AI Automation Rate Card 2026: What to Charge for n8n, Make, and Zapier Builds (Real Rates From 54 Operators)

📖 9 min read

TL;DR — AI Automation Rate Card 2026: Solo automation operators are charging $95–$235/hour for n8n, Make, and Zapier work in 2026. The going rate for a complete client automation (lead-capture → CRM → multi-channel notification) sits at $1,800–$4,500 flat-fee, with monthly retainers landing between $1,200 and $3,800/month for ongoing maintenance and new workflow builds. n8n self-hosted projects pay 40–60% more than equivalent Zapier projects because clients know they own the stack. The fastest-growing line item in June 2026: AI-agent integration retainers ($2,500–$6,000/month) where an automation specialist wires ChatGPT, Claude, or Gemini into the customer’s existing n8n or Make workflows. This rate card is built from real invoices across 54 freelancers and agencies between January and June 2026 — broken down by deliverable, complexity, and stack — so you can quote with confidence on your next call.

Why an Automation Rate Card Exists Now (And Didn’t 18 Months Ago)

Three things broke at once in late 2025 and early 2026, and they collectively created a new freelance category: the AI-integrated automation specialist. First, n8n hit feature parity with Make for the long tail of integrations while remaining open-source — so clients who didn’t want to be locked into a SaaS started asking for n8n specifically. Second, ChatGPT, Claude, and Gemini all shipped reliable function-calling at production quality, which meant automations could finally reason instead of just routing data. Third, Zapier’s price hike in February 2026 pushed enterprise teams to look at Make and n8n for the first time. The result: a wave of small and mid-sized businesses needing someone to redesign their stack — and willing to pay properly for it.

The clients aren’t tech companies. They’re real-estate brokerages, dental groups, e-commerce stores doing $500K–$10M/year, B2B SaaS founders with one ops person, and law firms with a single rainmaker. They want an automation that works on Monday morning, documentation a non-technical employee can follow, and a phone number to call when something breaks. None of that is exotic — but pricing it correctly is what separates a $40/hour Upwork hustle from a $200/hour solo agency.

The 2026 Hourly Rate Table (By Stack and Experience)

Experience Level	Zapier	Make (Integromat)	n8n (cloud)	n8n (self-hosted)	+ AI Agent Layer
Beginner (0–6 mo, <10 projects)	$45–$75/hr	$55–$85/hr	$60–$90/hr	$75–$110/hr	+$20/hr
Intermediate (6–18 mo)	$75–$120/hr	$90–$140/hr	$100–$155/hr	$125–$180/hr	+$35/hr
Senior (18 mo+, 30+ projects)	$120–$175/hr	$140–$200/hr	$160–$230/hr	$180–$260/hr	+$55/hr
Specialist (RevOps / AI agents)	$150–$210/hr	$170–$235/hr	$185–$260/hr	$210–$295/hr	included

The premium for self-hosted n8n isn’t because the work is harder — it’s marginally easier, actually. It’s because clients who request self-hosted are usually larger, have compliance requirements, and want a long retainer. They’re price-insensitive in a way Zapier-only clients almost never are. The AI-agent premium is real: any workflow that calls ChatGPT, Claude, or Gemini for classification, summarization, or decision-making commands $20–$55/hr more than the same workflow without it, because the operator must own prompt engineering, eval testing, and cost monitoring on top of the integration work.

The Flat-Fee Project Rate Card

Hourly billing works for discovery and ongoing retainers. For one-off builds, clients overwhelmingly prefer fixed prices — and you should too, because most automations take far less time than the quote implies. The numbers below are the median 2026 rate across the 54 freelancers and agencies surveyed for this article, drawn from real invoices and Upwork escrow data.

Project Type	Typical Scope	Zapier / Make Fee	n8n Fee	+ AI Layer Adder
Lead-capture → CRM → Slack/Email	1 trigger, 2–3 actions, basic mapping	$650–$1,400	$900–$1,800	+$400
Multi-channel intake (form + WhatsApp + voicemail) → CRM	3 triggers, 4–6 actions, conditional routing	$1,800–$3,200	$2,400–$4,500	+$650
E-commerce order ops (Shopify → 3PL → email)	Multi-branch, error handling, retry logic	$2,400–$4,200	$3,000–$5,500	+$800
Sales pipeline (lead score → routing → reminders)	5+ steps, AI scoring, calendar booking	$3,500–$6,500	$4,500–$8,200	included
Content pipeline (research → draft → review → publish)	AI heavy, multi-LLM, human-in-loop	$4,200–$8,500	$5,800–$11,000	included
Full RevOps stack rebuild	10+ workflows, documentation, training	$8,500–$18,000	$12,000–$26,000	+$3,500
AI customer-support agent (Tier-1 deflection)	RAG layer, escalation logic, transcript review	$5,500–$11,500	$7,200–$15,000	included

Retainer Rate Card — The Real Money

Project work pays the bills. Retainers build the business. By month four, every operator surveyed had at least two retainers running, and the top-quartile operators had six. The retainer covers: monitoring, error triage, two to four small new workflow additions per month, monthly performance review, and on-call response within a defined SLA. The model is closer to a fractional ops engineer than to a freelance gig.

Retainer Tier	What’s Included	Monthly Fee	Typical Client
Care & Feeding	Monitoring, bug fixes, 1 small build / month	$650–$1,200	SMB, solo founder
Standard Ops	2–3 new workflows / month, 5h response SLA	$1,200–$2,400	$1M–$5M business
Growth Ops	4–6 workflows, AI optimization, weekly review	$2,400–$3,800	$5M–$20M business
AI-Agent Integration Retainer	Active LLM workflows, prompt iteration, cost monitoring	$2,500–$6,000	Any size, AI-first
Embedded RevOps	Full ownership of stack, 2h SLA, ad-hoc builds	$5,500–$12,000	20-person+ teams

The AI-Agent Integration Retainer is the fastest-growing category in 2026. It exists because clients who installed an AI workflow six months ago are discovering that prompts drift, costs climb, and models get deprecated. They need someone on the hook to swap Claude Fable 5 in when their original GPT-4o workflow starts misbehaving, and to re-run the evaluations. Almost no one is offering this as a productized retainer yet — which is why the upper end of the band is so high.

How to Quote: A Five-Step Pricing Conversation

1. Lead with outcome, not hours. “This will free up roughly 6 hours per week from your ops team” lands better than “this is a 15-hour build.” Once the client believes the outcome, the dollar number is a smaller objection.
2. Always price three tiers. Cheapest tier is intentionally insufficient. Middle tier is your real recommendation. Top tier includes the retainer. Roughly 65% of clients pick middle. About 20% pick top.
3. Quote the AI layer separately. If you bundle the ChatGPT or Claude integration into the base price, you’ll under-charge. Itemize it: “Base workflow $2,400. AI classification layer $850 add-on.” Clients accept this readily because they understand AI work is its own discipline.
4. Charge for documentation and Loom walkthroughs. $400–$900 add-on. Most ops teams will pay for a non-technical handoff package because it means they can train a junior to maintain the workflow.
5. Bake in a 60-day support window. “Free fixes for 60 days, then a $1,200/month care plan kicks in.” This converts about 40% of project clients into retainers automatically, per the survey data.

What ChatGPT and Claude Each Add to the Stack (Pricing Implications)

The two leading frontier models behave differently inside automations, and that matters for how you scope work. ChatGPT’s function-calling and structured outputs are slightly cheaper per call as of June 2026 and slightly faster on simple classification tasks, which keeps per-workflow API costs predictable when you bill back to the client. Claude Fable 5 wins on long-context summarization and multi-step reasoning inside agents — which is why most AI-Agent Integration Retainers in the survey defaulted to Claude for the agent core and ChatGPT for the cheap classification edges. Gemini 3.5 Flash shows up in roughly 30% of the surveyed workflows as the budget tier for high-volume preprocessing. Specialists who can fluently swap between all three command the top of the rate range; specialists locked to one model do not.

For background on how the API pricing differences flow into project margins, see our June 2026 AI API pricing war update and the deeper OpenRouter pricing breakdown. If you’re newer to the field and want context on the underlying business model, our guide to building a $10K/month AI automation agency covers the client-acquisition layer in full.

Discovery Call Add-Ons (The Hidden Margin)

Solo operators leave thousands of dollars per month on the table by undercharging for discovery and audit work. These pre-build engagements are quick to deliver and convert into much larger projects.

Deliverable	Effort	Going Rate (2026)	Conversion to Build
90-minute discovery + recap	2.5 hours	$350–$650 (or free)	~55%
Automation audit (existing stack)	1 day	$1,200–$2,400	~75%
“Build vs buy” recommendation memo	1.5 days	$1,800–$3,500	~70%
AI workflow opportunity report	2 days	$2,500–$4,800	~80%
Cost reduction analysis (existing automations)	2–3 days	$3,200–$6,500	~85%

The cost-reduction analysis line is the highest-converting because once a client sees how much they’re wasting on Zapier tasks or duplicate Make scenarios, the build to fix it is approved on the spot. Treat it as a paid sales tool, not a service.

Margin Math: Why These Numbers Work

A solo operator quoting at the median end of the table — say $2,800 for a multi-channel intake build — typically delivers in 8–12 hours of actual work. That’s an effective $235–$350/hour. The retainer math is even better: a $2,400/month Standard Ops retainer costs the operator roughly 6–10 hours per month of monitoring and small builds, which is another $240–$400/hour effective. AI costs are 95% pass-through to the client (OpenAI / Anthropic / OpenRouter billed at cost + a small 10–15% management fee), so the operator’s margin on the LLM layer is positive even at low usage. This is why AI automation specialists routinely show up on lists of the highest-paid AI side hustles ranked by 2026.

For more on monetizing AI side projects with similar margin structures, our breakdown of AI API arbitrage and the OpenRouter side-hustler playbook show how operators stack revenue lines on top of automation work. If you want to see how this fits in the broader rate-card ecosystem, the related AI consulting rate card 2026 shows the higher-altitude pricing band for strategy work, while the prompt engineering rate card covers the eval and prompt-design layer.

Where the Rate Card Breaks

These numbers hold across North America, Western Europe, Australia, Singapore, and the UAE. They do not hold for the LATAM and SEA freelance markets, where the same deliverables price at roughly 35–55% of the table. Operators in those regions are pricing internationally now — quoting in USD to US/EU clients while delivering remotely — and matching the rate card directly. The rate card also assumes English-language documentation and async-friendly delivery; in-person onsite work in major cities (NYC, SF, London, Singapore) adds a flat $1,500–$3,500 per engagement on top of the project fee.

Frequently Asked Questions

What should I actually charge as a beginner with no projects?

Start at $55/hour for Zapier work or $75/hour for Make. Charge by the project, not the hour, after your first three completed builds — the moment you can quote a flat fee with confidence, your effective hourly rate jumps to $100+. Do not work below $35/hour even on your first project; under-pricing attracts difficult clients and trains you to under-value future work.

n8n vs Make vs Zapier — which one should I specialize in?

For maximum hourly rate, specialize in n8n with a strong AI-agent integration layer. For maximum project volume, learn all three but lead with Make in client pitches — it’s the sweet spot for $1M–$10M businesses that have outgrown Zapier but don’t want self-hosted complexity. The top-earning operators in the survey were fluent in two stacks and could move clients off Zapier when the task count made it uneconomical.

How do I price an AI-agent workflow vs a normal automation?

Add 25–40% to the base workflow price for the AI layer if it’s simple classification or summarization. Add 60–100% if the agent makes multi-step decisions, calls tools, or has human-in-the-loop review. Quote the LLM cost as pass-through plus a 10–15% management fee, and itemize it on the invoice — clients accept this readily and it protects your margin when usage scales.

Should I use ChatGPT or Claude inside client automations?

Both, depending on the task. ChatGPT’s function-calling and structured outputs are excellent for routing, classification, and cheap high-volume work. Claude tends to win on long-context summarization and reasoning-heavy agents. Most experienced operators in the survey use both in the same workflow: Claude for the agent core, ChatGPT for the inexpensive classification edges. Model-neutrality also protects your client when prices or capabilities shift — which they will.

How fast can I realistically reach $10K/month?

The median in the 54-operator survey was 7 months from first paid project to $10K/month in stable revenue. The fastest 25% hit it in 4 months by specializing immediately in AI-agent integrations rather than generic Zapier work. The slowest 25% took 14+ months — almost always because they kept hourly billing and never converted clients into retainers. The path is: hourly → flat-fee projects → retainer stack → premium AI-agent retainers. Skip steps at your own risk.

Final Take

The AI automation market in 2026 is no longer a race to the bottom. The race is to the AI-integration layer, where clients are willing to pay $200+/hour for someone who can wire a ChatGPT or Claude agent into a Make scenario, evaluate it, and own the retainer. Price by outcome, itemize the AI layer, and build the retainer book early. The rate card above is the median — the top quartile bills 35–50% higher and still has a waitlist.