AI Workflow Builders 2026: 9 Platforms Compared

Make's pricing page warns that one scenario run can use "anywhere from two credits to thousands of credits." Other builders count one run, however many steps it has.

For finance, that gap decides whether the automation bill is forecastable or a monthly plot twist (and finance teams rank plot twists just below unreceipted expenses).

The 2026 wrinkle is AI steps inside workflows, which some platforms meter in a second unit.

Here are nine AI workflow builders 2026 buyers should compare, ranked by fit and labeled by billing unit.

If you only read one section, make it this one:

Zite wins this list because you describe a workflow, it writes the steps, and a run counts once, however many steps. It's the AI workflow builder of 2026 that builds the app and real database around the workflow and logs every run, so you can check exactly what ran.

Make handles branching visual scenarios, Zapier brings the widest app catalog, and n8n gives technical teams self-hosting. Microsoft Power Automate fits Microsoft 365 shops, while Workato covers governed enterprise integration.

Kissflow suits enterprise approval apps, Gumloop leads with AI agents, and Lindy works as a Slack-native AI teammate.

Platform	What you pay for	Free plan	Paid from (as of Oct. 2026)
Zite	Workflow runs	Yes	$25/mo
Make	Credits	Yes, 1,000 credits/mo	$10.59/mo
Zapier	Tasks	Yes, 100 tasks/mo	$29.99/mo
n8n	Executions	Self-hosted Community edition	€24/mo
Microsoft Power Automate	User or bot licenses	Free license, standard connectors	$15.00/user/mo (billed annually)
Workato	Credits	Yes, 50K one-time credits	$100/mo
Kissflow	Annual agreement	No	Custom quote
Gumloop	Credits	No, 14-day trial	$37/mo
Lindy	Credits per active user	$50 in credits for 7 days	$49.99/user/mo

How do pricing units change the bill?

Every platform on this list meters something, and the unit matters more than the sticker price because it decides what happens to your bill when a workflow gets longer or busier.

Tasks. Zapier bills tasks per month, with 100 on the free plan and 750 on Professional's entry tier. Its agents bill in a separate unit called activities, so an AI-heavy setup can need two budgets.

Credits. Make says that as of August 27, credits are its billing unit, and most actions use one credit while some AI features use more. Workato's credits cover workflow tasks, API calls, rows processed, and agent actions under one meter.

Gumloop prices credits at $0.005 each, and they pay for model tokens, compute, and tool calls plus an orchestration fee. Lindy's credits are pooled, but every active user adds a seat to the bill.

Executions. On n8n, an execution counts once "regardless of complexity," so a twenty-step workflow and a two-step one cost the same single execution.

Runs. The top pick on this list meters the same way, one run per workflow however many steps it has, and test runs in the editor never count.

Licenses. Microsoft Power Automate charges per user or per bot, paid yearly, which makes headcount the number to watch. (Somewhere, a procurement team is already building the spreadsheet.)

1. Zite: for business teams that need an app and database around each workflow

The strongest AI workflow builder for a business team is the one that builds the screen, the data, and the automation together, and this platform does all three from a plain-English description.

Fit: Operations and business teams whose workflows need somewhere to live, such as an approval screen, an intake form, or a dashboard, plus the records behind them. You describe what should happen, and the builder writes the workflow and the app around it.

Highlights: A workflow can start on a schedule, on a webhook from a service such as Stripe or Slack, on a click in your app, or on demand. Inside a run it can read and write your database, call 30+ native integrations, run AI to summarize or classify, render PDFs, and send email.

The editor shows each workflow as a visual diagram of readable steps, with the code there if you want it. Every run is kept for 90 days with its inputs, outputs, and any error, and you can ask the builder what went wrong, then replay the run with the same inputs.

The data sits in a real database behind a spreadsheet-style view, with AI fields that categorize, summarize, and score records. The Zapier and n8n connectors are on every plan, which covers apps outside the native list, and apps stay inside your organization until you share them.

Cost: Free is $0, Pro is $25/mo, Business is $89/mo, Team is $300/mo, and Enterprise is custom, all billed monthly, with unlimited users, editors, and apps on every plan. Runs per month come to 1,000 on Free, 5,000 on Pro, 50,000 on Business, and 250,000 on Team.

Scheduled workflows start at Business, while webhooks, in-app triggers, manual runs, PDF generation, and email sending work on every plan. Granular permissions and roles start at Team.

Drawbacks: There's no self-hosting, so your data stays on the platform's hosting in the US, or the EU on request. Every plan carries SOC 2 Type II, but SSO and audit logs are Enterprise-only, which matters if IT wants both before anything goes live.

2. Make: the fit for ops teams with branching, data-heavy scenarios

If your workflow looks less like a straight line and more like a subway map, Make is where it'll feel at home.

Fit: Operations teams that have outgrown one-path automations and want visual control over branches, filters, and data without writing code.

Highlights: The scenario builder is visual and drag-and-drop, so you can see every path on one canvas. Maia, Make's conversational builder, creates automations and AI agents from a chat, and Make AI Agents (Beta) run on Make's AI Provider or your own LLM key.

The AI Toolkit, AI Content Extractor, AI Web Search (Beta), and an MCP server are included on every plan. For agents, you can set rules, add manual approvals, and read a reasoning panel that shows each decision, and the catalog covers 3,500+ apps.

Cost: As of October 2026, Free is $0 for 1,000 credits a month, two active scenarios, a 15-minute minimum interval between runs, and 7-day log storage. Billed monthly at 10,000 credits, Core is $10.59/mo, Pro is $18.82/mo, and Teams is $34.12/mo.

Enterprise pricing is custom. Code steps cost two credits per second of execution time and aren't on the free plan.

Drawbacks: Because every module action spends credits, a busy scenario can drain a plan faster than the slider suggested. When credits run out, scenarios stop until you add more, and credits expire at the end of the term.

SSO, audit logs, and custom roles are Enterprise-only, and hosting is cloud only, on AWS in the EU or North America. What I'd test in a trial is your heaviest scenario's credit use across a full week of real data.

3. Zapier: built for companies that need the widest app catalog

Zapier is the pick when the app you need to connect is obscure, because its catalog lists 9,000+ apps and 66,000+ triggers and actions.

Fit: Non-technical teams that want an easy builder and a good chance that every tool they already pay for has a ready-made connector.

Highlights: Zapier now pitches itself as "AI automation for the agentic era," with Zaps, Agents, MCP, Tables, Forms, and Canvas in one platform. You can build in plain language with Copilot on every plan, and paid plans add AI fields for steps that summarize or classify.

Every Zap gets an AI manager agent that, according to Zapier, detects issues and fixes them automatically. Zapier MCP lets Claude, ChatGPT, Cursor, and other AI clients take action across that app catalog, and the Human in the Loop app (paid plans) can pause a Zap for approval.

Zapier says 3.4 million companies trust it, and it lists SOC 2 Type II, SOC 3, GDPR, and CCPA on its site.

Cost: As of October 2026, Free is $0 for 100 tasks a month with two-step Zaps and one seat. Professional starts at $29.99/mo billed monthly for 750 tasks, and Team is $103.50/mo billed monthly with 25 users, shared Zaps, and SAML SSO.

Enterprise goes through sales. Zapier Agents bill separately in activities, with 400 a month on the free tier.

Drawbacks: Any multi-step workflow means paying, since the free plan stops at two steps. Heavy agent use puts you on two meters at once (tasks and activities), and bring-your-own-model and log streaming are Enterprise-only.

4. n8n: a fit for technical teams that want to self-host

For technical teams that want automation running on their own servers, n8n is the strongest option on this list.

Fit: Developers and technical ops teams who want visual nodes, code steps in JavaScript or Python, AI agents, and the choice to host it all themselves.

Highlights: The AI Workflow Builder runs on the tagline "Describe it. Build it. Ship it." Its assistant plans, builds, runs, and fixes a workflow, and the output is a normal n8n workflow.

MCP works both ways, so other AI platforms can call your workflows and your workflows can call MCP-enabled tools. You can also require human approval before an AI agent uses a specific tool, routed through Slack, Telegram, or n8n Chat.

The free Community edition is self-hosted with "almost the complete feature set," every plan includes unlimited users and workflows, and the project counts 207k GitHub stars. Cloud data is hosted in Frankfurt.

Cost: As of October 2026, n8n Cloud prices are in euros and count executions. Billed monthly, Starter is €24/mo for 2,500 executions, Pro is €60/mo for 10,000, and Business is €800/mo for 40,000 (self-hosted only), with Enterprise on a sales quote.

Drawbacks: SSO starts on Business, which is self-hosted only today, so SSO on n8n Cloud means Enterprise. Dedicated support is Enterprise-only too, and the free Community license bars hosting n8n for clients who want to build their own workflows.

Self-hosting also means you run the infrastructure. The company describes its license as fair-code and says it doesn't use an open source license, which your legal team will want to read before anyone installs it.

5. Microsoft Power Automate: strongest inside Microsoft 365 organizations

Microsoft Power Automate makes the most sense when your company already runs on Microsoft 365 and Teams.

Fit: Organizations standardized on Microsoft that want cloud flows, desktop RPA, and approval processes managed alongside the rest of their Microsoft 365 setup.

Highlights: Power Automate covers cloud flows, desktop flows (RPA), and process mining, and Copilot lets you create and edit flows in natural language. AI Builder models handle document and image processing, prediction, and analysis.

It offers more than 1,000 API connectors, or you can create your own custom connector. You can build and share approval processes across the organization, and attended RPA runs with a person at the keyboard.

Agents live in Microsoft Copilot Studio, sold separately with Copilot Credits, and Microsoft says an Azure subscription is required to use them.

Cost: As of October 2026, Premium is $15.00/user/mo (billed annually) for cloud flows, attended RPA, process mining, and premium connectors. Unattended RPA needs Process at $150/bot/mo (billed annually), and Hosted Process is $215/bot/mo (billed annually).

A free license for work or school accounts runs cloud flows with standard connectors only and no sharing, and there's also a free trial.

Drawbacks: Premium connectors, including non-Microsoft apps such as Salesforce, need a Premium or Process license. Each unattended bot runs one desktop flow at a time, and agents mean a second product with its own credits.

6. Workato: built for IT teams that govern enterprise integrations

Workato is built for the IT team that has to approve automation at enterprise scale and still know what's connected to what.

Fit: IT and enterprise architecture teams running integration, workflow orchestration, and AI agents across many systems at once.

Highlights: Workato pitches itself as the platform to govern and scale enterprise AI, and the toolset follows: Agent Studio, agents it calls Genies, AIRO for multi-agent work, Enterprise MCP, App and Data MCP Servers, and an AI Gateway.

Even the free plan includes Enterprise MCP and AI agents, and the tiers page lists 10,000+ app and system integrations. A Recipe Optimization Copilot is part of the toolset too.

Capped Recipes, introduced in July 2025, stop a recipe's cost from rising once it hits its task cap, and Genies are priced on a fixed-cost basis. That's welcome news for anyone who has watched an integration bill creep up quarter after quarter.

Cost: As of October 2026, Free is $0 with 50,000 one-time credits, and Pro is $100/mo for 3,500 credits a month, with up to 3 users per workspace and custom connectors. Enterprise is custom.

Workato doesn't label Pro's billing period, and its pricing FAQ describes Pro credits from 2,500 up to 50,000 a month, so confirm your allowance in writing before you sign.

Drawbacks: SSO and unlimited users require Enterprise, since Pro caps at 3 users per workspace. Pro data is hosted only in the US, with other regions reserved for Enterprise.

7. Kissflow: a fit for large enterprises replacing email approvals

Kissflow fits the large enterprise whose purchase approvals still travel by email and spreadsheet.

Fit: Large enterprises building workflow-heavy operational apps (procure-to-pay, approvals, and claims) with IT governance on top.

Highlights: You can describe an app in plain language or hand over a requirements document, and Kissflow says it generates complete applications "in less than 5 minutes." Agents can research and apply policy decisions, document extraction is built in, and AI usage is listed as unlimited.

Its AI page says Kissflow primarily uses Gemini 1.5 Pro. Plans include unlimited applications, unlimited integrations, 25+ AI-powered automation features, public and private cloud hosting, and implementation services.

You can pick a data region in the US, EU, India, or Australia. Kissflow says 10,000+ companies use it, including 50+ Fortune 500 companies.

Cost: As of October 2026, Kissflow publishes no prices, so you set up a consultation and get a quote for a fixed annual agreement. The pricing page promises no surcharges for AI or API usage, which makes the year easier to budget once the contract is signed.

Drawbacks: It's sales-led with annual contracts only, so there's no self-serve plan to poke at over lunch. Its AI page doesn't mention MCP or bring-your-own-model options, so ask about both, along with SSO and audit log details, during the sales process.

8. Gumloop: for ops and GTM teams building AI agents

Gumloop now leads with AI agents, and its pricing table labels the classic builder "Workflows (Legacy)."

Fit: Ops and go-to-market teams that want agents working in Slack, Teams, Gmail, and a real browser, built by the non-technical experts who know the work.

Highlights: You tell Gumball what an agent should do, and it builds or updates the agent, a feature from Gumloop's October 7, 2026 changelog. Agents come with skills, connectors, and triggers, and the browser agent supports human-in-the-loop handoffs.

The platform offers 400+ connectors, and Pro includes 35+ models, bring-your-own API keys, and a hosted MCP server, with unlimited agents, seats, and teams. Enterprise adds spend caps and approvals for costly actions.

Gumloop lists SOC 2 Type II and GDPR, and its enterprise page quotes Instacart crediting it with helping all of its teams adopt AI.

Cost: As of October 2026, there's no free plan, only a 14-day trial, and Pro starts at $37/mo for 20,000 credits (7,400 plus 12,600 bonus). Credits are $0.005 each and cover model tokens, compute, and paid tool calls, plus an 8% orchestration fee that rises to 16% with your own keys.

Drawbacks: Spend follows model usage, so it's harder to predict than a flat task count. When you hit 20,000 credits, agents stop until you switch on pay-as-you-go (off by default), and unused credits don't roll over.

Teams that want fixed, step-by-step workflows should also weigh the legacy label on the classic builder.

9. Lindy: the pick for teams that delegate work in Slack

Lindy is the pick when you'd prefer to ask a teammate in Slack over drawing a flowchart.

Fit: Teams that want to hand off work by chatting with an AI teammate, from answering questions to running scheduled routines and joining meetings.

Highlights: Lindy calls itself "The AI Teammate That Lets You Do More," and it works inside Slack, acts across your tools, and lets you choose the model. It ships with 40+ skills (plus any you create), MCP support, computer use, and 1,000+ integrations.

Approvals are built in, so anything with outside impact, such as sending an email or updating a ticket, waits for a person to approve it. Lindy lists SOC 2 and GDPR and says your data is never used to train models.

Cost: As of October 2026, Free is $0 with $50 in credits that last 7 days and no card required. Team is $49.99/mo per active user with 5,000 credits each, pooled across the workspace, and top-ups cost $10 per 1,000 credits.

Lindy's credit guide puts everyday asks at 2 to 250 credits, deep work at 250 to 1,000, and big builds at 1,000 to 2,500. Enterprise is custom and adds SSO, SCIM, audit logs, and a signed HIPAA BAA.

Drawbacks: Anyone who uses Lindy takes a seat, including people who only @mention it in Slack, so the bill grows with adoption. When the credit pool runs low, credit-using actions pause until it resets, and monthly credits don't roll over.

A buying checklist for 2026

A plan page tells you what a platform wants to sell, so walk every shortlisted option through the same seven questions.

Name the meter. Write down whether each platform bills tasks, credits, executions, runs, or licenses, then estimate your monthly volume in that unit.

Count every person who touches it. Lindy bills each active user and Power Automate licenses per user, while n8n includes unlimited users on every plan.

Find the second meter. AI often bills on its own line, since Zapier Agents use activities, Power Automate agents need Copilot Studio credits, and Gumloop adds an orchestration fee to model costs.

Ask where it runs. Self-hosting puts n8n on your own servers, Make hosts on AWS in the EU or North America, and Workato keeps Pro data in the US.

Check the approval point. Look for a real pause before an AI action touches a customer, such as Lindy's built-in approvals or the human approval n8n offers for agent tool calls.

Price the security tier. SSO is Enterprise-only on Make and Workato, comes with Team on Zapier, and means Enterprise on n8n Cloud.

Read the logs. Find out how long run history is kept (Make's free plan keeps 7 days) and whether you can replay a failed run with the same inputs.

Price one real workflow first

Plan pages are built for side-by-side comparison, which makes them a poor place to start (they're the dating-profile photo of software).

Pick one workflow you'd run anyway, such as a hypothetical invoice flow that reads a new invoice, has AI summarize it, updates your CRM, posts to Slack, and logs the result. Write down its steps, its monthly volume, which steps call AI, and who needs access.

Then price that one workflow on two or three platforms. On Make, five actions at one credit each across 2,000 runs a month lands at 10,000 credits, Core's default allowance, as long as the AI step stays at one credit.

On n8n Cloud, the same 2,000 runs are 2,000 executions, which fits inside Starter's 2,500. I'd shortlist whichever platforms come out cheapest at your real volume while still giving you the approval point and the logs you need.

FAQ

What's the cheapest way to start with an AI workflow builder?

The cheapest start is a free plan with a usable allowance, such as Make's 1,000 credits a month or n8n's self-hosted Community edition. Zapier's free plan also costs nothing, but it stops at two-step Zaps and 100 tasks a month.

Do AI workflow builders support MCP?

Most of the big names do: Zapier, Make, n8n, Gumloop, Workato, and Lindy all list MCP support as of October 2026. Kissflow's AI page doesn't mention MCP, so ask its sales team before you count on it.

Which AI workflow builders suit regulated industries?

Start with platforms that publish their controls and data locations, such as Kissflow's US, EU, India, or Australia regions and Lindy's Enterprise plan with a signed HIPAA BAA. If data can't leave your infrastructure, n8n's self-hosted editions keep the whole platform on servers you run.