AI vs Hiring

This week brings together everything you have learned and applies it to the most important resource allocation decision founders face: when to automate, when to outsource, and when to hire.

The three ways to get tasks done

Every task in your business can be handled one of three ways:

AI automation: Software handles the work with minimal human oversight. Best for repetitive, rule-based tasks where consistency matters more than creativity.

Freelancers: External specialists handle specific projects or ongoing work. Best for specialized skills you need periodically but not constantly.

Full-time hires: Dedicated team members who become part of your organization. Best for core functions that require deep context, ongoing judgment, and cultural alignment.

The mistake most founders make is defaulting to one approach. They either try to automate everything, outsource everything, or hire for everything. The right answer depends on the specific task.

The decision matrix

When deciding how to handle a task, evaluate it across these criteria:

Predictability

How consistent is the task each time?

• High predictability (same inputs, same outputs): AI automation works best
• Medium predictability (similar but varying): Freelancers with clear SOPs work well
• Low predictability (requires judgment each time): Full-time hire needed

Frequency

How often does this task occur?

• Daily or continuous: Consider automation or full-time hire
• Weekly or monthly: Freelancer often makes sense
• Occasional or project-based: Freelancer is usually best

Strategic importance

How core is this to your competitive advantage?

• Core differentiator: Full-time hire to protect and develop
• Important but not differentiating: Could be freelancer or automation
• Commodity task: Automate if possible

Context required

How much business knowledge is needed?

• Deep context required: Full-time hire who lives in your business
• Moderate context: Freelancer with good onboarding
• Minimal context: Automation with clear rules

Error tolerance

What happens if the task is done imperfectly?

• Low tolerance (legal, financial, customer-facing): Human oversight needed
• Medium tolerance: Can use AI with review
• High tolerance (internal, experimental): Automation is fine

The 80% rule

Here is the philosophy that should guide your decisions:

Hire full-time when the person will do this task 80% or more of the time.

If someone spends less than 80% of their role on their core specialty, you are paying for capacity you are not using. A full-time graphic designer who only designs 40% of the time is an expensive choice.

How to apply this

Before hiring full-time, ask:

• Will this person do this specific work 80%+ of their time?
• Do I have enough volume to keep them busy?
• Is the work predictable enough to forecast?

If the answer is no to any of these:

• Use freelancers for the variable portion
• Automate what can be systematized
• Only hire when you hit the 80% threshold

Example calculation

You need design work. You estimate 15 hours per week of design tasks.

• Full-time designer: 40 hours available, you use 15 = 37% utilization
• Decision: Use freelancers until you consistently need 32+ hours per week

The 80% rule prevents premature hiring that creates overhead before you have the volume to justify it.

AI toolkit by business function

Different business functions have different AI readiness. Here is what works today:

Legal and compliance

• Contract review: AI can summarize terms and flag unusual clauses
• Policy drafting: AI can create first drafts of standard policies
• Compliance research: AI can research regulations and requirements
• Limitations: Always have human review for anything binding

Lead generation

• Prospect research: AI agents can gather information about potential customers
• Lead list building: Automated scraping and enrichment
• Initial outreach drafting: Personalized at scale
• Best tools: Claude for Chrome, Browse AI, Clay

Marketing

• Content creation: Blog posts, social media, newsletters
• Ad copy variations: A/B test copy generation
• SEO optimization: Content optimization and keyword research
• Image generation: Product shots, social graphics, hero images
• Best tools: ChatGPT, Claude, Midjourney, DALL-E

Finance

• Bookkeeping categorization: Transaction classification
• Report generation: Standard financial reports
• Expense analysis: Pattern identification and optimization
• Forecasting assistance: Scenario modeling
• Limitations: Human review for decisions, compliance, tax

HR and operations

• Job description writing: Role documentation
• Interview question generation: Structured interview prep
• Policy documentation: Standard operating procedures
• Onboarding materials: Training document creation
• Best tools: Claude for detailed writing, Notion AI for documentation

Documentation

• Meeting notes: Transcription and summarization
• Process documentation: SOP creation from descriptions
• Knowledge base articles: Help documentation
• Best tools: Fireflies.ai, Otter.ai, Claude

Web and digital

• Website copy: Landing pages, about pages, product descriptions
• Email sequences: Automated drip campaigns
• Chatbot responses: Customer service automation
• Best tools: Claude, ChatGPT, Jasper

Image and design

• Product photography: AI-generated product shots
• Social media graphics: Branded visual content
• Presentation visuals: Pitch deck imagery
• Best tools: Midjourney, DALL-E, Canva AI

Application development

• Prototype building: No-code app creation
• Code assistance: Development support
• Bug fixing: Code review and correction
• Best tools: Cursor, Claude, Replit Agent

Working with freelancers

When AI cannot handle a task and you do not have enough volume for full-time hire, freelancers are the answer. Here is how to work with them effectively:

Step 1: Define the scope precisely

Before posting anything, document:

• Exact deliverables with specifications
• Timeline with milestones
• Quality standards and examples
• What success looks like

Vague scopes create problems. The more specific you are upfront, the better results you get.

Step 2: Write a comprehensive brief

Your brief should include:

• Background on your business
• The specific problem or need
• Deliverables in detail
• Timeline and deadlines
• Budget range
• How you will evaluate submissions
• Examples of work you like

Step 3: Choose the right platform

• Upwork: Best for ongoing relationships and diverse skills
• Fiverr: Best for quick, defined tasks
• Toptal: Best for high-end technical talent (premium pricing)
• 99designs: Best for design competitions
• Specialized platforms: Industry-specific marketplaces

Step 4: Screen effectively

Look for:

• Portfolio work similar to what you need
• Clear communication in their proposal
• Realistic timeline estimates
• Questions that show they read your brief
• Reviews from similar projects

Red flags:

• Generic proposals that could apply to any job
• Unrealistically low prices
• No relevant portfolio examples
• Poor communication quality

Step 5: Start with a paid test project

Never commit to a large project without a smaller test. Pay for a real but limited scope that lets you evaluate:

• Work quality
• Communication style
• Deadline reliability
• Revision handling

Step 6: Set clear milestones

Break larger projects into checkpoints:

• Review work at each milestone
• Pay incrementally based on deliverables
• Course-correct before too much work is done

Step 7: Provide thorough feedback

When something is not right:

• Be specific about what needs to change
• Provide examples of what you want
• Explain why it matters
• Give clear direction, not just criticism

Step 8: Document everything

Keep records of:

• Original scope and agreement
• All communications
• Deliverables and versions
• Payments and invoices
• Feedback and revisions

Step 9: Build ongoing relationships

Good freelancers are valuable. When you find one:

• Pay fairly and promptly
• Give clear feedback
• Provide repeat work when possible
• Write honest reviews
• Maintain the relationship for future needs

Step 10: Know when to transition

If a freelancer is working 80%+ of their time for you:

• Consider converting to full-time or contract
• Discuss whether they want that arrangement
• Formalize the relationship appropriately

Red flags and quality standards

Red flags when hiring freelancers

In proposals:

• Copy-pasted responses that do not address your specific needs
• Prices far below market rate
• Promises that seem too good to be true
• Poor grammar in a writing-focused role
• No questions about your project

During work:

• Missed deadlines without communication
• Deliverables that ignore your specifications
• Defensive responses to feedback
• Scope creep or surprise charges
• Radio silence between milestones

Quality standards to set

For written content:

• Matches your voice profile
• Factually accurate (you verify)
• Properly formatted
• Free of obvious AI giveaway patterns
• Meets specified word count or length

For design:

• Matches brand guidelines
• Provided in required formats
• Includes source files
• Responsive or adaptable as needed
• Looks professional at all sizes

For development:

• Works as specified
• Code is documented
• Tested before delivery
• Follows your technical standards
• Includes deployment instructions

What NOT to automate or outsource

Some things should stay in-house regardless of efficiency:

Core strategic decisions

• Pricing strategy
• Product direction
• Market positioning
• Hiring key roles
• Partnership decisions

AI can inform these decisions but should not make them.

Customer relationships at critical moments

• Crisis communication
• High-value customer negotiations
• Complaint escalations
• Partnership discussions

Human judgment and emotional intelligence matter here.

Anything requiring your signature or liability

• Legal agreements
• Financial commitments
• Regulatory filings
• Public statements

Always have human review and explicit approval.

Creative direction and brand voice

• Overall brand strategy
• Visual identity decisions
• Core messaging
• Brand values and positioning

AI can execute but humans should direct.

Confidential or sensitive operations

• Personnel decisions
• Competitive strategy
• Financial planning
• Investor communications

Keep these close for security and judgment reasons.

Building your decision framework

Use this process for any new task:

Step 1: Classify the task

• Predictable or variable?
• Frequent or occasional?
• Core or commodity?
• High context or low context?
• Error tolerant or sensitive?

Step 2: Evaluate options

• Can AI handle this with acceptable quality?
• Do I have enough volume for full-time hire?
• Is there a freelancer who specializes in this?

Step 3: Start with the simplest solution

• Try automation first if task is predictable
• Use freelancers for variable, specialized work
• Only hire when you have 80%+ utilization

Step 4: Review and adjust

• Track time spent and quality delivered
• Measure cost per output
• Adjust approach as volume changes

Key takeaway

The goal is not to hire people or avoid hiring people. The goal is to match the right resource to each task based on its specific characteristics. Use AI for what it does well, freelancers for specialized and variable work, and full-time hires only when you have the volume and strategic importance to justify the commitment.