Explore Problems

A homeowner hired a contractor endorsed by Angi to pave and paint their driveway, paid the full amount upfront as required, and ended up with a ruined driveway due to wrong paint and poor work. The platform's endorsement provided no protection against the contractor's failure.

AT&T shipped equipment to an address the customer never lived at and never accepted delivery of, yet the carrier's system logged the equipment as delivered and accepted, resulting in a $400 non-return fee. Despite 15 months and many hours across multiple representatives and supervisors, AT&T has refused to correct the error.

After Allstate offered a competitive home insurance rate, its self-inspection app flagged the customer's roof (under 4 years old) as failing, leading Allstate to decline coverage and cancel the policy after already collecting payment through the mortgage escrow. The original insurer, unaware of the switch, kept covering the customer in parallel, leaving them fighting to recover a year of duplicate payments.

A carrier advertises a $400 device credit as a signup incentive, but customers report there is no functional way to actually redeem it, requiring hours of unresolved phone and online support.

Developers writing UI test automation must hunt for selectors and maintain brittle scripts that break as the app's interface changes, and existing AI-assisted testing tools charge per test run rather than per test created. This creates both a maintenance burden and an unpredictable, usage-scaling cost problem for teams adopting AI-driven test automation.

A marketer paid $149 for 1,000 B2B leads and generated zero customers, then discovered the failure wasn't due to messaging quality but the lead source itself. This points to a recurring problem where purchased contact lists are low-fit or low-intent, wasting acquisition budget before any copy or outreach quality comes into play.

A Shopify app user reports that an app advertised as a free 3-day trial began charging $1/month without clear upfront disclosure, prompting an immediate uninstall. The complaint reflects a recurring trust problem in app-store billing transparency around trial-to-paid conversions.

A developer scraping Shopee product data repeatedly hits a verify-traffic anti-bot error that redirects requests to a login wall, even when using ISP or premium residential IPs with Playwright. This reflects the broader difficulty of reliably bypassing e-commerce anti-scraping defenses, which persists despite paying for higher-quality proxy infrastructure.

A buyer newly purchased certified vehicle developed a check-engine fault within a week, caused by the dealer own service overfill, followed by a second failure with nine warning codes just after 30 days. Repeated corporate escalation attempts went unreturned, with the earliest inspection slot offered 17 days out.

A small company using a customer service platform notes that the product design assumes large teams of agents working around the clock, an assumption that does not hold for small businesses without dedicated or continuous support staff. This mismatch between platform design assumptions and small-team realities limits usability for smaller customer service operations.

A customer who switched auto insurance policies for a car purchase that fell through within the same day saw their rate jump $800 after reinstating the old policy, with no advance warning that mid-cycle changes would trigger a full rate recalculation. This reflects a structural gap in how insurers communicate the pricing consequences of policy changes before customers make them.

Companies knowledge about their customers lives fragmented across many disconnected documents and systems, so teams cannot easily answer relational questions like what is driving churn or why customers adopt or abandon a feature. A launch post (136 upvotes) addressing this drew strong engagement, though customer-intelligence tooling is an already competitive space.

Teams maintaining internal documentation or wikis struggle to keep them current, often relying on hacky manual processes to reflect changes in underlying files and systems. A self-updating wiki tool addresses this by auto-generating and refreshing documentation from uploaded sources, with agent-native access via CLI, SDK, and MCP.

A business needs a list of 200 independent healthcare practices, such as doctors offices, dental, and vision clinics with 7 to 25 employees, within a 30-mile radius, including decision-maker names and contact details, and is hiring a freelancer over 1 to 3 months to manually build the list.

A long-time Xfinity/Comcast customer who wanted to leave for available fiber internet found the self-service web cancellation option removed, and was instead required to schedule a callback with a retention-focused sales representative just to cancel service.

A borrower says they explicitly protested a vehicle repossession in person, which under state self-help repossession law should have stopped the action, but the lender proceeded anyway and allegedly used intimidation tactics. The company has since refused to respond to a written dispute.

Customers who relocate internationally and lose their US phone number are locked out of online banking because secondary verification is hard-bound to that number, with no alternate recovery path.

When someone else reports a received transaction as fraudulent, banks can restrict the recipient account access even though the transaction was authorized. Affected customers have no clear, fast way to prove legitimacy and restore access.

Sales teams using HubSpot on mobile face a notably degraded experience compared to desktop, with tasks feeling slower and less intuitive. This is an acknowledged industry-wide gap in CRM tools where desktop-first design still dominates. Field sales and remote teams are disproportionately affected.

The Freshdesk-Jira integration is unreliable, leaving support teams unable to track escalated issues across systems. Analytics fail to capture service gaps when tickets move through multi-team escalation workflows. This blind spot makes it impossible to measure true resolution time or identify bottlenecks.