Explore Problems

GEICO enrolls customers in telematics-based discount programs then retroactively bills them if they remove the tracking app, even for valid medical reasons that prevent app use. Customers receive bills for discounts already paid off, creating surprise debt. This program structure penalizes customers without accommodating legitimate exceptions.

Banks allow consumers to withdraw funds from deposited checks before they clear, then hold consumers fully liable when checks prove fraudulent. This practice is particularly damaging in peer-to-peer selling contexts where fraudulent payment methods are common. The bank policy of enabling early access while shifting all fraud risk to consumers creates a predictable harm pattern.

Engineering teams shipping AI in production report a failure category where standard metrics stay green while the system loses coherence or drifts in non-reproducible ways. The root cause is structural: verification built on the same model that generates creates blind spots that existing observability tooling cannot detect.

A disabled senior homeowner reports that State Farm's estimate for water-damage repairs, after applying the deductible toward water removal, leaves a $3,000-6,000 shortfall against an independent contractor's repair estimate. This reflects a broader, structural pattern in property insurance claims where insurer damage estimates are systematically lower than real-world contractor quotes, disproportionately burdening policyholders least equipped to dispute them.

A truck renter was billed for damage after return with no walkthrough inspection performed at pickup and no documented condition report issued, leaving no baseline to dispute against. Requests for the photo or video evidence the operator claimed to hold went unanswered, and only a bare damage receipt was provided. Without a shared, timestamped condition record at handover, the renter carries the burden of disproving a claim they cannot see the evidence for.

Teams using monday.com's AI Work Platform have no way to set upfront AI credit limits per user, so a single team member who builds boards inefficiently or overuses AI fields can consume the shared credit pool for the whole workspace. Admins are left with no granular control to cap or throttle individual AI usage before costs accumulate.

Teams building internal tools with AI app builders like Lovable report escalating credit costs to fix bugs the AI itself introduces, plus a deeper architectural risk: role-based permissions enforced only in the UI rather than the database, letting users see data outside their role, such as drivers viewing other drivers' deliveries. This combination of recurring cost and fragile access control pushes some builders to seek alternatives with real backend ownership and row-level security.

Educators report that building class schedules manually is extremely time-consuming and routinely produces suboptimal results due to the combinatorial complexity of constraints. Existing tools are either too rigid or too manual for most school contexts. There is clear demand for software that can efficiently generate and adjust schedules while respecting teacher, room, and student constraints.

Open-source MCP servers commonly contain critical security flaws like unrestricted file access and insufficient SQL guards. Manual code review is infeasible at scale as the MCP ecosystem rapidly grows. Automated scanning tools are needed before these servers reach production AI agents.

A builder describes repeating a tedious cycle dozens of times a day: copying text from whatever app they're using, opening ChatGPT, pasting, writing a prompt, copying the answer, and returning, just to summarize or rewrite something. They built an overlay assistant that works directly on top of any app to eliminate that context-switching, on both mobile and desktop.

A founder paid $1,750/month to an organic marketing agency for six months before discovering the actual deliverables were a single AI-written article every few days, purchased low-quality backlinks, and roughly 15 minutes of human effort per week. This points to a broader trust and transparency gap for small businesses paying agency retainers without visibility into what work is actually being done.

A homeowner's wind-damaged garage was deemed a total loss by their insurer's adjuster, but the payout was roughly half of what local contractors quoted to rebuild it, with the insurer disputing coverage for foundation repair despite visible damage from the same incident. This illustrates a recurring gap between insurer damage assessments and real market rebuild costs, leaving policyholders to either accept shortfalls or fight a drawn-out appeals process.

Roughly 1.4 billion unbanked people, including tens of millions in Nigeria, live in areas with unreliable internet connectivity that standard online payment rails cannot serve. This creates demand for payment systems that can securely process and sync transactions during extended offline periods.

After accepting a rep's offer to switch to a lower-cost plan, a customer's entire account -- multiple phone lines -- was left able to make emergency calls only, due to an internal system code error the carrier could not bypass. After nine days of repeated troubleshooting and unfulfilled callback promises, the lines remained non-functional, exposing a systemic risk in how carriers execute plan migrations.

Site owners discover that ISPs like Xfinity use third-party domain-reputation blacklists (e.g. SURBL) to silently block newly launched websites at the gateway level, with no notification to the site owner. Multiple independent site owners confirm hitting the same block across unrelated domains, sometimes even for domains over a year old, making new product launches invisible to a subset of users without any warning.

A Shopify store owner was charged $41 for an app they never used and did not intend to keep, and struggled to find a way to stop the unwanted charge and cancel the app. This highlights confusion and friction in Shopify app billing and cancellation flow, leaving merchants exposed to unexpected recurring charges.

A customer replacing a faulty handset was told in store that the phone and an accompanying tablet would be free, then received bills for both. The fraud case they opened was denied and the return request refused, leaving them liable for devices they accepted on a verbal representation. Nothing in the transaction record captures what the associate said, so the dispute reduces to the customer''s word against the billing system.

Builders of no-code Selenium-based test runners cannot reliably confirm that a click actually affected the target web application, since apps may attach event handlers asynchronously and a successful click() call only means the event fired, not that it was handled. This causes false test failures and makes generic, app-agnostic retry logic risky to implement without duplicating actions.

Homeowners seeking FHA loan modifications or COVID forbearance relief receive shifting and contradictory denial reasons from their servicer while foreclosure proceedings continue in parallel. The lack of a clear, consistent modification review process leaves borrowers unable to secure an affordable payment before losing their home.

An indie developer building a social media post scheduler found that Meta requires formal business verification (with legal/corporate registration) before any user besides the developer can authenticate via the app, effectively blocking solo/unregistered developers from launching multi-user products on Meta's platform. He worked around it by open-sourcing and self-hosting the tool instead.