Problems to solve
Problems to solve, from people who have them
Looking for problems to solve? Skip the brainstorm. The ones worth your time are already written down by the people who have them, in posts and comments, with what it costs them. Here is how to tell a real one from a loud one, and where to start.
Free to read, with people's own words.
What problems to solve: three tests
Do many different people describe it?
Count people, not posts. Fifty complaints from one person are a bad week. Fifty people who never met, using their own words for the same trouble, are a market you can reach.
A viral thread is one mood, on one day.
Does it hit them at least weekly?
Something that bites every week stays in mind long enough to get fixed, and it keeps someone paying month after month. A yearly annoyance gets forgiven by the time your email arrives.
Rare and painful is hard to sell on a subscription.
Is anyone already paying?
Look for money: a workaround they pay for, staff hired to do it by hand, a tool they resent but keep. People who spend on a problem have already decided it matters to them.
“I would use that” costs nothing to say.
The problems most people describe
The widest ones. Many builders will see them too, so look for the part of the group nobody serves.
- Unreliable AI Coding Agents
AI coding agents make out-of-scope, unverified, and unmaintainable changes that force constant human review and rework.
1.5K people · 9% pay or would pay · 89% at least weekly
- Uncoordinated Coding Agents
Developers struggle to coordinate, steer, and maintain context across multiple AI coding agents, which drift, loop, and produce unreliable work.
958 people · 7% pay or would pay · 92% at least weekly
- AI Code Review Burden
AI-generated code is unreliable, bloated, and poorly understood, forcing developers into heavy manual review, debugging, and rework before it can ship.
939 people · 6% pay or would pay · 91% at least weekly
- Managing Multiple AI Agents
Teams struggle to deploy, manage, monitor, and govern multiple AI agents without constant manual effort and rebuilding infrastructure.
796 people · 6% pay or would pay · 87% at least weekly
The problems people pay to fix
The highest share of people who pay or say they would, among problems at least 150 people describe.
- Inaccurate B2B Contact Data
B2B sales and outreach teams waste budget, dials, and email credits on stale, inaccurate contact data that causes high bounce rates and low connect rates.
188 people · 44% pay or would pay · 82% at least weekly
- Shared and Fake Paid Leads
Businesses waste money on purchased leads that are shared, fake, stale, or unresponsive, with no refunds or recourse.
221 people · 39% pay or would pay · 68% at least weekly
- AI Coding Token Cost Overruns
Developers using AI coding assistants and agents burn through token quotas and budgets unpredictably, hitting limits mid-task and losing productivity.
175 people · 38% pay or would pay · 85% at least weekly
- Small Law Firm Case Management
Solo and small-firm legal practitioners lack an affordable, usable system to track cases, deadlines, documents, and billing, causing missed tasks and lost time.
191 people · 27% pay or would pay · 90% at least weekly
The problems that hit every week
The highest share of people who meet it at least weekly: habits to replace, and a reason to keep paying.
- Overwhelming Task Backlogs
Users struggle with tasks scattered across multiple tools, manual prioritization, and lack of AI-driven automation, leading to lost follow-ups and overwhelming backlogs.
192 people · 5% pay or would pay · 100% at least weekly
- Unpaid After-Hours Paperwork
Clinicians across specialties lose time to excessive documentation and insurance paperwork, causing burnout and unpaid after-hours work.
169 people · 2% pay or would pay · 100% at least weekly
- Unverified AI Test Results
Teams struggle to trust and scale AI-assisted testing because agents produce unverified, unreliable results that require manual checking.
155 people · 3% pay or would pay · 100% at least weekly
- Unpaid Session Note Writing
Therapists and clinicians spend excessive unpaid time writing session notes and documentation, causing burnout, compliance risk, and lost billable time.
152 people · 5% pay or would pay · 100% at least weekly
Problems to solve in your own field
You will spot a real problem faster where you know the work. Each field below has its own list, with the number of problems in it.
Found a problem? Turn it into something to sell
Each idea on this site starts from problems like these, then adds who buys, what they use today and a first test. Or check your own with the same questions.
The figures come from posts and comments people wrote on Reddit and other platforms.