I build chatbots, RAG knowledge bots, and AI agents for the specific parts of your business that need to read intent and make a decision — not as a buzzword layered on top of automation that didn't need it.
When do I actually need AI instead of regular automation? AI is only needed when a step requires real judgment, like reading an open-ended message and understanding intent. If a process is genuinely rule-based, plain automation is cheaper and more reliable than adding AI unnecessarily.
Most "automate this" requests I get don't actually need AI — they need a Set/Edit Fields node and a Switch statement.
✓ No AI needed for: moving data between apps, sending scheduled reminders, updating a CRM field, generating a report on a schedule. These are rule-based and AI would only add cost and fragility.
AI genuinely helps with: reading an open-ended customer message and understanding what they want, answering questions from your actual documents, classifying intent before routing, drafting a first response for a human to review.
An AI agent isn't a standalone chatbot — it's the brain sitting in the middle of your actual tools, reading, deciding, and acting across all of them.
A business wanted to handle customer conversations on WhatsApp without a person manually replying to every message, and needed a way to re-engage customers who went quiet mid-conversation.
Built a WhatsApp Business API chatbot handling the core conversation flow, plus an automated follow-up sequence that re-engages customers who stop responding.
A short conversation about which part of your process actually needs AI, not just automation.
What it reads, how it decides, where the human review gate sits.
Prompt design, integrations, testing against real edge cases.
Handling the judgment calls, routing what it shouldn't handle to a human.
AI automation isn't a chatbot bolted onto your website — it's an AI step inserted into an otherwise ordinary automation workflow, exactly where a judgment call is needed. I build these using n8n connected to OpenAI, Claude, or other models, scoped narrowly so the AI handles interpretation while rule-based logic handles everything else reliably. See when you actually need AI versus simple automation, and how this applies to lead qualification specifically.
Automation and AI agents aren't a one-time build. APIs change, tools update, workflows fail, and agents need refinement as your business changes.
For businesses that want ongoing technical support after launch, I offer:
Need someone to keep your automation running? Get a Free Automation Audit →
Tell me what you're trying to solve. I'll give you an honest answer — even if it's simpler than you expected.
Get a Free Automation Audit →I built ProofMyAI, a SaaS that checks every chatbot answer against your docs and alerts you on made-up answers, missed escalations and AI agent loops. Free plan, no card needed.
See ProofMyAI →Want a team of AI agents like this?
I build automations that qualify leads, update your CRM and send replies while you sleep. Get a free automation audit →