Question 1

What does an AI automation consultant actually do for a small business?

Accepted Answer

I take a process that eats hours every week — lead routing, proposal drafting, support triage, invoice chasing — and rebuild it as an n8n workflow with a Claude or Gemini step where judgment is needed. You get a working automation in your stack, documentation, and a handover so your team can edit it. No managed-service lock-in unless you want it.

Question 2

n8n or Make or Zapier — which one should I pick?

Accepted Answer

Zapier is easiest for one-off two-step automations and the most expensive at scale. Make is cheaper and visually richer but gets clunky past ~20 modules. n8n is self-hostable, has native AI nodes, and is roughly 5-10x cheaper at volume — which is why I default to it for anything mission-critical. If you already live in Zapier and the workflow is small, I'll tell you to stay there.

Question 3

How long does a typical n8n automation project take?

Accepted Answer

Most single-workflow builds ship in 1-2 weeks. A multi-workflow system with a RAG knowledge base, a Claude agent, and CRM sync is usually 3-6 weeks. I scope flat-fee, not hourly — so the timeline is fixed before we start. If discovery reveals it's bigger than that, I'll tell you before invoicing.

Question 4

When should I use an AI agent versus a simple AI workflow?

Accepted Answer

Most of the time — use a workflow. A deterministic n8n flow with one or two LLM steps is cheaper, more reliable, and easier to debug than a fully autonomous agent. Agents make sense when the path through the work genuinely changes per input — research tasks, multi-tool browsing, complex triage. If a vendor is selling you an agent for something a 5-node workflow handles, you're paying for buzzwords.

Question 5

What's a RAG system and do I actually need one?

Accepted Answer

RAG (Retrieval-Augmented Generation) means the AI answers questions using your documents, not its training data. You need it if you want a chatbot that knows your product docs, an internal assistant that searches contracts, or a sales tool that quotes your own case studies. You don't need it if your use case is generic writing or analysis — vanilla Claude or Gemini handles that fine.

Question 6

How much does this cost?

Accepted Answer

Flat-fee, project-based. A focused n8n workflow starts around 400 EUR. A Claude-powered system with RAG and multiple integrations is typically 1,000-4,000 EUR depending on scope. I do hourly retainers too, if needed. Training engagements (1:1 or team) are priced separately via direct quote.

Question 7

Do you offer team training on n8n and Claude, or just delivery?

Accepted Answer

Both. I run 1:1 mentoring through MentorCruise for solo PMs, founders, and engineers learning to ship AI features. For teams I do paid workshops — usually a half-day or full-day on n8n fundamentals, Claude workflow patterns, or when to build agents vs. workflows. Trained teams cost less to support afterward, which is why I push for this.

Question 8

Are you a developer? Can you build a full SaaS app?

Accepted Answer

No — I'm not a web developer and I won't pretend to be. I build automations, AI workflows, and RAG systems on no-code and low-code platforms (n8n, Lovable, Claude, OpenAI APIs). If you need a custom React app with a backend, I'm the wrong person. If you need the automation layer behind your existing app or SaaS, I'm exactly the right one.

Question 9

How do you handle data privacy and prompt injection when using Claude or GPT-5.5?

Accepted Answer

I treat any external content (emails, scraped pages, user uploads) as untrusted — it never gets executed as instructions. For self-hosted n8n I run on your infra or mine, your choice. For client data in LLM calls I use Anthropic and OpenAI's zero-retention/no-training settings where available. If your industry needs more (healthcare, legal, regulated finance), say so upfront — some use cases are out of my lane.

Question 10

What kind of clients do you usually work with?

Accepted Answer

SMBs, agencies, and solo operators. Founders who need their first AI feature shipped. Ops leads drowning in manual work. Agencies that want to white-label automation for their clients. I don't work well with enterprises that need 6-month procurement cycles — the speed advantage of small teams is the whole point.

Ship AI automations without hiring a team.

Core Expertise

n8n Automations

AI Agents & RAG

Claude Workflows

About Gene

What I actually do

n8n & workflow automation

Claude & AI agent builds

Where I've shipped

What can I build for your business?

n8n Workflow Engineering

AI Agents & RAG Systems

Claude Workflow Architecture

Frequently asked questions

Get in Touch

Contact Information

Email

LinkedIn

Upwork