The problem

Every conference slide says "AI will transform your industry." Your LinkedIn feed is full of people claiming 10x productivity gains with ChatGPT. Meanwhile, your team copies and pastes ChatGPT outputs into emails and calls it a strategy.

Or worse, nobody touches it at all because no one knows where to start. You've heard about automation, internal chatbots, AI-powered analytics. Sounds great in theory. In practice, you're stuck between hype and paralysis. Your competitors claim they're "AI-powered" and you wonder if you're falling behind. Here's the thing: most of them are bluffing too.

Not a technology gap. A clarity gap. You don't need more AI talk. You need someone who builds AI integrations daily to tell you what's worth doing and what's a waste of money.

What I do

AI audit and maturity assessment

Before plugging in any API, I look at how your company actually works. Your processes, your tools, your bottlenecks, the tasks your team quietly hates doing every week. Most AI projects fail because they solve the wrong problem. I identify the high-ROI use cases first. And I'll tell you straight if a use case isn't worth the investment. Some things are better solved with a spreadsheet formula. Knowing the difference is what an AI consultant is actually for.

LLM and API integration

This is where it gets concrete. I connect language models (OpenAI, Claude) to your existing tools and workflows. Internal chatbots trained on your documentation. Automated content generation pipelines. RAG systems that let AI answer questions using your own data, not generic internet knowledge. At Google (through Teleperformance), I built automated reporting workflows that cut 10 hours of manual work per week. Same logic applies to AI: find the repetitive cognitive task, build the pipeline, remove the bottleneck.

Training and adoption

A tool nobody uses is a tool that doesn't exist. I train your teams on prompt engineering, responsible AI use, and the actual limitations of these models (because yes, they hallucinate, and your team needs to know when to trust the output and when not to). The goal is simple: when I leave, your team keeps going without me. Not dependency. Autonomy.

How it works

1. Discovery call

30 minutes on a video call. You tell me what you're trying to solve. I ask questions, poke at assumptions, and give you an honest first read. If AI isn't the right answer for your problem, I'll say so. Free, no commitment.

2. AI maturity audit

I map your current tools, workflows, and team capabilities against what's actually possible with AI today. You get a clear document: here are the opportunities ranked by impact and effort, here's what to ignore, here's where to start.

3. POC then deployment

I build a working proof of concept on the highest-value use case. You test it with real data, real users. If it works, we deploy. If it doesn't, we learned something valuable before spending the full budget. No 6-month project plans that deliver a PDF at the end.

4. Training and autonomy

Every integration comes with documentation and hands-on training for your team. I walk them through how it works, how to adjust prompts, how to spot errors. The project is done when your people can run it themselves.

Tech stack

OpenAI APIClaude APILangChainRAGPythonn8nGoogle Apps Script

Frequently asked questions

Is AI relevant for my company size?

Size doesn't determine AI readiness. The use case does. A 15-person company with a clear, repetitive process to automate will get more value from AI than a 500-person company chasing the buzzword without a defined problem. I work with SMEs, mid-size companies, and startups. The common thread is always the same: a specific pain point, not a vague desire to "do AI."

How much does an AI project cost?

I don't quote before understanding the problem. The discovery call is free, and after the audit I'll give you a clear scope and price. Some projects are a few days of work. Others take a few weeks. I bill by day or by project, depending on what makes sense. And if the honest answer is "this isn't worth the investment right now," I'll tell you that too.

Is my data kept confidential?

Yes. I work exclusively with private API instances. Your data is never sent to third-party services without your explicit agreement. No training data leaks, no shared model endpoints. We define the data handling rules together before anything starts. Confidentiality isn't a feature I sell. It's a baseline.