The short version. Make and Zapier are the same decision made two ways, and the pricing gap between them is larger than anyone tells you. Roughly $12 for 10,000 operations against $19.99 for 750 tasks. Airtable is where the operational data should live once a spreadsheet stops coping. ChatGPT or Claude writes the SQL and the formulas and, more importantly, explains them. Notion is where the process gets written down, which is the thing that survives you.
Prices below were read off each vendor’s own pricing page on 15 August 2026, except the assistant tier. OpenAI’s page blocked our request, so we describe rather than quote it.
Building the shortlist for a team rather than yourself? We maintain directory shortlists for operations managers, reviewed on a schedule.
How we ranked a builder’s stack
Operations changed shape in the last two years. The person who used to file a ticket now writes the query, wires the automation and ships the internal tool. That is a genuine expansion of what one competent person can do, and it comes with a failure mode nobody warns you about: the distance between something that works and something the business can depend on.
So we weighted:
- Does it explain itself? Generated code you cannot modify when requirements change is a dependency you created, not a capability you gained.
- Does the pricing survive success? Automation tools are cheap until the thing you built gets used, and then the meter is the cost.
- Does it work without an engineer? If the setup assumes an IDE or a terminal, it is not for this reader, whatever the marketing says.
- Does it leave something behind? Ops work that lives only in one person’s head is a risk the organisation is carrying without knowing it.
1. Make: best automation value by a wide margin
Free tier with 1,000 operations a month. Core $12/month for 10,000 operations. Pro $21/month. Teams $38/month. Higher operation tiers scale from 20,000 upward, and annual billing is advertised at 15% or more off. Verified 15 August 2026.
Start with the arithmetic, because it is the single most useful thing in this article. Make’s Core plan is $12 a month for 10,000 operations. Zapier’s entry paid tier is $19.99 a month, billed annually, for 750 tasks. The units are not perfectly comparable, an operation and a task are counted slightly differently, but they are close enough that the gap is roughly an order of magnitude in cost per unit of work.
For an ops manager, that difference decides what you are willing to automate. On a 750-task budget you ration: only the high-value routines are worth wiring up, and anything firing on every record is out of the question. At 10,000 operations you stop thinking about it, and that changes which problems you are willing to solve.
Make’s visual scenario builder also handles branching, iteration and error handling more naturally, which matters as soon as your automation has to do something conditional. And real operations work is almost entirely conditional.
Who it’s wrong for: teams needing an app integration Make does not have, which is the one place Zapier reliably wins. Check your specific stack before committing. Also anyone who will be handing these automations to a non-technical colleague. Make’s power comes with more concepts to understand.
2. Zapier: best app coverage, and worth the premium only for that
Free tier with 100 tasks a month. Professional from $19.99/month billed annually for 750 tasks ($29.99 month-to-month), scaling to $49 at 2,000 tasks and $89 at 5,000. Team from $69/month annually for 2,000 tasks. Verified 15 August 2026.
Two automation platforms on a five-item list breaks a rule we normally hold to, and here it is unavoidable: this is one decision, it is the most consequential one on the page, and the honest answer depends on a fact about your stack rather than a preference.
Zapier’s case is integration breadth and maturity. If a tool exists, Zapier probably connects to it, and the connector probably works properly rather than technically. When the alternative is building against an API yourself, that is worth paying for, and paying a premium for a connector that exists is cheaper than a project.
What you should not do is default to Zapier because it is the name you know and then ration your automations against a 750-task ceiling. That is the common outcome and it is an expensive one. You end up paying more for less capacity and quietly deciding not to automate things that would have paid for themselves.
Who it’s wrong for: high-volume automation of any kind. Anything triggered per-record, per-order or per-event will exhaust these tiers fast, and the upgrade path is steep. Model your real event volume before choosing, not your number of automations.
3. Airtable: best for when the spreadsheet stops coping
Free tier. Team $20 per seat per month billed annually. Business $45 per seat per month billed annually. Enterprise Scale is custom. Verified 15 August 2026. Its pricing page does not state AI credit allowances per tier, so check that before buying if the AI features are your reason.
Every operations function has a spreadsheet that has outgrown itself: the one with the colour-coding convention only one person understands, the one that breaks when someone sorts it, the one three teams depend on and nobody owns. Airtable is the natural next step because it keeps the spreadsheet feel while adding the things that make data survive contact with colleagues. Typed fields, relationships, views per audience, and an API so your automations can reach it.
That last point is why it sits here rather than in a general productivity list. Airtable is the piece that makes Make or Zapier genuinely useful, because automations need somewhere structured to read from and write to. A spreadsheet in a shared drive is not that.
Who it’s wrong for: teams whose data belongs in the actual system of record. Rebuilding your CRM in Airtable is a well-trodden path to a shadow system that nobody maintains. And watch per-seat pricing. At $45 a seat on Business, a twelve-person ops function is a real budget line.
4. ChatGPT or Claude: best for the code you have to maintain yourself
Both sell an individual tier around $20 a month. We describe rather than quote: OpenAI’s pricing page returned a 403 to us on 15 August 2026.
An enormous share of the code written inside companies is written by people who are not developers. SQL, spreadsheet formulas, regular expressions, small scripts. That work has always been done by pattern-matching against Stack Overflow, and this replaces that with something considerably better.
The high-value uses are specific: writing a query in the exact dialect your warehouse expects, unpicking a nested formula somebody left behind, converting a manual process description into a specification you can build against, and debugging an automation that fails silently on one branch.
The discipline that matters is asking it to explain, not just produce. Working code you do not understand is the fastest way to accumulate systems that only run until the requirement changes. Ask for the explanation, read it, and only ship what you could modify next quarter. That single habit is the difference between an ops manager who has gained a capability and one who has quietly created an outage waiting to happen.
Who it’s wrong for: anyone about to run generated SQL against production without reading it. A confidently wrong UPDATE is a different category of mistake from a wrong answer, and the model has no idea which of your tables matter.
5. Notion: best because the process should outlive you
Free tier. Plus $10 per seat per month. Business $20 per seat per month, which is where the full Notion AI sits, including agents and enterprise search. Custom Agents are $10 per 1,000 monthly credits. Annual billing advertised at 20% off. Verified 15 August 2026.
The characteristic operations failure is not a broken automation. It is that the person who built it left, and nobody can reconstruct what it does or why the exception on the third branch exists. Undocumented operations is an organisation quietly taking on risk it has not priced.
Notion’s value here is unglamorous: somewhere for runbooks, process documentation and the reasoning behind decisions, that people will actually open. The AI layer becomes useful precisely because that material exists. Search across your own documented processes answers questions a general model cannot.
Note where the AI is: full Notion AI is on Business at $20 per seat, not Plus at $10. Budget for the tier that has the feature.
Who it’s wrong for: teams with a working documentation system already. The migration is expensive and the win is marginal. Notion rewards the teams that never had one, which is most of them.
Ops teams this isn’t for
Engineering teams. You have an IDE, a repo and a deploy pipeline. Every constraint that shaped this list is one you do not have.
RevOps specifically. Your problem is revenue leakage between contract, billing and provisioning, which is a different toolset entirely.
Ops at enterprise scale. Your automation runs through an integration platform chosen by IT, and your constraint is governance rather than capability.
Anyone hoping to avoid learning anything. Every tool here raises your ceiling and none removes the need to understand what you built. The ops managers getting the most out of this shift are the ones who read the explanation.
What we passed over
n8n. A serious third option, particularly self-hosted, and genuinely cheaper at volume. Left out because self-hosting puts you back in the territory this list is defined against. If you can run it, you are closer to an engineering team than an ops one.
Power Automate. If you are a Microsoft shop it may already be paid for, and that is a real argument. It is left out because the answer is entirely determined by your licensing rather than by the product, so a recommendation adds nothing.
Excel and Google Sheets AI features. Improving fast and free with what you have. Check them before buying Airtable. For a lot of teams the spreadsheet has not actually stopped coping yet.
Dedicated AI SQL tools. Narrower than a general assistant at a similar price. The exception is when a tool produces code in your specific tool’s dialect and explains it, which is a real differentiator over generic output.
Operations questions
Make or Zapier, which should I actually buy?
Check whether Make connects to every app in your stack. If it does, Make is materially cheaper per unit of work. $12 a month for 10,000 operations against $19.99 for 750 tasks, and the gap widens with volume. If a critical integration is missing, Zapier’s breadth is worth the premium. That is the whole decision; almost everything else about them is a preference.
How do I avoid building things nobody can maintain?
Two habits. Ask the model to explain every piece of generated code and do not ship what you could not modify. And write down what each automation does and why, somewhere colleagues will look. Both are boring, and they are the entire difference between capability and risk.
Is it safe to run AI-generated SQL?
Against a read-only replica, generally yes. Against production with write access, not without reading it line by line. The model does not know which of your tables are load-bearing, and a confidently wrong UPDATE executes exactly as fast as a right one.
What does this stack cost?
Roughly $50 a month for one person. Make’s Core at $12, an assistant at around $20, Notion’s free tier, and Airtable free until you need seats. The costs that grow are per-seat pricing on Airtable and Notion, and automation volume on Make. Model those two before the team expands.
Should ops own automations, or should IT?
Ops should build them; the organisation should know they exist. The common failure is not ops building things, it is ops building things nobody else can see. Which is why documentation is on this list at all.
Will AI replace operations managers?
It is removing the manual execution layer that ops teams were often sized around. What it does not do is decide which process should exist, notice that a workflow has stopped matching how the business works, or own the outcome when it breaks. The exposure is to headcount doing the steps, not to the person designing them.
If you would rather compare a maintained shortlist than a listicle, see our directory picks for operations managers who build.
Last reviewed: 15 August 2026. Prices read off vendor pricing pages on that date, except the assistant tier, which we describe because OpenAI’s page blocked our request. Airtable’s page does not state AI credit allowances per tier; verify that yourself if it drives your decision. We re-review quarterly.
Leave a Reply