Yako is a tool for organisations that live on public work: the tenders, grants, and consultancy calls that, because they involve public money, have to be published by law. It watches that public market and tells you which notices fit the work you actually do. It is built by Kitsuno, the same team behind the AI career agent, and it follows the same rules: it reads and explains, it never applies on your behalf, and your data stays yours.

TL;DR. Yako reads the public market — government tenders, grants, and consultancy calls that must be published by law — across more than 30 sources in 23 languages, every night, and shows an organisation or an individual the few notices that fit the work they actually do. It reads each notice against your own words, not keywords; it applies for nothing; and it sharpens every time you keep or dismiss a find. Early access: yako.kitsuno.ai. The longer story: The map has no legend for what you do.

The job looks simple

On the surface, finding public work is simple. An organisation that wants a public contract or a grant goes to the places where those are published, reads the notices, and pitches for the ones that fit. Bookmark a few portals. Set up an email alert. Ask the colleague who always seems to spot the good ones.

That is how almost everyone does it. It is simple. And then, as we found out building this, it is not.

Five questions behind “finding public work”

To see why, it helps to slow down and name what “finding public work” is actually made of.

We borrowed a lens for this from our own history. Years ago, in a report on children’s learning co-authored at the Zurich University of Teacher Education, we used a model from activity theory (adapted from Yrjö Engeström) that describes any purposeful activity with five plain questions: who is doing it, what they are doing, why, how, and where. The same five questions describe an organisation looking for public work:

  • Who — the organisation, and the people in it who know the work. A small firm, a research group, a charity, a loose network of independents.
  • What — the meaningful activity: pitching for and winning the public work that genuinely fits.
  • Why — the reason behind it: to keep doing the work the organisation exists to do, and to fund it.
  • How — the means of finding and judging opportunities. Today, almost always by hand.
  • Where — the place all this happens: the public market. Published by law, in principle open to everyone, spread across dozens of national and international sources in many languages.

Written out like that, one thing jumps out. The organisation is doing all five at once, by hand. It has to be the instrument itself (the how), cover an entire market (the where), and still pick the meaningful work (the what) out of an enormous amount of noise. No small organisation has the hours for that. So in practice it watches a slice of the market and hopes the rest is not where the good work was.

Why it isn’t simple

There is a harder problem underneath, and it is the one that surprised us.

The obvious fix is to automate the search: type in keywords, or let software rank notices by how similar they look to what you do. For a lot of organisations that half-works. For the ones whose whole value is how they work, designed with a community rather than delivered to it, participatory rather than off-the-shelf, it fails.

The reason is that the systems that route public money can say what is being bought. They cannot say how the work is meant to be done. The official codes have an entry for “training services.” They have no way to tell apart designing a programme with a community and renting a trainer for a day. So keyword search and similarity both reach for the surface of a notice and miss the thing that actually decides whether it fits.

The organisation always knew what kind of work it does. Every system between it and the public money had an opinion, and none of them was built to hold that answer.

What Yako changes

Yako’s job is to take that five-part model apart and put most of it back together on the organisation’s behalf.

Yako takes the how. It becomes the instrument. Every night it reads the market and reads each notice against the organisation’s own words.

Yako takes the where. As of August 2026 it watches more than 90,000 notices across more than 30 sources in 23 languages, so the organisation no longer has to choose which slice to follow. And the map is never finished. Yako is a cartographer: the public market has no single directory, so Yako is always finding new ground, checking it, and adding it. The territory it covers keeps growing.

Yako takes the what. Out of that whole pool, it surfaces the handful of notices that genuinely fit, each with a short, plain reason.

Then it asks the organisation to do one thing, and to do it well: declare. This is the who speaking for itself. In its own words, the organisation says what work it does and, just as importantly, what it does not. Yako reads every notice against that declaration, not against a guess about the sector it sits in.

And the why gets bigger. Once the whole market is watched instead of a corner of it, an organisation can pursue work it could never have found by hand. The mission it exists for has more room.

Two things stay firmly with the organisation. Yako applies for nothing: it surfaces and explains, and the organisation decides what to pitch for. And the organisation’s own judgment always wins. Every time someone keeps a notice, or dismisses it as not for us, that decision overrules the machine on that notice from then on and sharpens the next reading. The declaration sets the starting point; the keeps and dismissals tune it, night after night. Whoever holds the declaration, an organisation or an independent professional working on their own, is quietly sharpening their own results the whole time.

Yako’s own five questions

There is a neat symmetry here, and it is worth stating plainly, because it is how we think about what Yako is. Turn the same five questions on Yako itself:

  • Who — Yako: a reading system that watches the public market on an organisation’s behalf.
  • What — reading each notice against a declaration and returning a clear yes, no, or unsure, with a reason.
  • Why — so an organisation can be recognised for the work it actually does, and so public work reaches the organisations built for it.
  • How — rules and structure first, and a small AI model only where judgment is genuinely needed, always inside checks it cannot skip.
  • Where — the whole public market, every night, across sources and languages, never just a shortlist, and a map Yako keeps expanding as it finds new sources.

The organisation is the who of finding work. Yako is the who of reading it. The declaration is the hinge between them.

How Yako is different from a tender alert or database

If you already chase public work, you have probably tried one of a few things, and hit the edge of each.

The most common is a keyword alert on a national portal, or a large tender database. Both match the surface of a notice, so they either miss the calls that fit or bury you in ones that do not. A newer kind of tool goes to the other extreme and offers to draft the whole bid for you in one click, which takes the human out of the one step where judgment matters most.

The tools that come closest to Yako do read each notice and give a verdict, the way Yako does. But they stay in the easy corner to keep that manageable: one sector, usually IT, one or two countries, and public tenders only. That serves an IT firm bidding only on Dutch or German public tenders well. But as soon as your work reaches into more countries, or into grants and consultancies too, you are back to running a separate narrow tool for every corner.

Yako is built to do that reading across the whole, messy union: tenders, grants, and consultancy calls, across many countries and 23 languages, from dozens of sources rather than one clean feed. It reads each notice against your own words instead of a keyword list, keeps you in charge of what you pitch for, and gets sharper every time you keep or dismiss a find. Reading a narrow, uniform stream is the corner others have already solved. Reading the wide one is the hard problem, and it is the one Yako exists for.

Bring your own AI

One more choice sets Yako apart, including from Kitsuno itself. Kitsuno has a built-in assistant, called Kitso, that explains its thinking to the person using it. Yako does not put an assistant inside the product. Instead it opens your own Yako data, your declaration, your saved searches, and the decision and reason on every notice, to the AI assistant you already use, over an open standard called the Model Context Protocol (MCP). The matching stays rule-based and checkable. Your assistant reads the results; it does not get to invent them.

Where Yako is today

Yako is early. Two organisations use it today, and their declarations have been read against tens of thousands of notices. It is honest about its limits: it can only find what publishers actually publish, and the thinnest notices, a title and a deadline and nothing else, are still hard. There is a list to join at yako.kitsuno.ai. A list, not a launch.

If the five-question lens feels familiar from the rest of what we build, that is not an accident. We first used it for a child’s learning, then for a person’s career in Kitsuno, and now for an organisation’s public work in Yako. The subject changes. The conviction underneath does not: the person, or the organisation, doing the work knows best who they are, and the tools around them should be built to recognise that, not to flatten it.

Frequently asked questions

What is Yako?

Yako is a tool from Kitsuno that helps organisations and independent professionals find public work: any opportunity that has to be published by law. It watches government tenders, grant calls, and consultancy opportunities across more than 30 sources in 23 languages, and shows you the notices that match the work you actually do. It reads and explains, never applies on your behalf, and keeps your data yours.

What kinds of opportunities does Yako cover?

Any work that has to be published by law. Because it involves public or publicly funded money, it is put out in the open: government and public-sector tenders, grant and funding calls, consultancy and framework opportunities. Sources range from national procurement portals across Europe to international bodies such as the EU institutions, the World Bank, and UN agencies. As of August 2026, Yako tracks over 90,000 notices, about 14,000 of them open at any one time.

How is Yako different from a keyword tender alert?

Keyword alerts and similarity search match the surface of a notice. Yako reads each notice against your own written declaration, in your own words, including the work you do not do, and returns a plain yes, no, or unsure with a short reason. Your own decision on a notice overrules the system from then on and teaches it.

Does Yako submit bids or applications for me?

No. Yako surfaces the opportunities that fit and explains why. You decide what to pursue and you write the pitch. This is a deliberate design choice, not a missing feature: you stay in control of what goes out under your name.

Can I use my own AI assistant with Yako?

Yes. Yako has no built-in chatbot. Instead it opens your own Yako data — your declaration, your saved searches, and the decision and reason on every notice — to the AI assistant you already use, through the Model Context Protocol (MCP), an open standard for connecting assistants to data. You choose which assistant, and you choose where your data goes.

Is Yako GDPR compliant?

Yes. Yako is operated by SF4L S.R.L. in Bucharest, Romania, under EU law, on servers in Germany. There are no ads and no third-party tracking. Your data stays yours: Yako does not pass it to anyone else, and when you connect your own AI assistant, you decide where it goes.

Who is Yako for?

Anyone who chases public or publicly funded work across more than one platform or country: small and mid-sized firms, IT and engineering companies, research groups, civil-society organisations, networks of independent professionals, and consultants or freelancers working on their own. Because the work is scattered across many sources, languages, and countries, Yako helps all of them, whether or not standard categories describe what they do. It is simply at its sharpest where those categories fail.

How does Yako relate to Kitsuno?

Same company, same principles, a different task. Kitsuno reads the job market for an individual and helps them apply. Yako reads the public market for an organisation and surfaces the calls worth its time. Both do most of their work with rules and structure and use AI only where judgment is genuinely needed, both refuse to apply on your behalf, and both treat your data as yours to keep and export.

Is Yako available yet?

Yako is in an early, invitation-based phase, in use with a small number of organisations. You can join the list at yako.kitsuno.ai.


Published 19 August 2026. No affiliate links. No sponsored placements.