Guide · AI and the meter
How AI steps change the meter — tasks, credits, executions
An AI or agent step can multiply tasks or credits, or leave the job on a different meter. Totals are not published yet. Self-hosted is a separate lane. A score is not a monthly bill. This page does not add new scores.
Use this page if AI changed the count
Use
Use this page if…
You added — or you are about to add — an AI classify, summarize, or agent step to a small-team job, and the Zapier, Make, or n8n count no longer matches the non-AI version. You will stay in units. You will verify the module you actually run. This page names how that step counts on each meter, accessed 2026-09-18.
Skip
Skip if…
You still need to pick a product — that is Which automation tool for small teams. You want the same 5-step job without an AI step — Same job, three meters. You want loops, retries, and polls, not AI rates — How to meter automation. You want a monthly dollar from a /500 score. This page does not rank products.
A use-or-skip check for an AI or agent step on the meter.
- You added an AI or agent step and the meter no longer matches the non-AI job.
- You will count in the vendor’s unit — including a separate activity or gateway meter.
Read more
What this page tells you. How an AI or agent step changes Zapier tasks, Make credits, and n8n executions — or moves the job onto a separate activity or gateway meter. This page shows units only, with dated quotes from primary vendor pages, accessed 2026-09-18. It does not invent a monthly dollar, an overage dollar, or annual savings. This page publishes no annual savings. Verify before buy. It does not add new scores.
GuideSome scores not shown yetVendor pages accessed 2026-09-18Live scorecard
This page vs the two meter reads
Tasks vs credits vs executions is the dictionary of the three words. This page does not reprint that dictionary. It is the AI / agent delta.
Same job, three meters is one 5-step job counted as Zapier tasks, Make credits, and n8n executions at 50 / 500 / 5,000 runs. This page does not reprint that volume ladder.
How to meter automation is what to watch — usage, volume, credits — and what multiplies a non-AI job: loops, retries, and polls. This page does not reprint that watch list.
This page is only the AI / agent delta. Units only. It does not invent a monthly dollar.
How AI changes each meter
Each vendor still meters a different thing. An AI step can multiply that unit, or it can leave the job on a second meter. Do not treat one task as one credit as one execution as one activity. Dated quotes below are from primary vendor pages, accessed 2026-09-18. Zapier publishes this formula for AI by Zapier: Tasks used per run = (1 × model rate) + (number of tool calls × model rate).
Task or activity
Zapier
A task is any successful action that runs in Zapier. Only successful actions count toward your task usage.
AI by Zapier uses a model-tier formula. Zapier Agents do not use tasks — they use a separate activity quota. Copilot build and test does not use plan tasks.
Credit
Make
Credits are the currency you buy and consume to use Make. Only features triggered by scenario runs (apps, modules, and some in-app features) or the AI agent’s chat use credits.
Non-AI apps: 1 operation equals 1 credit. Make’s AI Provider bills operations plus tokens. A custom AI provider connection bills 1 credit per operation; you pay that provider for tokens.
Execution
n8n
An execution is a single run of a workflow.
One turn with an n8n agent counts as one execution and shares the workflow quota. Gateway credits are a separate prepaid balance on n8n Cloud Starter and Pro — not extra executions.
What an AI or agent step does to the count
A surprise invoice is often a unit surprise. An AI step can use more than one task or credit. An agent product can leave the task or credit meter entirely. This table stays in units. It does not invent an overage dollar.
Swipe sideways on a narrow screen. Not a scored table.
| What you add | Zapier | Make | n8n |
|---|---|---|---|
| AI step inside the job | AI by Zapier: tasks per run = (1 × model rate) + (tool calls × model rate). Standard 1x, Advanced 3x, Premium 5x. New paid steps default to Premium. A Premium step with two tool calls is 15 tasks. | Third-party AI you connect: 1 operation equals 1 credit; you pay that provider for tokens. Make’s AI Provider: credits from operations plus input and output tokens. Token-per-credit rates are on Make’s credits page. | The workflow is still one execution. n8n’s executions page does not add an execution per AI tool loop. Gateway credits are a separate prepaid balance if you select that option on a supported node. |
| Agent product | Zapier Agents do not use tasks. Each trigger, knowledge lookup, action, web browse or scrape, web search, or Chrome-extension message is 1 activity. Using activities does not affect task usage. | Run an agent: 1 credit per operation plus token credits on Make’s AI Provider, or 1 credit per operation on a custom provider. Chat also bills 1 credit per called tool. | One turn — you send a message and the agent responds — is one execution. Agents share the same execution quota as workflows. |
| Builder assistant | While you build and test with Copilot, you do not use plan tasks or add-ons. Tasks start when the Zap runs live. | Make’s credits page meters scenario runs and the AI agent’s chat. It does not name a separate Copilot-style builder quota on the pages verified for this guide. | n8n Assistant credits are a separate editor balance. Topping up Gateway credits does not add Assistant credits, and using Assistant does not spend Gateway credits. |
Sources and access date sit under Sources, accessed 2026-09-18. Zapier’s task-usage article also names Code, Lead Router, and MCP as different task rates — those are not the AI job below. n8n’s Gateway dollar rates live on a service pricing page that was not re-verified this pass.
Same 5-step job, plus one AI step
Start from the non-AI job on Same job, three meters: form trigger, sheet or CRM write, confirmation email, Slack, follow-up task. That is 4 Zapier tasks, 5 Make credits, and 1 n8n execution per successful run. Then add one AI classify or summarize step. Units only. This is not a monthly dollar and not a rank.
Swipe sideways on a narrow screen.
| Product and step | Unit | Per successful run | What changed |
|---|---|---|---|
| Zapier — no AI step | Task | 4 tasks | Same 5-step job as Same job, three meters. The form trigger does not count. |
| Zapier — AI by Zapier Standard, no tools | Task | 5 tasks | Add one AI classify step. Standard rate is 1x and that step is 1 task. |
| Zapier — AI by Zapier Premium, no tools | Task | 9 tasks | The same added step at Premium (5x) is 5 tasks. New paid AI steps default to Premium. |
| Zapier — AI by Zapier Premium, two tool calls | Task | 19 tasks | Vendor example: (1 × 5) + (2 × 5) = 15 tasks for the AI step, plus the original 4. |
| Zapier Agents | Activity | 0 tasks | Agents use activities, not tasks. Each listed agent action is 1 activity. |
| Make — no AI step | Credit | 5 credits | Same 5-step job. The form-read is a credit. |
| Make — custom or third-party AI module | Credit | 6 credits + provider tokens | One extra module is 1 credit. Tokens are billed by the provider you connected. |
| Make — Make’s AI Provider | Credit | 6 credits + token credits | One extra operation plus credits from input and output tokens. Rates vary by model. |
| n8n — workflow with or without an AI node | Execution | 1 execution | One production run. An AI node does not add a second execution. |
| n8n — one agent turn | Execution | 1 execution | Shares the workflow quota. Gateway credits, if selected, are a separate prepaid balance. |
The 4 / 5 / 1 baseline is the non-AI inventory on Same job, three meters. The 5 / 9 / 19 Zapier counts use Zapier’s published formula and examples. Make’s 6-credit rows add one operation; token credits stay dynamic. n8n stays one execution. Do not copy the 50 / 500 / 5,000 volume ladder here. Verify before buy.
Don’t turn a unit into a bill
A unit count is not a monthly dollar. A /500 score is not a plan. This page does not invent an overage dollar or a converted list. It publishes no annual savings.
- Stay in units until you can name the AI or agent step in the vendor’s meter — including a second meter such as Zapier activities or n8n Gateway credits.
- Then open the 5-step job on Same job, three meters if you still need the non-AI count at 50 / 500 / 5,000 runs.
- Verify the module you run. Premium vs Standard, tool calls, Make’s AI Provider vs a key you bring, and Gateway credits vs your own credential change the unit count.
Not tax or billing advice. See editorial standards.
After you can name the AI meter
You can count the AI or agent step in the vendor’s unit. Now pick how the rest of the job runs — without treating this page as a ranking.
- The non-AI 5-step volume ladder: Same job, three meters — tasks vs credits vs executions.
- Loops, retries, and polls: How to meter automation.
- Four questions, then a pick: Which automation tool for small teams.
- When to pick each: Make vs n8n vs Zapier — when to pick each · live scorecard (totals are not published yet; self-host is a separate lane).
Primary CTA stays Same job, three meters. This page does not reprint that volume ladder.
How this sits on the scorecard
How this sits on the scorecard
Totals are not published yet.
If you still need to pick a product before you count units, see Which automation tool for small teams. When-to-pick copy lives on Make vs n8n vs Zapier — when to pick each.
Product links
Sources
Primary vendor pages. Access date for every citation on this page: 2026-09-18. n8n’s Gateway service-pricing page was not re-verified this pass — those dollar rates are omitted. We do not invent search volumes, testimonials, or a converted list.
- Zapier — how task usage is measured · accessed 2026-09-18
- Zapier — AI by Zapier model tier pricing · accessed 2026-09-18
- Zapier — how Zapier Agents usage is measured · accessed 2026-09-18
- Zapier — what is Zapier Copilot · accessed 2026-09-18
- Make — credits · accessed 2026-09-18
- Make — credit usage for AI agents · accessed 2026-09-18
- n8n — understand executions · accessed 2026-09-18
- n8n — build and manage agents · accessed 2026-09-18
- n8n — Gateway credits · accessed 2026-09-18
FAQ
How does an AI by Zapier step count tasks?
Zapier publishes this formula: tasks used per run = (1 × model rate) + (number of tool calls × model rate). Standard is 1x, Advanced is 3x, Premium is 5x. New paid steps default to Premium. A Premium step with no tools is 5 tasks. A Premium step with two tool calls is 15 tasks. Bring-your-own-key is 1x. A run that hits 75 tasks pauses for approval. Copilot build and test does not use plan tasks.
Do Zapier Agents use Zapier tasks?
No. Zapier writes that Agents usage does not count as tasks. Agents use a separate activity quota. Each trigger, knowledge lookup, action, web browse or scrape, web search, or Chrome-extension message is 1 activity. Using activities does not affect task usage. Free includes 400 activities per month and counts tests. Pro includes 1,500 and does not count tests.
How do Make AI modules and AI agents use credits?
Non-AI apps: 1 operation equals 1 credit. Third-party AI you connect: 1 operation equals 1 credit, and you pay that provider for tokens. Make’s AI Provider bills operations plus tokens. Run an agent is 1 credit per operation plus token credits on Make’s AI Provider, or 1 credit per operation on a custom provider. Chat also bills 1 credit per called tool. This page does not invent a monthly dollar from those token rates.
How do n8n AI agents and Gateway credits count?
An execution is a single run of a workflow. One turn with an n8n agent is one execution and shares the workflow quota. An AI node inside a workflow does not add a second execution on the pages verified for this guide. Gateway credits are a separate prepaid balance on n8n Cloud Starter and Pro. They are not available on n8n Cloud Enterprise or self-hosted n8n. n8n Assistant credits are a third, editor-only balance. This page does not reprint Gateway dollar rates — n8n points those to a service pricing page that was not re-verified here.
Does this page turn those units into a monthly dollar?
No. This page stays in units: tasks, credits, executions, activities, and token-per-credit rates as printed on the vendor pages, accessed 2026-09-18. It does not invent a monthly dollar, an overage dollar, or annual savings. It publishes no annual savings. Verify the module you actually run.
Does this guide change the Make vs n8n vs Zapier scores?
No. Totals are not published yet.
Where should I go next?
For the dictionary of the three words, see Tasks vs credits vs executions. For the same 5-step job without an AI step, see Same job, three meters. For loops, retries, and polls, see How to meter automation. To pick a product, see Which automation tool for small teams. Totals are not published yet.
Affiliate disclosure
Make Partner (pc=stackscorecard) is live elsewhere on the site and is not linked from this page. n8n and Zapier stay official-site URLs with UTM only — no invented IDs and no affiliate ID. Rankings are not sold. Policy: Affiliate disclosure · Disclosures.