← All analysisSource reading · The Ken, The Great Rewiring
Everyday commerce still runs on someone’s follow-up.
A reading of 16 everyday problems, from choosing insurance to receiving a parcel, and what they show about discovery, commerce and the growing habit of letting someone else act for us.
The Ken published these openings as prompts for its Case Competition 2026, which asks student teams to design AI agents for everyday consumer problems. The Ken is a subscription publication. This reading draws only on the short saved description of each opening and does not reproduce The Ken's text, images or reporting. Read the openings at the source for their full framing.
In each pattern below, “In the source” summarises what the saved descriptions share. “Our reading”, the tensions and the open questions are Bharat in Beta's interpretation. The openings are prompts, not evidence of demand, and a problem that recurs across them is not thereby validated.
Five patterns across the openings
01One person runs the household back office
In the source. The openings on meals, school and weddings each describe coordination that lands on one organiser or one family: preferences, messages, vendors, payments and schedules reconciled by hand.Our reading. The opportunity may be to share or hand over the organiser’s role rather than answer a single question faster. The unit of design is a household with several people and uneven workload, not one app user.
02Discovery fails before the search begins
In the source. Shoppers struggle to express fit, taste and budget; insurance buyers defer choosing cover; trip organisers juggle research and group decisions while availability changes.Our reading. The hard step is turning a vague or shared want into constraints a catalogue can use. Better ranking does not help a buyer who cannot yet state what they need, or a group that has not agreed. Preference memory and commitment may matter more than more choice.
03The purchase is the middle of the job
In the source. Appliance service, claims and refunds, tenancies and pet care all describe work that continues long after the first transaction: records, follow-ups, repeat visits, renewals and recovery.Our reading. Much of the friction sits after checkout, in work that a marketplace does not usually own. A service built here earns its place through a reliable history and follow-through, and its economics depend on provider participation and local density.
04Records and permission become the product
In the source. Household finances, delivery verification, family health and visa applications each turn on scattered records, and most on who may see or act on them. Several source directions name user control, consent or platform trust directly.Our reading. When software or another person acts on someone’s behalf, the design question becomes who is authorised, for what, for how long and what evidence remains. Permission is a product surface in its own right, not a setting added at the end.
05More prompts and more speed are not the answer
In the source. The goals opening notes that people abandon intentions despite reminders and trackers. The slots opening points toward fair allocation and venue-side coordination rather than faster booking bots.Our reading. Read together, these two openings caution against the obvious automation. Adding reminders or racing other users can make the experience noisier or less fair. The more useful change may sit at the moment of starting, or with whoever controls scarce supply.
Tensions to design around
These are Bharat in Beta's observations, not claims made by The Ken.
01Convenience against control
A service that remembers more and acts more needs deeper access to money, health records, deliveries and messages. Each step toward convenience raises the cost of a mistake and the bar for trust.
02Whose side is the service on?
The renting opening asks this directly. The same choice appears when an insurance comparison earns commission, when a slot tool serves venues or seekers, and when a delivery flow serves the purchaser or the platform.
03Frequent pain against willingness to pay
Meals and school messages recur daily but may feel too small to pay for. Visas, weddings and claims are rarer, with higher stakes. The first can build habit; the second may carry a clearer price.
04Working alone against needing institutions
Several jobs can start from what a household forwards or enters itself. Deeper value often needs schools, veterinarians, insurers, merchants or logistics providers to take part, and they need their own reason to.
Still unanswered
Questions the openings leave open
- Which of these jobs do people already pay an agent, planner, assistant or family member to handle, and how much of that payment is for coordination?
- Where does a household’s memory of preferences, documents and past decisions live today, and would people move it?
- What counts as consent when one family member acts for another, and how is it withdrawn?
- Which platforms or institutions must cooperate for each opening, and what do they gain from doing so?
- The source frames these as openings for AI agents. Which parts of each job need software, and which still need a person or an institution?
- How would someone tell a widely shared frustration apart from demand that will pay and stay?
Explore the source set
Open the problems behind the patterns
Each entry keeps The Ken's source link and recorded date, a faithful summary of the opening and a reviewed note with reasons to investigate, possible starting points and questions. The original prompts remain the primary record.
Browse all 16 entries