AI consulting for startup and SaaS founders
Vlad Tudor is an AI consultant and engineer in Bucharest who helps startup and SaaS founders decide what to build with AI, what to buy and what to leave out, before money is spent. He founded the AI studio Sapio AI in 2021 and has built speech, retrieval and customer-support systems for startups since before ChatGPT.
Reply within 24 hours · no sales pitch
What I hear from founders.
“I've asked ChatGPT and it thinks the idea is great. I'd like a second opinion from someone who isn't a language model.”
“Everything I want to build, a bigger company can ship next quarter. What's actually defensible?”
“The agency quoted a platform. I only wanted to know if the core idea works.”
Why a second opinion, and why early.
Ask a chatbot whether your business will succeed and it will almost always say yes. That isn't malice; it answers in the likeness of the question, and the models are tuned towards the answer the user prefers. It will never tug at your sleeve the way a person would and say: stop, you're on the wrong track.
That is the job I do in the first conversation. It is cheaper to hear it before the first line of code than after the first round of hiring.
From the REWIRED podcast.
What I'd do with you.
- 01
Test the core assumption first.
The one thing the product depends on, tested on real data in days rather than built into an MVP for months. Sometimes it holds. Sometimes the idea changes, which is cheaper now.
- 02
Build versus call an API.
Which part should be a call to an existing model, which part needs your own data or model, and which part is ordinary software that doesn't need AI at all.
- 03
Find the part only you can build.
Your data, your domain or your workflow. That's usually where defensibility lives, rarely in the prompt.
- 04
A technical story that survives due diligence.
An architecture and a plan an investor's technical adviser can question without it falling apart.
- 05
Getting your developers started with AI.
Not by mandate. Let them start on a side project they've postponed, away from the fragile main code base, until they know how to steer it.
What I've built.
A Romanian speech-to-text model
For a voice startup, Sapio trained a Romanian speech-to-text model from about 5,000 hours of recordings. At delivery it had the best word-error rate on the market. The team received the model, the full training pipeline and a guide, kept improving it themselves, and it became the centre of their technology offering and helped them raise their next round.
Support, rebuilt around retrieval
For a startup serving retail brands, I worked on a customer-support system that was redesigned around retrieval in under four weeks after ChatGPT's launch, becoming one of the first production services answering with generated text that cites its source.
Advisory roles
I advise dConnection, an Austrian personal-AI-assistant startup, and act as technical coordinator at CogniSync, a Romanian legaltech startup.
Only 25% of AI initiatives have delivered the ROI expected of them, and only 16% have scaled across the business.
Tell me the one assumption your product can't survive without. We'll start by testing it.
A few short questions. The first conversation is free.
Reply within 24 hours · no sales pitch
What an engagement can cover.
A first conversation
About the problem, before any technology. Free.
An audit of processes, data and tools
How the work runs today, what data exists, which tools are already paid for.
A prioritised roadmap
What to build, what to buy and what to skip, in order, with the reasoning written down.
The build, or a clean hand-off
Hands-on build through Sapio AI, or a roadmap your own team can take from there.
Team enablement
A workshop for the people who will use the result, so it gets used.
Follow-up
A check on what changed once the work is in place, and what comes next.
For founders: the audit becomes a technical due diligence of the idea or the existing prototype; the roadmap separates what to build from what to buy via API; the hand-off is usually to your own developers.
Related pages.
AI consulting for companies in Romania
How an engagement runs, step by step, and what it can cover.
AI process automation for operations teams
Manufacturing, energy, distribution and back office: fix the process, then automate it.
AI for retail, e-commerce and customer support
What AI should answer, what stays with a person, and when a ready-made widget is enough.
Questions that come up.
Should my startup build its own model?
Usually not at the start. Most products begin on an existing model called through an API, with your own data and workflow around it. Your own model makes sense when the data is yours, the task is narrow, and the API is too expensive, too slow or not allowed to see the data.
Can you act as our technical co-founder or CTO?
Not as a co-founder. I can advise and review the technical decisions, and when something needs building, Sapio AI can build it or hand it to your own developers.
Do you sign NDAs?
Yes, before you share anything that needs one.
Can you look at a prototype someone else built?
Yes. Often that's the most useful review: whether it works on real data, what it will cost to run, and what breaks when usage grows.
Is it too early to talk if I only have an idea?
No. That's the cheapest moment to change it.
Tell me what you're working on.
A few short questions, two or three minutes. The reply comes from me, within 24 hours on working days.
Reply within 24 hours · no sales pitch
Help with
What would you like help with?
Pick all that apply.