Explore Problems

A homeowner hired a contractor through Angi for a lighting fixture installation, but the contractor disregarded manufacturer handling instructions, stood on the customer's toilet and vanity in work boots, argued about missing hardware, and left the finished installation loose with exposed wires.

A customer describes AT&T's phone support as unusable: the automated phone tree makes it hard to reach a live person, and once connected, representatives lack the knowledge to resolve issues and repeatedly transfer callers, resulting in 90+ minutes on the phone with no resolution.

A fragrance shopper describes having to open around 15 browser tabs to compare price and check stock availability for a single bottle across retailers, since availability and price both vary significantly by site. The maker built a tool aggregating price and availability across roughly 17 fragrance retailers to solve their own version of this problem.

A customer was billed for the period after canceling Xfinity Mobile service, but could not resolve the overcharge online because cancellation deactivated the account chat access, and was instead told to visit a retail store in person. This reflects a structural gap where cancellation workflows disable the very support channels customers need to fix resulting billing errors.

Job hunting requires repetitive manual work: finding relevant postings, filling out applications, and tracking status across many companies. An autonomous AI agent that searches, applies, and tracks applications end-to-end addresses this, though several AI-assisted job-application tools already compete in this space, and automated application/captcha-bypass approaches raise platform-ToS risk.

People running several AI agents and chat tools end up with outputs spread across separate sessions, with nowhere durable for results to land. Each output type (places, schedules, todos, notes) needs a different format to be usable, so results stay trapped as chat text. The more agents someone runs, the more work it takes to find and reuse what they produced.

A customer whose appliance broke down under an Allstate protection plan was told replacement was warranted, then offered reimbursement below the current replacement price after costs rose, with escalation to a supervisor going nowhere. This points to a gap between protection-plan promises and actual claim payouts.

Users save location-tagged reels (e.g., restaurant or travel recommendations) to themselves but rarely return to them because they are buried in a flat DM or bookmark list. The problem is retrieval and actionability of saved content, not the saving itself.

A developer needs to run a Python-trained neural network on an embedded controller with no Linux, limited CPU/memory, minimal runtime dependencies, and predictable real-time behavior, but existing frameworks like ONNX Runtime mostly target desktop or embedded-Linux environments. This points to a broader gap in tooling for generating standalone, dependency-light C code from trained models for bare-metal or RTOS embedded targets.

A commenter praises a tool for working without installation, since many office laptops block downloading unapproved software, and for tying reminders to actual laptop activity rather than a fixed timer. This surfaces two real constraints: corporate device lockdown limiting what employees can install, and generic timer-based nudges feeling irrelevant compared to activity-aware ones.

Accountants using QuickBooks Online report recurring price increases and unnecessary interface or format changes that appear aimed at justifying higher fees. They feel the vendor undervalues accountant relationships. This erodes trust among the professionals who recommend the tool to clients.

An experienced engineer spent six months finding no customers because the problem was not validated up front. He struggles to find and validate small, specific problems with a cohort before building, and Reddit mining has not worked.

A Freshdesk user finds advanced features hard to master without extra training time. Costs rise sharply when moving beyond the basic tier. This affects support teams outgrowing entry-level plans.

A Freshdesk reviewer dislikes that skills-based routing and deeper integrations require plan upgrades, and notes occasional lag with heavy ticket volume. It affects growing support teams.

Slack users receive so many notifications that they need discipline to separate what needs attention. It affects anyone in busy workspaces.

The same question often needs answers from several different platforms, such as search engines, video sites, and community forums, requiring users to manually re-enter the same query in each one. This interrupts the flow of research for anyone trying to gather a complete picture from multiple types of sources.

Builders who prompt a single AI model to build an app often get results without the planning, design review, and staged approvals a real product team would apply. There is demand for an AI-driven process that mirrors a full team, planning, designing, and building in visible stages the user can approve at each handoff.

Web developers often keep custom JavaScript or CSS workarounds in production long after modern CSS gained native support for the same behavior, adding unnecessary code weight and maintenance burden. Teams lack an easy way to audit a live page and identify which custom code could be replaced with browser-native CSS, with evidence to justify the change.

A State Farm customer was pressured into a renter's insurance policy they didn't need, were assured it would be removed, but continued being charged premiums for nearly a year. When they discovered the ongoing charges, State Farm only offered a one-month refund instead of reimbursing the full period since the policy's inception.

A customer's vehicle has been at a CarMax service center for a full week after a wrong replacement part was ordered, with the dealership providing inconsistent updates and no proactive loaner vehicle. The delay is disrupting the customer's work and daily responsibilities as a single parent.