Skip to content
AETHERAI
Aether AI · Mats Ponte

You know your business.I know AI.

Together we look at where your business gets stuck, and whether a solution can be built for it. You bring the knowledge of your business, I bring the AI expertise. Without the hype.

Process automation, knowledge bases, websites and MVPs for SMEs. In days to weeks, not months.

Available for new projects

In the cloud or entirely on your own servers.

Independent AI specialistThe Netherlands

You do not have to take my word for it.

This site

Designed and built myself

What you are looking at is my work: design, code and infrastructure. No template, no agency.

Five working demos

Technology you operate yourself

Knowledge base, invoice extraction, quotes, lead triage: genuinely working AI to try out, not a screenshot.

See the work

n8n · Next.js · React · Supabase · PostgreSQL · Python · Pinecone · OpenAI · Anthropic · Resend · TypeScript · Tailwind

01 · In your business

Every business owner knows it: the day is full of work that is not running the business. Email, handovers, quotes, the same questions. That is working in your business, while the gain is in working on it.

Extra staff does not solve that, and neither does working longer hours. It is in the set-up: processes that look right on paper but keep landing on your desk in practice. Knowledge that only lives in your head. Follow-up that waits until you have time.

That is exactly where the value is, and nobody sees those spots more sharply than you. What is usually missing is someone who can judge which of them AI can solve. That is what I bring.

02 · Together

The best analysis is one we make together.

What you bring

  • How your business really runs
  • The places where hours leak away
  • The judgement whether a solution holds up in practice

What I bring

  • AI expertise, from university to shop floor
  • The building capacity to actually make it
  • The option to run everything on your own servers
  • An honest answer, even when that answer is 'no'

That overlap is where the value lies.

03 · Working principle
I will also say so when AI is not the answer.

No sales call: an honest answer, even if it is 'no'.

04 · Why different

No agency, no packages. One craftsman.

One point of contact

You talk to the person who also designs and builds it. No account manager between you and the technology, no noise in the handover.

No packages

I do not sell a subscription or a standard solution. What I build is what your situation asks for, not what happens to be on the shelf.

Honest, even when it is 'no'

If there is no opportunity, I say so. No sales funnel, no pressure: an honest judgement is worth more than a solution sold.

Science to shop floor

AI knowledge from the university, translated into what your business can use tomorrow. And, if you prefer, entirely on your own servers.

05 · Where I build

If there is an opportunity, I build the solution.

Click a row for detail

Replacing manual business processes with AI-driven workflows. Lead follow-up that runs itself, quotes generated automatically, knowledge bases that answer questions.

What I build in practice

  • Incoming leads that are automatically qualified and followed up
  • Quotes generated from an email or a form
  • Documents that are automatically summarised, routed and archived
  • Knowledge bases that answer internal questions via chat or email

Tooling

n8n · Python · OpenAI · Anthropic · REST/Webhooks · PostgreSQL

Lead time: one to three weeks per workflow.

What that looks like

Manual step by step → one trigger that does it all. (own animation)
Leads that follow themselves up instead of piling up. (own animation)

Knowledge bases, RAG systems, vector databases: the foundation AI needs to be usable inside an organisation.

What I build in practice

  • Company knowledge structured in a searchable RAG layer
  • Vector databases (Pinecone / pgvector) fed with your documents
  • Secure, authenticated access per user role
  • Monitoring of quality, hallucinations and cost

Tooling

PostgreSQL · Supabase · Pinecone · pgvector · Python

Lead time: one to four weeks, depending on the sources.

Commercial websites and MVPs: fast, conversion-driven, connected to your CRM. No months-long projects.

What I build in practice

  • Marketing sites with measurable conversion and clear CTAs
  • MVPs for ideas that still need validating
  • Customer portals, dashboards, intake applications
  • Integration with CRM, email, payments, AI layer

Tooling

Next.js · React · TypeScript · Tailwind · Supabase · Resend

Lead time: one to three weeks for an MVP.

Safe, privacy-compliant AI roll-out without hallucinations. GDPR-aware, with monitoring, evaluation and audit.

What I build in practice

  • GDPR-compliant deployment strategy on your infrastructure
  • Evaluation sets and monitoring to prove quality
  • Audit trail of AI decisions for regulators
  • Guidance on choosing open/closed models and hosting

Tooling

Compliance · Monitoring · Evals · AVG · On-prem / EU

Lead time: depends on scope, a conversation first.

Rather keep everything out of the cloud? Open AI models running on your own servers, and fully custom-built systems for maximum control over data and behaviour. An option we can discuss together.

What I build in practice

  • Open AI models running on your own servers, with no data going to third parties
  • A fully custom-built system around your specific process
  • A private knowledge base or assistant that never leaves the company
  • Hybrid setup: sensitive work locally, the rest in the cloud

Tooling

On-premise · Open-weight modellen · EU / eigen servers · Maatwerk

Lead time: depends on scope, a conversation first.

06 · What that looks like

Three set-ups, one pattern: the work keeps going. Without you.

Illustrative set-ups. Your situation determines the final shape. No standard packages.

Set-up 01

The quote that writes itself

Today

Every request means copying from old quotes, looking up prices, fixing the layout, and doing it all over again next time.

  1. Request comes in
  2. AI reads & structures
  3. Draft in your house style
  4. You approve
  5. Customer receives

With this set-up

You only approve. The rest is ready, in your house style, with your prices.

n8n · your CRM · email

Set-up 02

The knowledge base that answers itself

Today

The same questions come back every week, and the answer is in your head, or somewhere in ten documents.

  1. Question is asked
  2. Searches your documents
  3. Checks the source
  4. Answer, with reference

With this set-up

Staff and customers get an immediate answer from your own documentation, with a source reference.

RAG · vector database · your documents

Set-up 03

The lead that never sits idle

Today

A request on Friday afternoon waits until Tuesday, and by then has become someone else's customer.

  1. Lead comes in
  2. AI assesses & enriches
  3. Personal reply
  4. In your CRM, followed up

With this set-up

Every lead gets an immediate, substantive, personal reply, and you see everything back in your CRM.

n8n · AI agent · your CRM

No promises, proof

Try the working demos
07 · How it works

No sales call. An honest answer.

even if it is 'no'

You email

Describe where it gets stuck. One paragraph is enough. You know your business best.

I read

I look at it through an AI lens. Honestly, even if there is no opportunity in it.

We look together

If I see an opportunity, we discuss together whether it is worth pursuing.

Free and without obligation. One paragraph about where it gets stuck is enough; you get an honest judgement on whether there is an AI opportunity in it, even if that answer is no.

Send your challenge
08 · Honest answers

What you are probably wondering.

Q1Do I need to be technical myself?

No. You know your business, I handle the technology. Everything I build is documented in plain language, and you get one point of contact: me.

Q2Is my data safe?

Solutions run on your own infrastructure where possible, and I work GDPR-aware: no customer data goes to third parties without agreement, and it is clearly recorded what goes where. Safe, privacy-compliant deployment is one of my four core areas.

Q3Can AI also run entirely on my own servers?

Yes. Alongside cloud models, I can run open AI models entirely on your own infrastructure, so sensitive data never leaves the company. That can be fully local or a hybrid setup: sensitive work on-premise, the rest in the cloud. We look at what fits best together, it is always open for discussion.

Q4What does AI automation cost for SMEs?

That depends on the solution, but we make a clear agreement about it up front, so you know where you stand before I build. No agency rates, no billable-hour surprises. And a first email with an honest answer costs nothing.

Q5What if AI is not the solution?

Then I say so. That is the working principle of Aether AI: rather an honest 'no' than a project that delivers you nothing. You are then one email further along, with no budget spent.

Q6How long does a project take?

Days to weeks, not months. I work solo with AI tooling as leverage, so there is no agency overhead between your question and a working solution.

Q7Will I stay dependent on Aether?

No, no vendor lock-in. It runs on your infrastructure, I document what I build, and you can always take it to someone else. That keeps me sharp.