The AmplifierReal cases · Personal · Extended family
Vikram, 45: twelve relatives, four generations, one week in Rajasthan
Parents who want temples, teenagers who want nothing, a brother-in-law with a budget and a sister with opinions.
In brief
The situation. Vikram has been made, by family consensus and against his will, the planner of a week-long trip for twelve people: his parents in their seventies, his and his sister's families, two teenagers, a nine-year-old, and a brother-in-law who will audit every rupee. Last time this happened, in Goa, three people did not speak for a month. His first instinct was a travel app that "builds itineraries instantly". His second instinct, better, was to close it.
Step 1 — Define the End State
He writes it on a Sunday: "My parents come back saying they saw the temples they wanted without being rushed. The teenagers have one day that was theirs. The nine-year-old is not bored for more than an hour at a stretch. My brother-in-law's spreadsheet balances. Nobody stops speaking to anybody. And I get two evenings when I am not the tour manager." Six audiences, each named, each with one thing they want.
Step 2 — Map the Flow
His list: dates, budget ceiling, one non-negotiable per person (he phones each of them — a task in itself), route, stays, travel between, food logistics. AI review adds: buffer days (he had none), the temple closing days in the week he chose, a mobility check for his parents at each site, and a shared document so the brother-in-law can see the numbers before he audits them. It suggests a group chat poll for everything; he declines — polls are how the Goa silence started.
Step 3 — Assign Tasks
What the week is for: his. Collecting each person's one non-negotiable: his, by phone. Generating three route options that fit the non-negotiables and budget: AI, low cost. Choosing between them: his. Checking timings, closures, distances: AI with web access, then verified. Persuading the teenagers: human, unavoidably. Booking: human, because the cost of a wrong booking is high.
Step 4 — Select the Tool
A conversational tool for the plan (iterative, moderate context — he pastes the twelve non-negotiables), and a tool with live web access for anything factual. Nothing needs setting up; this happens once.
Step 5 — Prepare and Direct
Criteria: no two travel days in a row; every site on the route has step-free access or a bench; each person's non-negotiable appears on a named day; total under the ceiling with 10% held back; two evenings marked "no plan". He briefs the tool as if it were a travel agent who has never met his family, and lets it write the prompt.
Steps 6 and 7 — Execute and Evaluate
Option one fails "no two travel days". Option two fails the budget by 15% once real prices land. Option three passes and is dull — every day the same shape. He iterates: asks for variety within the constraints. The timings check finds the fort his father most wants closed on the day the plan has it. Small fix, moved.
Step 8 — Integrate and Stress-Test
Assembled and read as each of the twelve. Small fixes: the nine-year-old has a four-hour car day with nothing; a stop is added. Structural: when the brother-in-law sees the shared sheet, the numbers do not add up, because two stays were priced per room and one per person — a Step 3 assumption mismatch, two tasks done on different bases. Re-priced consistently. He cuts one night to hold the 10%.
What changed
The trip happens. The teenagers' day is a desert camp they chose. Nobody stops speaking to anybody — or at least not for a month. Vikram gets his two evenings, one of them spent explaining to his sister how he planned it, which she does not believe took him a Sunday.
Name every person and their one non-negotiable before you open any app. Hold back 10%. Price everything on the same basis — the family budget's structural problem is almost always two tasks that assumed different units.
A composite case, not a client engagement; the person is invented, the situation is not.