.

Once you've eliminated a bottleneck for yourself, you can eliminate it for your entire team. Democratizing the solution multiplies your impact.

1. Save your AI prompts
Create a doc titled "Quick [Task Type] for [Your Role]"
Example: "Quick SQL Queries for Designers"
- Store the prompts that worked
- Note any gotchas or tips

2. Document the 3-5 steps

Write down exactly what you did, so someone else can follow:

# Quick Data Queries for Designers

**When to use this:** You need basic user behavior data and don't want to wait for analytics

**What you need:**
- Read access to database (ask #engineering, takes 5 min)
- Database schema (link here: [internal doc])
- Claude or ChatGPT

**Steps:**
1. Identify your question ("How many users did X?")
2. Ask Claude: "I need to [question]. Here's our schema: [paste]. Write the SQL query."
3. Copy query into [database tool]
4. Run and export results

**Common queries:**
- Count page views: [sample query]
- Count button clicks: [sample query]
- Filter by date range: [sample query]
- Group by user segment: [sample query]

3. Share with team (Notion page, Slack message, team wiki)
Post it where teammates will find it. Ideally, tag people who have the same bottleneck.
Example Slack message:
"Hey team! I just learned to run basic SQL queries with Claude's help, and it's been a game-changer. I'm no longer blocked waiting for analytics for simple questions.
I documented the process here: [link]
If you find yourself waiting for data queries, this might help unblock you too."
4. Now anyone can self-serve
The first time someone uses your template, offer to help. After that, they can do it themselves.

Example: What Sarah created

Sarah created "Quick SQL Queries for Designers" template:

5 common query types (page views, button clicks, user segments, conversion funnels, time series)
Prompt templates for each
Where to run queries (internal tool + credentials)
Troubleshooting tips

Result:

Shared with 12-person design team
5 people started using it in first week
Team dependency on analytics for basic queries dropped 60%
Analytics team appreciated focusing on complex analysis over simple queries
Sarah became known as "person who unblocks teams"

The blocker disappears for your entire team, not just you.

Common Blockers & How to Overcome Them

You will hit resistance. Here's how to handle it.

"I don't have access to the tools"

Solution: Ask once for read-only access. Frame it as "exploring to better understand our data/systems" not "doing someone else's job."

Most teams grant read access easily because it doesn't risk breaking anything. If you get resistance, explain:

"I'm trying to unblock myself on simple [queries/prototypes/etc.] so I can move faster. I'm not trying to replace [specialists], just handle basics myself. Can I get read-only access to try?"

If they say no, ask: "What's the concern?" Address it directly.

"What if I do it wrong?"

Solution: Start with low-stakes tasks. Verify output with the expert the first time. Build confidence.

Your first few attempts should be:

Low-stakes: Not mission-critical, not customer-facing
Verifiable: Easy to check if you got the right answer
Reversible: Can be undone if wrong

Example progression:

Run a simple query, verify result matches existing dashboard
Run a new query for your own curiosity, don't share yet
Run a query for a decision, verify with specialist before using
Run queries independently, consult specialist only for edge cases

"I'm afraid I'll look incompetent"

Solution: Shipping quickly shows more competence than waiting 3 days for a 5-minute task.

Consider what your manager sees:

Scenario A (waiting):

Monday: "I need data, asked analytics"
Wednesday: "Still waiting on analytics"
Friday: "Got the data, moving forward"
Time to decision: 5 days

Scenario B (self-service):

Monday: "Learned basic SQL, ran query, got answer, made decision"
Time to decision: 1 day

Which person looks more capable?

Ship imperfect > wait for perfect

"AI gives me wrong answers sometimes"

Solution: That's why you start with simple, verifiable tasks. Check the output. Learn to spot errors.

AI isn't perfect, but neither are humans. The key is verification:

Start with tasks where you can check the answer (e.g., query should match existing dashboard)
Learn to sense-check results ("Does 500,000 users sound right?")
Consult specialists when something looks off
Build judgment over time

You're not replacing expert judgment. You're eliminating waiting for simple execution.

"This feels like I'm doing someone else's job"

Solution: You're not replacing them. You're unblocking yourself on basics so they can focus on complex work.

Reframe it:

Old mental model:

"Data queries are the analytics team's job. If I do it, I'm stepping on their toes."

New mental model:

"The analytics team's job is strategic analysis. If I can self-serve simple queries, they can focus on the complex stuff only they can do."

Most specialists appreciate this. They want to work on interesting problems, not repetitive requests.

Example conversation with the specialist:

"Hey, I'm learning to run basic SQL queries so I can unblock myself on simple questions. This means I'll only come to you for the complex analysis that requires real expertise. Cool?"

99% of the time, they'll say: "Yes! Please stop asking me how many button clicks we had last week."

Tools & Resources

AI Tools

Claude (claude.ai) - Best for code generation, data queries, and learning new concepts with detailed explanations
ChatGPT (chatgpt.com) - Good for quick learning and iterative problem-solving
Cursor (cursor.com) - For designers/PMs learning to code prototypes (AI-powered code editor)

Learning Resources

For SQL/Data Queries:

Mode Analytics SQL Tutorial (free, practical)
SQLBolt (interactive lessons)
Your company's database schema documentation (ask your data team)

For Basic Coding/Prototyping:

"Automate the Boring Stuff with Python" (first 3 chapters, free online)
Cursor documentation (cursor.com/docs)
Replit tutorials (for quick web prototypes)

For Design (if you're non-designer):

Figma's official tutorials
Refactoring UI (paid, but excellent)
Your company's design system documentation

Community & Support

Generalista Newsletter - Weekly insights on capability expansion
Community - Coming soon: Connect with other professionals expanding capabilities

Ask for Help

When you get stuck:

Ask Claude/ChatGPT to debug (paste error message, ask for explanation)
Ask the specialist on your team (they usually appreciate curiosity)
Search for the error (someone else hit this before)
Join the Generalista community (coming soon)

What's Next

Congratulations! You've learned to eliminate individual bottlenecks.

But something bigger is now possible: collaborative discovery.

When your whole team can eliminate blockers (designers coding, PMs querying data, researchers building dashboards), you can work in parallel.

That's when discovery accelerates. That's when cycle times compress from months to weeks.

Learning Plan 2: Collaborative Discovery

Shows you how to work in parallel, not sequential handoffs. It unlocks 40-60% faster discovery cycles.

Ready for the next level?

You'll automatically receive more learning plans over time
We'll email you when new plans are available
In the meantime, focus on eliminating your bottlenecks with this plan

Your Action Plan This Week

Don't just read this. Do it.

Today:

Map your top 3 bottlenecks (30 minutes)
Pick 1 to eliminate first

This week:

Spend 30 min learning basics with AI
Attempt 1 real task with AI assistance
Verify the output

Next week:

Complete the task independently
Document your process
Share template with team

Set a reminder now to complete these steps. Put 30 minutes on your calendar this week.

The difference between people who expand capabilities and people who stay stuck isn't talent. It's action.

You just learned how to eliminate bottlenecks. Now go do it.

Share Your Progress

We'd love to hear how this worked for you!

Tag John Garvie on LinkedIn: What bottleneck did you eliminate?
Email John: john@generalista.ai

The best case studies get featured in future content (with your permission).

About this Learning Plan

This is part of the Generalista Capability Expansion Framework: a structured system for building cross-discipline skills without burning out or losing your specialty.

Created by John Garvie, Head of Design at Evisort AI (Workday). Former design leader at Uber, Amplitude, and LinkedIn. First UX Researcher to transition into design leadership at Uber.

Generalista.ai