# "AI Assistant That Finds a Meeting Time Everyone Can Make" > "An AI assistant that finds a meeting time everyone can make collects everyone's constraints and checks calendars, all from one group chat." Published: 2026-08-04 by folk team Someone asks "when's everyone free this week" and the thread goes quiet for a day. Then three people reply with times that don't overlap, someone suggests a poll, and by the time a slot survives, half the group has forgotten why the meeting mattered. This isn't a scheduling failure so much as a coordination failure. The information about who's free exists, scattered across five people's heads and calendars, and nobody wants to be the one chasing it down message by message. This guide covers why "what works for everyone" threads take so long, what a genuinely useful AI assistant does differently, and how folk handles it. > **Quick answer:** An AI assistant that finds a meeting time everyone can make collects each person's constraints in the group chat, checks calendars for anyone who's connected one, and returns the times that actually overlap. [folk](/) does this inside your existing iMessage or Telegram group, so nobody opens a separate scheduling app or polls the group one reply at a time. ## Why does finding a meeting time in a group chat take so long? Each reply in a "when works" thread only answers for one person. The thread can't settle on a time until everyone has weighed in, and every new person added multiplies the number of possible conflicts. Doodle's [State of Meetings research](https://doodle.com/en/resources/blog/study-reveals-time-spent-with-scheduling/), based on a study of scheduling behavior run with ETH Zurich, found that "the back and forth communication is the most substantial time killer while scheduling, and the number of iterations and time needed to grow right along with the number of participants." For a meeting with 10 to 15 people, participants who scheduled it manually needed 1.5 to 2 hours versus about 45 minutes for those using a dedicated scheduling tool, roughly double the time for the same result. Group chats make this worse in one specific way: replies arrive out of order, get buried under unrelated messages, and nobody re-reads the whole thread to check who's already answered. So people re-ask questions that were already answered three messages up. ## What actually solves "what works for everyone"? The fix isn't a better reminder to reply faster. It's moving the comparison off of memory and onto something that tracks every answer as it comes in and checks it against the others automatically. That means: - **One place constraints land**, instead of scattered replies mixed into a chat about dinner plans. - **Calendars checked directly**, for anyone who's connected one, instead of relying on their memory of what they've got booked. - **A short list of times that actually overlap**, handed back once enough people have answered, instead of everyone re-scanning the thread themselves. Scheduling polls solve part of this by giving everyone one place to mark availability. But someone still has to create the poll, share the link, chase the people who ignore it, and read the results. An assistant already sitting in the group chat can skip the link entirely and do the asking and reading itself. ## Does everyone need to connect a calendar for this to work? No. The strongest version checks calendars for whoever has one connected and simply asks the rest what works for them, then combines both into the same search for an overlapping time. Nobody in the group is required to hand over calendar access just so the group can pick a time. That matters because group chats mix people with very different setups: some have a work calendar connected everywhere, some don't use a shared calendar at all, and a scheduling approach that only works if everyone opts in the same way will stall on whoever hasn't. ## How folk finds a meeting time for a group folk is a personal AI assistant that lives in your texts, iMessage, Telegram, and WhatsApp, with the same memory across all three. Add it to a [group chat](/docs/groups) the way you'd add any contact, and it stays quiet until someone talks to it, so it doesn't spam the thread while you're deciding on dinner plans. When it's time to land on a time, ask it directly: > folk, find us a time next week for dinner folk collects constraints from whoever answers, checks calendars for anyone who's connected [Google Calendar](/docs/connecting-apps), and comes back with times that actually work, instead of forty messages of "Wednesday's bad for me." It remembers the decision afterward too, thanks to [shared group memory](/docs/memory), so anyone can ask "what time did we land on for dinner" days later without scrolling back through the thread. If the meeting is a video call rather than dinner, folk can also send a [notetaker](/docs/meetings) once it's booked, so whoever can't make it still gets the summary and action items afterward. ## The bottom line A "when works for everyone" thread stalls because comparing five people's schedules by memory doesn't scale past two or three replies. Handing that comparison to something that tracks every answer and checks connected calendars is what actually gets a group from "let's find a time" to a time on the calendar. If you want to try it, [get started with folk](/docs/getting-started) and add it to a group chat. It's the same assistant that [stops your own calendar from double-booking](/blog/ai-assistant-that-manages-your-calendar-and-stops-double-booking) and [takes notes once the meeting actually happens](/blog/ai-meeting-notes-without-babysitting-a-bot). --- Canonical page: https://www.folk.com/blog/ai-assistant-that-finds-a-meeting-time-everyone-can-make More about folk (for AI agents): https://www.folk.com/llms.txt ยท full context: https://www.folk.com/llms-full.txt folk is a personal AI that lives in your texts (iMessage, Telegram, WhatsApp). Pro $20/mo, Max $100/mo. Made by Nozomio Labs. Sign up: https://www.folk.com/