When Do We Talk to AI? A 10,000-Conversation Study: Tuesday at 11 AM Wins

Artificial intelligence gets talked about as entertainment, a toy, or even a replacement for human contact. The data tells a different story: people talk to AI during office hours. Alongside our AI chat, we also run a free ChatGPT alternative in 22 European languages — and it’s from this free version that we anonymously analyzed 10,000 conversations to measure when people start talking to artificial intelligence, how long they keep going — and who writes it the longest messages.

A note on privacy up front: we worked exclusively with aggregated, anonymized data — no IP addresses, no publishing the content of a single conversation. Paying customers’ data never enters our studies. Details are in the methodology at the end of the article.

Key findings at a glance

Daily rhythm: AI keeps office hours

For every conversation, we know the moment the user started it. Converted to local time in each market, Europe’s day with artificial intelligence looks like this:

The curve tracks a classic workday: a steep rise between seven and nine, a peak at 11:00, a full morning and afternoon — and just a wind-down in the evening. The quietest hour is four in the morning. Only 10% of conversations fall in the nighttime hours (22:00–06:00), while morning and afternoon together account for 75%.

Night owls: Swedes and Greeks stay up, Germans sleep

At the extremes, the gap is fivefold: Germans are the picture of nighttime discipline (3.6%), and Czech is also low (6.7%) — Central Europe works with AI, then goes to sleep. The south and the Balkans stay up noticeably more.

The Spanish mystery: the language that never sleeps

Spanish set us a lovely trap in the data. If we mapped it to Madrid time like the other European languages, it would appear that 38% of Spanish-language conversations start at night — peaking at six in the evening. That’s not Spanish insomnia. Six in the evening in Madrid is noon in Mexico and 1 p.m. in Buenos Aires: most Spanish-language conversations with AI come from Latin America. The same holds, on a smaller scale, for Portuguese (Brazil) and, of course, English (the US, Australia, the whole world). It’s a useful reminder that language isn’t country — for Czech or Slovak, mapping language to market is reliable; for global languages, it isn’t. That’s why we excluded English, Spanish, and Portuguese from all time-based metrics entirely: any single local-time conversion would just distort the numbers for them.

The week: Tuesday rules, Saturday is quiet

To keep the day-of-week comparison from being skewed by the fact that our one-month window contains some days five times and others four, we normalize traffic per occurrence of each day. Result: an index of 1.00 = an average day.

Tuesday is the strongest day of the week (1.26) — and the whole Monday-to-Thursday block stays high. By Friday, traffic already drops noticeably (0.95), and the weekend is a cliff: Saturday sees 60% of an average day’s traffic. Overall, a weekend day has just 55% of a weekday’s traffic. If AI were mainly entertainment, we’d expect the opposite.

The weekend dip holds across all seventeen measured languages without exception — only its depth differs. Czechs have one of the deepest weekend slumps in Europe (0.42): on weekends, they really do switch off from AI. Estonians and Greeks, by contrast, carry more of their traffic into the weekend.

How long conversations last: half of people ask one question

We measure conversation length by the number of user messages, in three tiers: one message, two messages, or a longer exchange of three or more.

Overall, 52.8% of conversations end after a single message — for half of people, AI is a one-off tool: question, answer, done. At the other end, 29.8% of conversations have three or more messages — nearly a third of users hold a real conversation with AI.

And here’s the study’s nicest contrast: the French (41.5%) and the Czechs (38.4%) hold longer conversations the most often — the Germans almost never (15.3%). Fully 79.6% of German conversations are a single message. The next chart shows this isn’t a lack of interest.

Who writes AI the longest messages

Germans write the third-longest messages in Europe (206 characters) — combined with the previous chart, this forms a distinct communication style: a German writes everything at once, in one long, precise message, and that’s it. Czechs and the French are the opposite: shorter messages (121 and 144 characters respectively), but a real dialogue — they ask, clarify, react. Two ways to work effectively with AI: a monologue with instructions, or a conversation.

By the way — that a longer conversation also changes behavior, we know from our politeness study: in longer conversations, people are noticeably more polite to AI.

How we measured this

• Sample: 10,000 anonymized records from the free ChatGPT alternative by GuideGlare (no-registration version), June 23 – July 24, 2026, 22 language versions; after removing 6 technical duplicates, we worked with 9,994 conversations — the same sample as our AI etiquette study. Rankings include languages with at least 50 conversations (Danish and Finnish were excluded). Sample sizes range from 53 conversations (Slovenian) to 2,775 (English); full counts by language are in the downloadable data.
• Time: for each conversation, we know the start time (UTC); we convert it to local time using each language market’s summer time zone (Central and Western Europe +2; Romanian, Greek, Bulgarian, the Baltics and Finnish +3). English, Spanish and Portuguese are excluded from all time-based metrics (n = 3,774 combined) — these are global languages whose users write from multiple continents (Latin America, the US, Australia, Brazil…), and a single local-time conversion would be an artifact for them. We illustrate the artifact (hypothetically mapping Spanish to Madrid time → 38% “nighttime” conversations) in a box in the article. The time-based metrics therefore rest on 6,220 conversations in languages with a clear home market.
• Days of the week: normalized to full local days from June 24 to July 23 (30 days; incomplete edge days excluded). In this window, Wednesday and Thursday each occur five times, other days four — so we compute averages per occurrence of a day, not raw sums.
• Conversation length: measured by number of user messages, given in three tiers: one message, two messages, three or more — longer conversations are counted as a single group. Conversation duration isn’t measurable — individual messages carry no timestamps.
• Statistics: percentages are shares within a given language; for key values we compute Wilson 95% confidence intervals (full figures in the downloadable data). The weekend index is a ratio of averages; we don’t report a confidence interval for it — treat it as indicative for small samples.
• Privacy: we worked exclusively with aggregates — no IP addresses, no user identification, no publication of conversation content. Paying customers’ data never enters our studies.
• Limits: the sample consists of users of one free tool — the numbers describe behavior on our chat, not the population of an entire country. This is a one-month summer snapshot; the holiday period may actually dampen the work pattern, so the real “office” character of the traffic could be even more pronounced. Repeating the study will show the trend.
• Data downloads and citation: download the complete aggregated data (hourly distribution, days of week, conversation lengths × 22 languages × confidence intervals) as a CSV. You’re free to reuse the figures and charts in articles, blogs, and presentations — just credit “Source: GuideGlare” with a live link to this page; details in our citation guidelines. We’re happy to send print-quality charts — just email info@explicaire.com.

Try it yourself

The data for this study comes from our free ChatGPT alternative — it works in all 22 languages covered by this study, needs no registration, and the full version is GuideGlare’s AI chat. Feel free to try it at three in the morning on a Saturday — you’ll improve the statistics. And if you’re curious what all those conversations are actually about, read our companion study What People Ask AI.