Last week I did something stupid.
I pitted four AI models against each other. Cage-match style. To see which one could actually help someone land a job in 30 days.
Why? Because Iβm always talking about using the right AI tool for your job search, but I wanted to know: soup to nuts, which tool would reign supreme?
So I tested:
-
ChatGPT5
-
Claude Sonnet 4.5
-
Manus Agent
-
Gemini 2.5 Flash
And scored them out of 10 (where 10 = this will absolutely land you a job and 0 = total garbage).
The results? Two clear winners. One total face-plant. And some surprising tactics that actually work.
Here’s what you’ll learn today:
-
Which AI model gives the most actionable 30-day job search plan
-
The critical mistake 3 out of 4 models made (and why it kills your chances)
-
Specific tactics to steal from each model (including a copy-paste cold email template)
β¦and more!
π₯ The Setup: One Prompt, Four AI Gladiators
I chose Product Manager as the test role because it’s one of the most competitive remote positions with limited resources. If AI could crack this, it would work for other roles too.
I gave each AI model the exact same prompt:
Act as a career strategist and product coach. I want to land a Product Manager role within 30 days. Here are my constraints and context:
I have 2 years of Product Manager experience.My professional network is not very strong, so I cannot rely much on warm introductions.I can dedicate 4 hours per day to the job search.
Please design a step-by-step 30-day plan that maximizes my chances of landing interviews and offers. The plan should include: 1/ Weekly goals and milestones (e.g. positioning, outreach, interviews).
2/ A breakdown of daily tasks within my 4-hour limit.
3/ Concrete deliverables for each week (resume updates, outreach targets, interview practice, etc.).
4/ Tactics to compensate for my weak network (e.g. cold outreach, creating proof of work, leveraging communities).
5/ Pros and cons of each major tactic so I can prioritize effectively.
Some models asked follow-up questions. I answered them. Then I let each one cook π¨βπ³
Letβs check out the results.
π The Winners: ChatGPT Crushes It, Claude Close Behind
πΒ ChatGPT: 8/10
ChatGPT didn’t mess around. It gave a practical, day-by-day breakdown with clear deliverables, e.g.:
Week 1 deliverables: Updated resume and LinkedIn profile; target company list; membership in PM communities; and a proof-of-work project plan.
Week 2 deliverables: 15-20 tailored applications; 10-15 new LinkedIn connections; and a published proof-of-work piece showing your skills.
The best part?
A detailed “Networking Without a Strong Network” section with specific tactics and honest pros/cons for each approach. This included:
-
Using PM communities like Mind the Product and ProductCoalition to build relationships before asking for anything
-
Commenting on thought leader posts with genuine insights (not generic “great post!” responses)
-
Creating content that demonstrates your PM thinking, which naturally attracts connections
The fatal flaw: ChatGPT skipped market research before jumping into resume updates. You can’t tailor your materials if you don’t know what companies actually want.
π Claude: 7.5/10
Claude nailed the fundamentals. It started with market research FIRST, then moved to positioning.
It asked smart follow-up questions nobody else did:
-
What company stage/type do you want?
-
Geographic flexibility?
-
Product domain focus?
Claude also gave brutal 2024-2025 market context and compensation benchmarks. No sugar-coating the state of the market. I thought this was a brilliant addition to the plan.
The contingency plans were gold, too. If you’re not getting responses after Week 2, Claude offered three backup approaches:
-
A/B test different LinkedIn message templates with specific variables to test
-
Shift from PM roles to adjacent roles (growth manager, customer success) with clear positioning
-
Add technical case studies to your portfolio with templates for what to include
π€ The Disappointments: Manus Stumbles, Gemini Face-Plants
πΒ Manus: 6/10
Manus jumped straight to “positioning and branding” without market research. It told you to spend 80% of Week 1 on positioning but never mentioned understanding market demand first.
It also gave an hour-by-hour breakdown, which was way too detailed. Nobody needs that level of micromanagement.
One bright spot: Manus provided this solid cold email template.
I read it and immediately thought: βOh, this would work.β
(I share the email below.)
ποΈΒ Gemini: 3/10
Gemini’s output was practically unreadable. Instead of clear advice, it gave me section headers like:
-
“The 30-Day Product Management Velocity Strategy: Maximizing Offer Conversion Under Constraint”
-
“Framework Mastery for Cognitive Efficiency (Days 11-13)”
Translation: practice talking about your work. Why not just say that?
Gemini obsessed over proof-of-work projects while ignoring market research and networking. It put negotiation tactics before you even had interviews. Cart, meet horse.
The one thing Gemini got right: focus on one specific role instead of spraying applications everywhere.
π― The Best Tactics to Steal from Each Model
Each AI had at least one killer recommendation worth using. Here’s what to take from each:
ChatGPT’s “Networking Without a Strong Network” section was the most practical guide I’ve seen.
πΒ Here’s the framework:
Week 1-2: Build visibility
Join 3-5 PM communities (ProductCoalition, Mind the Product, Product School)
Comment thoughtfully on 5 posts per day from industry leaders
Share one insight or observation weekly that shows your thinking
Week 2-3: Create value first
Offer to review someone’s product roadmap or do a mini teardown
Write a post solving a common PM problem you’ve faced
Participate in community discussions without asking for anything
Week 3-4: Make specific asks
Reach out to people you’ve engaged with consistently
Reference specific conversations or shared interests
Ask for 15-minute informational chats, not job referrals
π Do this: Pick one PM community this week. Spend 20 minutes daily engaging genuinely, not just lurking.
Claude’s backup plans were brilliant because they gave you clear decision points.
πΒ Contingency plans
After Week 1: Check your materials
If zero profile views: your LinkedIn headline is invisible to recruiters
If views but no messages: your experience section doesn’t match what hiring managers want
If messages but wrong roles: your positioning is unclear
After Week 2: Check your outreach
If zero responses to cold emails: A/B test subject lines and opening hooks
If responses but no meetings: your ask is too vague or too big
If meetings but no interviews: you’re not demonstrating enough value in conversations
After Week 3: Pivot strategy
Consider adjacent roles that value your PM skills
Double down on proof-of-work instead of applications
Focus networking efforts on smaller companies with faster hiring
From Manus: The Cold Email That Actually Works
πΒ Use this cold email
Hi [Name], I’ve been following [Company Name]’s work in [specific area, e.g., AI-driven analytics] and was particularly impressed by [mention a specific product/feature or recent achievement]. As a Product Manager with 2 years of experience in [your area of expertise, e.g., B2B SaaS platforms], I’m keen to learn more about how your team approaches [specific PM challenge or domain]. I’d love to connect and potentially hear your insights on the PM landscape at [Company Name] when you have a moment. Best, [Your Name]
Why does this work?
-
Show you’ve done research (specific product/feature)
-
Establish credibility (your relevant experience)
-
Express genuine curiosity (what you want to learn)
-
Make a small ask (insights, not a job)
Send this exact template to 5 people this week. Customize the bracketed sections. Track your response rate.
Buried in Gemini’s jargon was one smart insight: commit to one specific role identity for at least 10 working days.
Instead of applying to “Product Manager,” “Growth Manager,” “Product Marketing Manager,” and “Strategy Manager” all at once, pick ONE. Tailor everything to that identity:
-
Your LinkedIn headline
-
Your resume positioning
-
Your portfolio projects
-
Your outreach messages
-
Your community participation
After 10 days, if you’re getting zero traction, pivot to a different specific role. But give each position enough time to generate signal.
π« The Bottom Line
ChatGPT and Claude are your best bets for creating a job search plan. This week, ChatGPT wins on tactical execution, Claude wins on strategic thinking.
But the way you keep winning?
Keep testing different models and tools to give you that edge on your job search.
You got this.