Explore Problems

A Google One subscriber cancelled their subscription after Google Docs kept prompting to summarize every file, even after they explicitly turned the AI feature off. The inability to fully disable AI suggestions pushed the user to look for an alternative cloud storage provider.

A customer was told by their agent that a policy was cancelled, but continued to be billed and was eventually sent to collections when they didn't pay. Follow-up conversations with the agent and owner resulted in denial rather than resolution, highlighting a gap in reliable confirmation of policy status changes.

A long-time policyholder filing a claim was not clearly told upfront about rental car coverage caps, leading to unexpected out-of-pocket costs and a poorly coordinated claims process across the agent, claims department, and rental provider. The lack of proactive, upfront communication about coverage limits made an already stressful claim worse.

A customer canceling a home insurance policy requested a refund by mail but the insurer instead sent it to an expired, discarded gift card used for a past payment. The insurer refuses to provide transaction records or an alternate resolution, leaving the customer unable to recover the funds.

Users doing repetitive work report spending time copying the same fields between different tools by hand. Because each instance takes only a couple of minutes, it stays below the threshold that would justify existing integration platforms, leaving a persistent low-grade friction for individuals and small teams.

AI-driven project management platforms generate risk assessments and forecasts from ticket status data, but predictions become inaccurate when project managers fail to keep statuses current. The underlying problem is that AI insights are only as reliable as the manual data feeding them, undermining trust in automated risk calls for teams with inconsistent update habits.

A customer was overcharged $90 and told on multiple occasions — via chat and in-store — that the error would be corrected, only to have the provider ultimately refuse to fix it. This shows a gap between what front-line billing support promises and what actually gets resolved, leaving customers with no reliable escalation path.

A developer using Cursor reports the tool increasingly nudges users toward its own cloud agent and Grok models through frequent prompts and UI changes, including a silent default-model switch that wasted ten minutes of token spend on a task another model solved in thirty seconds. The friction pushed the user to migrate to an alternative stack (Pi plus OpenRouter) for more predictable model control.

Carvana canceled a customer's vehicle purchase just before delivery and promised a $1,290 shipping deposit refund within 7-14 days, but the money was never released. The company later admitted internally that the required Treasury ticket was never submitted, leaving the customer chasing the refund for over a month with no resolution.

A customer following prepaid return-envelope instructions given by an AT&T/Asurion representative continued to be billed monthly for the device because, unbeknownst to them, the carrier required in-store returns instead. The mismatch between what representatives tell customers and what the carrier's systems actually require results in months of wrongful charges before the customer discovers the discrepancy.

A customer was quoted an $85/month AT&T plan with no activation fees but was billed significantly more for three consecutive months, and customer service refused both to acknowledge the discrepancy and to issue the promised refund. This reflects a recurring telecom pattern where the quoted price at sign-up diverges from the actual bill with no enforcement mechanism for the customer.

Small creators using a DM-automation and link-in-bio platform were abruptly restricted to 25 contacts when the vendor quietly slashed its free tier, cutting off access overnight. The problem is a platform-driven pricing change that leaves creators without warning or an affordable path to keep using comment-to-DM automation and link-in-bio tools.

A prospective T-Mobile customer who never received service due to a shipping error was charged a deposit and endured a months-long refund process requiring five in-person store visits and escalation attempts by store staff. The dispute-resolution process for billing errors is slow and requires excessive manual follow-up even when store employees are cooperative.

A product creator describes noticing that existing AI products are either functionally useful but forget the user after each session, or emotionally warm and persistent but unable to take real action. This framing points to an unmet need for an AI assistant that combines lasting memory with actionable capability.

A customer's confirmed truck rental reservation had its pickup location moved an hour away the day before pickup, with no call or email, and the company then required an additional fee to use a closer location. This reflects a structural lack of notification when a rental company unilaterally changes reservation terms.

A telecom customer discovered an unauthorized phone line and device coverage charge had been billed for over four years without their knowledge or consent, and spent nearly 20 hours across calls and in-person visits attempting to get a $2,400 refund, with calls repeatedly dropping while on hold. This points to weak internal controls for unauthorized account changes and a dispute-resolution process that resists making customers whole.

A customer attempting to cancel a not-yet-active internet service order was given repeated conflicting information by support staff, with the account still showing the order as pending days later despite being told it was cancelled. This reflects a common friction point where cancellation requests are hard to confirm or verify, leaving customers unsure whether unwanted billing or equipment shipments will follow.

A customer's card was authorized for a partial large purchase that was then canceled, but the retailer held $7,000 of their funds well past the promised release window, and neither store nor customer support could locate the transaction or provide a timeline. This illustrates a common gap in high-value retail transactions: payment holds aren't reliably tracked internally, leaving customers unable to get resolution when funds are stuck.

A user reports that Notion works fine for basic tasks like a quick checklist, but becomes extremely difficult to work with once requirements grow even slightly more complex, leading to hours of wasted effort trying to fix a broken setup. This points to a structural complexity cliff in flexible no-code workspace tools.

Users of full-featured project management platforms like Jira find the interface and workflow and permission configuration complex to learn, requiring significant administrative effort, while support responsiveness is inconsistent. This creates friction for teams onboarding new users or setting up custom processes.