Langfuse

A new inquiry hall opened in the city, where guests came to ask about directions, accounts, and contracts. The clerks answered quickly, and the shopkeeper, hearing only the last sentence at the counter, often felt business was going much more smoothly.

But some customers came back complaining: some had been directed to the old dock, some had received expired price lists, and some had been refused things that should have been handled. The shopkeeper questioned the clerk about what he had been thinking at the time, but the clerk only remembered roughly; the details leaked away like water.

The shopkeeper first had the clerk write a retrospective summary. The summary was sometimes too flattering, and sometimes forgot key turning points. Which question caused the misunderstanding, which document influenced the answer, and at which step things began to drift—still unclear.

So he hung a long scroll on the back wall of the inquiry hall. For each reception, everything was recorded in order: the guest's original words, which registers the clerk consulted, how he reasoned, when he hesitated, how he finally answered, and whether the guest was satisfied.

For the first few days, the librarians found it troublesome. Later, after a misdirected route was reviewed, everyone realized the problem was not in the final sentence, but at the beginning, where “east dock” was heard as “old east gate,” and every later step followed the misunderstanding.

The long scroll also helped the manager compare different curators. Who was steady on complex questions, who was often led astray by similar place names, which types of questions were most prone to error — none of it relied on impressions anymore. When changing the rules, one could also see whether performance before and after had really improved.

The long scroll also recorded cost and waiting. The shopkeeper discovered that some answers, though correct, took too many detours; some questions should have been routed to a senior librarian from the start. After seeing the process, improvements no longer focused only on wrong answers but could also reduce waste. Later the shopkeeper added marks beside the long scroll: which step consulted which volume, which step changed the wording, which step waited too long. When new librarians joined, they no longer only studied excellent answers but could also see how a good answer was formed step by step, and where a bad answer began to go wrong. The shopkeeper also used the long scroll to compare old and new rules. After changing the reception script, whether wrong answers decreased and whether waiting time shortened could both be seen. The shopkeeper said: Only by seeing the process do you know whether to train people, change materials, or adjust the entry point. Later, improvements at the inquiry hall were all evidence-based and no longer decided by whoever spoke loudest.

The inquiry hall did not answer fewer questions because of the long scroll; instead, it became more willing to take on difficult matters. The shopkeeper said that looking only at results is like watching only the end of a play; to improve a question-and-answer business, you must see the entire path from question to answer.