Airtable, Notion and no-code for clinic scheduling: how far it goes.

Why the question comes up now

A manager who decides to abandon the spreadsheet rarely jumps straight to specialized software. She looks first at what is already lying around the clinic. Airtable often tracks equipment or supply requests, Notion holds the onboarding procedures and the meeting notes, and both look capable of housing a schedule without anyone having to sign anything new.

That instinct is largely sound, which is what makes the question interesting. No-code in 2026 has little to do with the tools of 2018: relations between tables hold up, calendar views are usable, automations fire on their own. The honest answer is therefore not binary, and it works up to a point that can be named precisely. For the coded version of the same question, when someone in the clinic offers to write a real application, our comparison of Synchro and a DIY tool covers that other path.

What no-code does well

Structure, in one evening

A workable structure takes two or three hours: one table for people, one for blocks, one for absences, one for offices, and links between them. The gain over a spreadsheet is immediate and it is structural. A physician's name stops being text retyped into forty cells and becomes a record everything else points to. Fixing a spelling, changing an assignment, removing someone who leaves the clinic, all of it happens in one place.

A spreadsheet never managed that, which is the first reason a base feels so convincing in the early weeks. The copy-paste errors genuinely disappear.

Views, without writing a line

The same table renders as a grid for the manager, a calendar for the team, filtered by person on the phone of a physician who only wants to know where he is on Tuesday. Nobody programmed anything, and picking a date field was enough. A clinic that used to email a monthly PDF finds something genuinely better here, and it is often the moment the team adopts the tool.

The form that replaces the email thread

Time-off requests and availabilities normally arrive as twelve emails, three texts and a note left on a counter. An Airtable form or a shared Notion base lands them all in one place, dated, with the person's name and the reason in the right fields. The manager stops digging through her inbox to reconstruct who asked for what. This is one of the rare places where no-code fixes a real problem instead of relocating it.

Automatic reminders

A message the day before an on-call shift, an alert when a time-off request has sat for five days, a notification when someone edits the published grid. These automations are configured without code and they work. They also attack the right half of the problem, propagating a change, which is where most schedules actually fall apart.

The exact line where it gives way

A base records, it does not refuse

This is the line, and everything else follows from it. Airtable and Notion validate the contents of a cell: a date is a date, a link points to a record that exists, a required field is not empty. What they do not do is evaluate a rule involving several records at the moment you create one.

An office that takes one person at a time, a physician on vacation from the 12th to the 19th, a new recruit who is never put on call alone: each of those rules is checked by looking at rows other than the one being written. Nothing in these tools fires to stop you. You place two people in office 3 on Tuesday morning, the base saves two valid entries and renders a spotless calendar.

You can of course add a formula that turns a row red when it detects a duplicate, and many well-built bases do exactly that. The problem is the delay. The warning appears after the entry, in a view someone has to open. An assignment engine works the other way, holding every constraint at once and refusing the impossible cell as it is proposed. That is the difference between a dashboard that flags and a system that prevents.

Nobody assigns on your behalf

The base solved storage without touching the decision. You are still the one opening the calendar view and placing names one at a time, holding vacations, hospital days and everyone's half-days in your head. Thirty physicians, some fifty blocks a week, a dozen constraints per person: the number of combinations passed what a brain arbitrates correctly at 9 p.m. on a Sunday a long time ago.

No-code has therefore moved the storage without touching the work. That is exactly what clinics report when they leave Excel for a base and, six months later, still spend the same number of hours building the period. Our article on the limits of the spreadsheet lays out that hours calculation, and it carries over almost unchanged to a no-code base.

A rollup counts shifts, it does not weigh them

Airtable and Notion are very good at sums. A rollup field shows that every physician has five shifts this quarter, and the manager breathes out. Then someone points out that he collected three Friday evenings and two 24 Decembers while a colleague accumulated Tuesday mornings, and the discussion restarts from zero.

Weighting means giving each type of shift a value, having the team accept it, and having the tool enforce it during distribution. A number displayed after the fact settles no dispute, it feeds one. Giving on-call shifts a value is what lets you answer with a calculation rather than an impression.

A walk-in block is not a row like the others

In a base it becomes an entry with a date, an hour and a linked physician, no different from a team meeting. Nothing says that an empty block is a problem of a different nature than an unused office, that those hours count toward the accessibility obligations of the MSSS funding and professional support program for FMGs, or that an approved withdrawal immediately raises the question of who picks the block up. Patients still show up at 1 p.m. Walk-in distribution follows rules that have no reason to exist inside a general-purpose tool.

The shared office meets the same fate. A clinic almost always has more professionals than rooms, and placing offices at the same time as people is what avoids the standing negotiation in the corridor on Monday morning. A base can store the list of offices, and it does not assign them.

What a physician sees when the base opens

A schedule mixes information everyone must see with information that is nobody's business. That Dr. Tremblay is away from the 12th to the 19th publishes fine. The reason for the withdrawal does not.

In Notion, permissions are granted at the page level and database entries inherit from the database, so a physician invited to the base sees everyone's rows with whatever fields they contain. Airtable solves part of this by giving read-only access away for free, but a reader sees the whole row, withdrawal reason included, and restricting field by field takes a higher plan. The usual workaround is a second, stripped-down base for distribution, synchronized with the first, which adds one more component to maintain.

November, when the rules change

A scheduling base ends up holding formulas, filtered views, four or five automations and two synced tables. All of it was built by one person, often a resourceful assistant or a physician who likes tinkering, and all of it lives in that person's head.

Then November arrives. The clinic adds a site, on-call distribution changes, a new type of block appears. Someone has to reopen the formulas, work out why that one filtered on that field, and redo the automations without breaking the views. If the person who built the base has left or simply no longer has the time, the clinic is left with a tool nobody dares modify, which is precisely the situation it left the spreadsheet to escape.

The arithmetic: limits and seats

Two numbers deserve to be on the table before you commit, because both are verifiable in the vendors' own published documents.

The first is volume. Airtable documents 1,000 records per base on the free plan, 50,000 on Team and 125,000 on Business. A 30-person clinic writing one row per person per half-day, across roughly 250 working days, produces about 15,000 rows a year. The free plan therefore lasts three weeks, and the Team plan a little over three years before you have to archive history or move up a tier. Nothing disqualifying in that, but it is a deadline worth knowing before you build on it.

The second is the seat, and this is where the pricing argument deflates. Airtable does not bill read-only collaborators on any plan, and a shared view link opens without an account at all. A clinic therefore does not pay thirty seats so that thirty people can read the schedule: two managers with Editor access are enough and the rest of the team consults for free. At US$20 per seat per month on the Team plan, the bill looks like US$40 a month, not US$600. The catch worth knowing is the Commenter role, billable on Team just like Editor and free only from Business up, so a physician who needs to write "I can't do that Thursday" under a block stops being a free reader. Notion works along similar lines, with free guest access limited to the pages a person is invited to.

Price should therefore not be what makes you hesitate. A well-built Airtable base costs less than just about any specialized tool, Synchro included at $10 CAD per user per month. What has to be settled is what the tool does with your thirty people once they are in it, and what the hours of manual assignment keep costing you.

Law 25 when you are the one building the base

A clinic schedule is full of personal information, even with no patient chart anywhere in it. A physician's availability, absences, preferences and withdrawal reasons all qualify under Law 25, whose main obligations have applied since 22 September 2023.

Section 17 of the Act respecting the protection of personal information in the private sector requires, before communicating such information outside Quebec, a privacy impact assessment weighing the sensitivity of the information, the purpose of its use, the protection measures provided including contractual ones, and the legal regime of the destination jurisdiction. The communication must also be covered by a written agreement. That is the part no-code shifts onto you: when you build the base yourself, your clinic makes the communication, and the assessment is yours.

On hosting, both vendors publish their position. Airtable hosts in the United States by default and offers data residency in Europe or Australia to Enterprise Scale customers only. Notion hosts in the United States by default and offers the European Union, Japan and South Korea to its Enterprise customers. No Canadian option on either side. A no-code build also multiplies the suppliers, since the vendor, the automation service and the email sender each come with their own chain of subprocessors to assess.

Where to start if you want to try anyway

The exercise has real value, if only for laying out rules the clinic never wrote down. A few precautions make it conclusive faster.

1. Run the conflict test on day one. Deliberately place two physicians in the same office at the same hour, then put someone you just marked as on vacation into a walk-in block. If the base saves both without a word, you know its limit before sinking three months into it.
2. Time the build of a real period. Build a full month with your actual staff, stopwatch running. Compare it to what the job takes today. If the number does not move, the base solved storage and not the work.
3. Write the rules before the tables. What settles two physicians wanting the same block, what happens when a block stays uncovered, who has the final say. That document stays useful whatever tool you end up with, unlike the base.
4. Decide immediately who sees what. Separate what gets published to the team from what stays with management, and check that the tool can hold that separation on the plan you intend to pay for.
5. Name a second person. Someone else has to be able to open the formulas and automations next September. If nobody volunteers, you have just recreated the file only one person knows how to open.

Where Synchro fits

Synchro builds schedules for medical teams in Quebec using the objects of the job rather than generic tables: walk-in blocks, weighted on-call shifts, shared offices, transfers between colleagues with a record of who accepted and when. Each physician writes their constraints in plain language and the engine decides what can be placed, which gives you the ease of entry of a no-code base without its silence at the moment of the error. Natural-language constraints rest on exactly that division. In our pilot FMG in Saint-Hyacinthe, about thirty people managed in a spreadsheet took roughly 30 management hours a month, corrections after publication included, and about one hour with the tool. Those are our own measurements, not a network average.

What Synchro does not do: no patient data flows through it and it does not replace your EMR, whether Ofys, Omnimed, Medesync or Kinlogix, all of which consume the schedule rather than produce it. The assessment required by section 17 still has to be done, here as anywhere else. The difference lies in how many suppliers you have to assess, and in the fact that a vendor you hold an agreement with answers for what it hosts.

Further reading

• Synchro and a DIY scheduling tool compared
• Why Excel is not enough for FMG scheduling
• 8 moments that wreck your FMG schedule
• Best scheduling software for a Quebec FMG: 7 tools compared
• FMG scheduling software: the complete guide