DEV Community

Matt Yonan
Matt Yonan

Posted on • Originally published at ideascout.polsia.app

I Tried 5 AI Business Ideas in 30 Days — Here's What Actually Worked

Real experiment testing 5 AI side hustles in 30 days. Revenue, time investment, what worked, what failed, and raw data from a founder.

In January, I decided to treat business ideas like a scientist treats hypotheses: test them quickly, measure results honestly, and kill the ones that don't work.

I picked 5 AI business ideas that showed up repeatedly in my research. I gave each one roughly 6 days of focused effort. I tracked time spent, outreach attempts, responses, and revenue generated (or not). Here's exactly what happened.

The Setup

  • Time available: ~10 hours/week (working full-time)
  • Budget: $200/month for AI tools (Claude Pro, ChatGPT Plus)
  • Background: Marketing professional, 6 years in B2B SaaS, decent writer, no coding skills
  • Goal: Find 1 idea worth pursuing seriously

Idea 1: AI Content Repurposing Service (Days 1–6)

The idea: Help B2B companies turn podcast episodes and webinars into social content — LinkedIn posts, Twitter threads, email blurbs.

What I did:

  • Identified 20 B2B podcasts with engaged audiences but low social media presence
  • Repurposed the latest episode of each into a 5-piece content package (uninvited)
  • Sent personalized emails: "I made something for you — no pitch, no strings. Here it is."

Results:

  • Outreach sent: 20 emails
  • Responses: 7 (35% response rate — unusually high)
  • Paid interest: 3 asked about pricing
  • Closed: 1 client at $600/month (trial, starting week 4)
  • Time spent: 22 hours

What worked: Leading with the work eliminated the credibility gap. Showing them something real with their own content was impossible to ignore.

What didn't: Production time per prospect was high — about 1.5 hours per package. Needs systematizing at scale.

Verdict: Strong signal. Pursuing.

Idea 2: AI-Powered SEO Blog Writing (Days 7–12)

The idea: Write SEO-optimized blog posts for B2B companies using AI to accelerate research and drafting.

Results:

  • Outreach attempts: 15 LinkedIn + 3 Slack posts
  • Responses: 4
  • Paid interest: 1
  • Closed: 0
  • Time spent: 15 hours

What I learned: The AI blog writing space is saturated. Every marketing manager I spoke to either had a solution or had been burned by quality issues. The trust gap is hard to overcome with cold outreach without a portfolio of measurable SEO wins.

Verdict: Poor signal. Killing it.

Idea 3: AI Resume and LinkedIn Optimizer (Days 13–18)

The idea: Charge job seekers $150–300 for an AI-enhanced resume and LinkedIn profile overhaul.

Results:

  • Free analysis takers: 6
  • Paid conversions: 1 at $150
  • Total revenue: $150
  • Time spent: 18 hours
  • Effective hourly rate: $8.33/hour

What I learned: Job seekers have real demand but limited budget and high price sensitivity. At $150/client and 3–4 hours of work per client, this would never scale without moving dramatically upstream (executive coaching, outplacement firms as B2B clients).

Verdict: Real demand, wrong pricing structure. Deprioritizing.

Idea 4: AI Social Media for Local Businesses (Days 19–24)

The idea: Manage social media for local service businesses — salons, gyms, restaurants. AI generates content; I schedule and report. $299–499/month per client.

Results:

  • Free trials started: 3
  • Paid conversions: 0 (trial still in progress at day 30)
  • Time spent: 12 hours
  • Engagement increase (avg): 47% in first week

What I learned: Local businesses are genuinely interested but slow to commit to paid. The 30-day free trial converts — but after the trial ends, not during. Time-per-account: about 2 hours/week per client with AI doing the content generation.

Verdict: Promising, too early to confirm. Monitoring.

Idea 5: AI Customer Intelligence Reports (Days 25–30)

The idea: Synthesize customer reviews, support tickets, and social media comments into actionable intelligence reports for product teams. Charge $500–2,500 per report.

Results:

  • Outreach sent: 8
  • Responses: 5 (62% — highest of any test)
  • Calls scheduled: 3
  • Paid: 0 (calls pending in week 5)
  • Time spent: 16 hours

What I learned: The product manager response rate was extraordinary. Instead of asking "how much does this cost?" they asked "how do we get this on a recurring basis?" That's the right conversation.

Verdict: Highest potential of the five. Pursuing aggressively.

Month 1 Summary

Idea Revenue Hours Signal Decision
Content Repurposing $600/mo signed 22 Strong Pursue
SEO Blog Writing $0 15 Weak Kill
Resume Optimizer $150 18 Moderate Deprioritize
Local Social Media $0 (trials active) 12 Promising Monitor
Customer Intelligence $0 (calls scheduled) 16 Very Strong Pursue aggressively

Total revenue in 30 days: $750

Total hours invested: 83 hours

Effective hourly rate: $9.04 (far below target — but recurring contracts change this math fast)

What I'd Do Differently

The biggest lesson: the ideas that worked best were the ones where I led with demonstration rather than description. Sending a finished content package or sample customer report made the pitch disappear. The work was the pitch.

The ideas that failed were ones where I described a service and asked people to believe I could deliver it.

If I were doing this again, I'd run 3 ideas instead of 5, go deeper on each, and focus entirely on creating sample deliverables before sending a single outreach message.


Originally published on IdeaScout — a tool that surfaces researched AI business ideas matched to your background so you spend less time evaluating and more time testing.

Top comments (0)