The ship
Every fleet needs a small, fast ship that goes where the big ones can't. Not consultants, not headcount: a two-man unit — Šarūnas inside, with an owner's mandate and nine years of context; Linas outside, with fresh eyes, zero political baggage, and industrial-grade AI leverage. Self-contained, own tools, one mandate: find asymmetry — the places where one unit of input, an hour or a euro, returns a multiple instead of a percent.
Where we hunt
Three concentric rings. We don't know yet which holds the asymmetry — we know precisely where we're looking, and every find comes back as a proposal with a decision gate. Your call, every time.
The current company
MN's team, products, and processes: the economic perimeter, where margin leaks, and the one or two places a small move creates the biggest change.
The team, aimed outward
Opportunities outside the current products that MN's existing team, infrastructure, and know-how could be pointed at — in part, at first. New lines on existing rails.
Capital & network
Asymmetric opportunities that never enter the MN team — deployed through its capital and, where useful, its network. Each arrives as a standalone proposal with its own gate; nothing is pre-authorized.
What we commit to
The honest answer to "what exactly will you deliver?" is: we don't know yet — and you already know that's the honest answer. Real boots-on-the-ground research will surface things none of us can see from here; fixed targets named today would be fiction dressed as rigor. What we commit to instead is the machinery that makes not-knowing safe to fund:
- A written update every Monday. Five fixed lines: decisions made · numbers moved · next 7 days · blockers and who unblocks them · spend. No meeting attached.
- Every idea becomes a killable hypothesis, tested at the cheapest possible price and kept or killed on evidence. The research behind this proposal already killed twelve of our own favorite ideas — the graveyard is public. That discipline is the product.
- A monthly decision gate. One page, in writing: what was found, what it argues for. Then your call — continue or stop. No open cheque, ever.
Returns, framed honestly
We hunt defensible asymmetries: bets where a unit of input can realistically return 3–10×, each with a clear hypothesis and a cheap way to be proven wrong early. Bigger upside exists — the research maps where — but it's a bonus, never a promise. And the first weeks are diagnostic by design: understand the foundation before betting on it. Plug the leaks first, then soar.
What we need
Two kinds of access, both from day one:
1 · The money floor. A named owner of the numbers who answers within 48 hours: revenue by product line, costs per line, the client list, pricing by geography, burn. Necessary — but not sufficient.
2 · Listening access. Most of what Mysterium knows is written down nowhere — it lives in the team. Slack presence and informal coffee conversations, opt-in, zero meetings added. Listening, not auditing: nobody's sprint gets disturbed.
Terms
| Retainer | €5,000 per month each, pre-tax — Šarūnas + Linas. Invoiced monthly, payment dates fixed in the calendar. |
| Capacity | Up to 100 focused hours each per month. A lighter month bills honestly: hours worked × €50. Full-time jobs and families are real — the meter only ever runs down, never up. |
| Tools | Our own AI stack, at our own cost — included in the retainer. Mysterium never sees a tooling invoice. |
| Experiments | ~€300/month traffic valve (≈€10/day) — enough to read first signals from live tests, and we say so plainly: it reads signals, it doesn't prove conversions. Anything bigger is proposed case-by-case with a budget and the expected signal, and runs only if Mysterium approves and funds it. |
| Start | On your word. |
Plumbing
Three boring sentences that keep everything clean. The agreement names the contracting entity and its authorized signer — so the invoice has an address. Payment dates are fixed in the calendar, unaffected by anything else. Neither side uses the other's name publicly without written approval — in both directions, for both our sakes.
And one page at kickoff with four names on it: executive sponsor · data owner · implementer · budget owner. Questions get answers within 48 hours, or the clock pauses. That is all the process there is.
The long game
If something we build produces sustained recurring revenue above a threshold we agree together, we open the upside conversation — revenue share, equity, or token, whatever fits. Named now, in one sentence, so it never has to be negotiated under pressure later.
The full research behind this proposal — everything we found, everything we killed, and why the optimism is earned — is open at kairos.leematulis.com/research.
Šarūnas Savickas · Linas Matulis