Strategy and a realistic plan
Which parts of the business are worth touching, what to do first, what it should cost, and how you'll know it worked.
Consulting · From someone who builds
advice · free
where most advice stops
build · fixed price
Most consulting ends at a recommendation. Mine continues: I scope the work, write the code, and hand it over. That shapes the advice itself — suggesting something I couldn't build would only embarrass me.
Getting AI wrong is expensive. But the costly mistake is rarely the technology — it's spending months on the wrong thing, or buying tools that don't fit how your team actually works. What's missing is usually not more AI. It's a clear read on which parts of the business are worth touching at all, and what should happen first.
You don't pay for advice. The first call, the workflow audit and the recommendation are free — whether or not we end up working together. If it's worth building, you get a written scope and a fixed price before anything starts. I get paid for the build, not for telling you what to build.
Which parts of the business are worth touching, what to do first, what it should cost, and how you'll know it worked.
Map how the work actually flows today, find where automation pays off first, and separate the quick wins from the bigger plays. Where the answer is to build something, that becomes the scope — not a document you take to someone else.
When data genuinely cannot leave your environment. I'll say plainly what local models can and can't do compared with the big ones — it's a real trade-off, not a sales pitch.
The landscape is loud. I evaluate what actually fits your workflows and your budget, and I'll tell you when a subscription tool is enough and custom work is not worth it.
If the answer is an agent, that's its own page: what an agent is, how I build one, and what it costs.
See the AI Agent page →A first conversation to something running — with no bill for the thinking.
Your business, your tools, where it hurts. No pitch, no commitment; an honest read on whether this is worth doing at all.
How the work really flows today, where the time goes, which steps are automation candidates and which need judgement.
In plain language: what to do, what to skip, what to buy instead of build. If the answer is “you don't need AI right now,” that's the answer.
Before anything starts. If the work grows later, we agree that separately instead of watching a meter run.
The code in your repository, documented, with a walkthrough so your team can maintain it without me.
Repetitive work eats the team's day — Data entry, scheduling, routine inquiries. If predictable patterns take hours, they're automation candidates.
Response times are slipping — Support or sales is becoming the bottleneck, and good enquiries wait their turn.
Documents get retyped by hand — Invoices, forms, contracts. Anything where information is extracted and re-entered is automation-ready.
Competitors are moving — Others in your space are shipping AI-assisted workflows. That gap compounds.
Your data can't go to someone else's cloud — Some businesses genuinely can't send sensitive data out. Local models deserve a serious look — with eyes open about the trade-off.
What the work actually looks like: the four shapes I see most often — enquiry triage, document extraction, internal knowledge, monitoring — each with the failure mode I design around. They're on the AI Agent page.
See the AI Agent page →When you use AI through a web interface, the conversation goes through someone else's servers. For most business data, with a sensible provider plan, that's fine. For data that can't leave — regulated, contractual, or just yours — local models keep everything in-house. That's a real trade-off: local models are cheaper to run and nothing leaves, but they genuinely don't match the frontier ones. Which way to go depends on the task, and I'll tell you my read directly.
Clients across time zones; the work moves by milestones, not by meetings.
Main working hours 06:00–21:00 (UTC+8); issues don't wait overnight.
No account manager relaying requirements, no handoff after the strategy is signed. The person who scoped it is the person who builds it.
Most consultants will tell you the answer is always consulting. It genuinely is not. Sometimes the useful thing is to skip the talking and build the one thing you already know you need.
Yes. Discovery, audit and recommendation cost nothing, and that stays true even if we never work together. I get paid when something gets built.
If it's not worth building, I say so — recommending work just to win it is a short-term win that shows up later as a bad reference.
One person, no deck, no handoff: the one auditing the workflow is the one who'd build it. If your project truly needs a big team and 24/7 SLAs, I'll say so and point you to a firm.
A plain-language read on what to do, skip, or buy — plus, if it's worth building, a written scope and a fixed price. No slide deck to decode.
The tools you use today and the one process that annoys you most. Half an hour is enough.
Tell me what's bugging you. If it's worth doing, you'll know what to do and what it costs. If it isn't, you'll know that too.
Book a free call →