# SkillMix — Full Content Index > Concatenated full-content version of all SkillMix pages, per the llms.txt spec. --- ## SkillMix - AI that works Source: https://skillmix.nl/ # SkillMix — AI that works > Elite talent consultancy delivering AI automation, autonomous AI agents, and quantitative modeling for businesses across the Benelux. Fixed price, fixed date, no-cure-no-pay. Belgian BV, Antwerp. We build AI that runs in production, not slides. We agree up front what it must do. We guarantee it does it, or you don't pay. - **Average pilot turnaround:** 2 weeks. - **Track record:** 10+ projects shipped into production. - **Promise:** works as agreed, or you don't pay. - **Team average age:** 23.8. AI-native operators, zero legacy consultants. --- ## Why we exist Most "consulting" today is recycled frameworks and decks no one uses. We don't sell hours or slides. We sell working solutions. Acceptance criteria, agreed up front. If the system doesn't pass, you don't pay. Worst case: you walk away with an honest diagnosis. That's it. > "We turn down more projects than we take. If we can't guarantee it, we won't sell it." --- ## Why clients choose SkillMix Three reasons clients pick us over a traditional consultancy or an internal hire: 1. **No cure, no pay — works, or it's free.** We define acceptance criteria up front, exactly what the system must do and how we'll test it. If it doesn't pass, you don't pay. Most firms won't stand behind their work this way. We do. 2. **Built, not briefed — ship, don't slide.** Every engagement ends in something running in production: an agent, a model, an automation your team actually uses. No frameworks, no roadmaps for the shelf. Working software, measurable ROI. 3. **Move fast — pilot in two weeks.** We scope tight, build tighter, and get a working pilot in front of your team in weeks, not quarters. Validate the business case before committing to scale. --- ## Three practices. One promise: it works. Every engagement is custom, but it always ends in a system running in production, doing what we agreed it would. Not a deck. Not a pilot you can't use. ### 1. AI Automation — your operations, on autopilot End-to-end automation of the processes eating your team's time: Lead-to-Cash, reporting, inbox triage, core workflows. AI-driven, integrated into your existing stack, measurable impact from day one. - Lead-to-Cash · Reporting & KPIs · Inbox & Triage · Core Workflow - Most automations achieve 60–80% efficiency gains. - See [/automation.md](https://skillmix.nl/automation.md) ### 2. Autonomous AI Agents — workers inside your stack Custom LLM agent swarms that handle workflows, process information, and support decisions safely. Specialised roles (Scout, Drafter, Critic, Router, Executor, Archivist, Escalator). Frontier models swapped as the field moves — your system stays. - See [/ai-agents.md](https://skillmix.nl/ai-agents.md) ### 3. Data & Modeling — forecasting, optimization, decisions From ML forecasts to optimization models, the math that gives you a measurable edge. Operational apps, forecasting & optimisation, messy-data plumbing. - See [/modeling.md](https://skillmix.nl/modeling.md) --- ## Selected work — running in production - **Real Estate · Cash flow, automated end-to-end.** Advanced cash-flow system with real-time data, customisable credit structures, and automated loan & interest calculations. - **Wealth Advisory · AI-driven lead gen for HNW clients.** Upgraded digital presence, AI lead-gen engine, and a discreet, streamlined appointment funnel. - **AI Startup · Consulting as a Service.** A data layer that unifies a consulting firm's institutional knowledge — playbooks, client work, expert reasoning — into one structured memory, so their AI agents act with real context instead of generic answers. - **BESS / Energy · Energy price forecasting, ML-driven.** Deep-learning forecasting engine ingesting market, weather, and grid signals in real time, predicting energy prices hours ahead so the trading desk defends margin through every spike and crash. - **Public Sector (EU) · GDPR-safe traffic intelligence.** Computer-vision system classifying vehicles, bikes, and pedestrians in real time, paired with an AI optimisation model that cut wait times at intersections — fully compliant hardware and deployment included. - **Construction · Every battery, fully utilised.** Energy management system built for a top-tier corporate, so every factory's batteries are fully utilised — charging, discharging, and trading automatically against real-time prices. --- ## Who builds it Built by the people who grew up in AI. Not consultant retrofits — operators who were coding in high school, shipping SaaS in university, and on their second AI startup by 23. Native fluency, not a learning curve. - **0** legacy consultants on the team. - **9** nationalities on the bench. - **100%** AI-native operators. - **Average age 23.8.** Yes, really. We don't train our people on AI. We hire people who never knew a world without it. That's why our pilots ship in weeks, not quarters. --- ## How we engage 1. **Free AI scan / exploration call (30 min, no obligation).** Tell us what needs to work. We map your highest-leverage opportunity. Book at [reclaim.ai/m/skillmix/exploration-call](https://app.reclaim.ai/m/skillmix/exploration-call). 2. **Paid Scope Sprint (mandatory, fixed fee).** A short discovery and design sprint producing a precise scope, written acceptance criteria, and a fixed price. Worst case: you walk away with an honest diagnosis. Typical: €1,000 – €2,500. 3. **Implementation under no-cure-no-pay.** Fixed scope, fixed price, weekly demos. Implementation fee owed only when agreed outcomes are met. 50/50 payment structure. 4. **Run (optional).** Lightweight monitoring and iteration after go-live, or full hand-off — you own the code, infra, and keys. Most automations and modeling engagements ship in 2–6 weeks after the Scope Sprint. Larger swarms and multi-system integrations: 6–12 weeks. --- ## Pricing — typical ranges Everything is fixed-price and quoted up front during the Scope Sprint. | Engagement | Typical range | |---|---| | Scope Sprint (Automation) | €1,000 – €1,500 | | Scope Sprint (Agents / Modeling) | €1,500 – €2,500 | | Automation prototype | ~€8,000 | | Automation full build | €10,000 – €15,000 | | Agent prototype | ~€12,000 | | Agent harden (build) | €15,000 – €25,000 | | Modeling build (small) | €20,000 – €40,000 | | Modeling build (typical) | €40,000 – €70,000 | | Modeling build (large) | €70,000 – €90,000 | | Frontier retainer / Run | monthly, scoped per system | Above €90k we'll tell you to split it. Engagements should be forecastable; past a certain size they stop being that. --- ## Differentiators - Top-tier analytical and technical talent from the Benelux. - Academic edge: direct access to current AI research. - Outcome-based pricing on a no-cure-no-pay basis (after a paid Scope Sprint). - Bilingual delivery: English by default, Dutch for NL/BE/SR. - EU-hosted by default; on-prem and local-swarm available for regulated data. - You own everything: code, prompts, models, infra, audit logs. No lock-in. --- ## Contact - **Email:** info@skillmix.eu - **CEO direct:** xander@skillmix.nl - **Booking:** https://app.reclaim.ai/m/skillmix/exploration-call - **Address:** Abel Janssenslaan 10, 2845 Niel, Belgium - **Legal entity:** Belgian BV — VAT BE1029020540 — PEPPOL 0208:1029020540 - **Jurisdiction:** Antwerp --- ## Related pages - [/automation.md](https://skillmix.nl/automation.md) — AI Automation practice - [/ai-agents.md](https://skillmix.nl/ai-agents.md) — Autonomous AI Agents practice - [/modeling.md](https://skillmix.nl/modeling.md) — Data & Modeling practice - [/get-started.md](https://skillmix.nl/get-started.md) — How engagements start - [/join.md](https://skillmix.nl/join.md) — Careers (the seat) - [/service-terms.md](https://skillmix.nl/service-terms.md) — Service terms - [/Collaboration-Framework.md](https://skillmix.nl/Collaboration-Framework.md) — Collaboration framework - [/privacy.md](https://skillmix.nl/privacy.md) — Privacy policy --- ## AI Automation for Operations · SkillMix Source: https://skillmix.nl/automation # AI Automation — SkillMix > Custom AI automation for the repetitive work quietly eating your team's week. Not templates. Not no-code duct tape. Systems built for your operation, integrated into your stack, running on day one. Fixed price, fixed date, no-cure-no-pay. Most automations achieve 60–80% efficiency gains. --- ## The observation — every team has a "Tuesday" problem Three patterns we see in almost every operations review: 1. **The inbox that never empties.** Requests, approvals, confirmations. Each one a 90-second tax on someone senior. Multiplied by 80 a day, 200 days a year. 2. **The copy-paste that refuses to die.** Two systems that don't talk. A human translating between them, field by field, every day, hoping they don't fat-finger a decimal. 3. **The report rebuilt from scratch. Again.** Friday afternoon. Same queries, same pivots, same spreadsheet someone's been maintaining for six years. No one remembers why the formulas are purple. You didn't hire your team to do this. And yet, that's exactly what they're doing. --- ## What we build Not a chatbot. Not a workflow wizard. A purpose-built system that takes the repetitive work off your plate, and shows its work every step of the way. Every system follows the same internal shape: 1. **Intake** — trigger fires. Email, form, webhook, schedule, signal from your stack. 2. **Understand** — LLMs and classifiers extract what matters. Validate. Match. Reason. 3. **Act** — post to systems. Notify humans. Create, update, route, approve. 4. **Verify** — acceptance tests run continuously. Exceptions flagged. Nothing silent. --- ## Patterns we've shipped before These are shapes the work has taken before. Yours will probably look different. ### 01 · Lead-to-Cash — €6K – €12K From inbound lead to signed invoice, without someone babysitting the pipeline. Qualify, route, follow up, draft, bill. **Deliverables:** lead capture & qualification · automated CRM updates · invoice generation · payment reminders · cash-flow visibility. **Typical:** 3–6 weeks · ERP + CRM. ### 02 · Reporting & KPIs — €5K – €10K The weekly report writes itself. Every number traceable to source. No more Friday afternoons. **Deliverables:** automated KPI dashboards · scheduled reports to Slack/email · cross-tool data sync · anomaly and forecast alerts. **Typical:** 2–4 weeks · warehouse + BI. ### 03 · Inbox & Triage — €7K – €9K Email or ticket streams classified, summarised, routed. Humans only see what actually needs a human. **Deliverables:** AI email classification · smart routing to people or systems · automated ticket creation · daily summaries · optional auto-replies. **Typical:** 2–3 weeks · email/helpdesk. ### 04 · Core Workflow — from €10K The thing only your company does. Your weird, specific, mission-critical process. The one a template could never capture. The starting point for most companies. **Deliverables:** workflow analysis & design · custom integration architecture · bespoke automation logic · multi-system orchestration. **Typical:** 4–8 weeks · custom. ### 05 · Something else entirely Most of the work we actually do lives here — a process that doesn't have a name yet. Bring the one you thought was too weird to automate. --- ## How it goes — four phases, no surprises ### Phase 01 · Scope Sprint — Paid · 1–2 weeks · €1,000 – €1,500 We map the workflow, shadow the people doing it, and write the acceptance criteria. You leave with a concrete plan: scope, timeline, fixed price, and exactly what "done" means. - Process mapping - Technical feasibility - Risk register - Written acceptance criteria Worst case: you walk away with a diagnosis. Still useful. ### Phase 02 · Prototype — No cure, no pay · ~2 weeks · ~€8,000 A scrappy, working MVP that proves the business value on real data. Fast, rough, honest. If it doesn't demonstrate the outcome we agreed on, you don't pay (except hard costs like compute). - End-to-end working slice - Run on your real data - Measurable outcome vs. baseline - Go / no-go decision at the end If the numbers don't add up, we stop here. No pressure to continue. ### Phase 03 · Build — Fixed price · 4–8 weeks · €10,000 – €15,000 Prototype becomes production. Fixed price, fixed scope, weekly demos. If it fails the acceptance tests we wrote together, we fix it until it passes. No overruns, no surprise invoices. - Weekly working demos - Deployed into your stack - Documented, tested, handed over - Acceptance gate before go-live ### Phase 04 · Run — Optional · month-to-month Lightweight monitoring and iteration after go-live. Or hand it off entirely: your team owns the code, the infra, and the keys. - Uptime & drift monitoring - Monthly iteration window - Cancel anytime --- ## Pricing summary | Phase | Pricing | No-cure-no-pay | |---|---|---| | Scope Sprint | €1,000 – €1,500 | No (always paid) | | Prototype | ~€8,000 | Yes | | Full build | €10,000 – €15,000 | Yes | | Run (optional) | month-to-month, scoped per system | — | Typical project range: **€4K – €14K**. Actual pricing varies with use case, integrations, and data complexity. --- ## Selected case studies - **Real estate developer** — cash flow system with real-time data and customisable credit structures; automated loan and interest calculations. - **Wealth advisory firm** — AI-driven lead generation system and automated onboarding funnel for high-net-worth clients. --- ## Frequently asked **What if our process doesn't look like any of your examples?** Good. Those are the ones we find most interesting. Most of our projects don't fit a template — they're the weird, specific thing your company does that no off-the-shelf tool understands. Bring it. **We don't have clean data. Is that a blocker?** Rarely. Data hygiene is part of the build. We've scoped projects where step one was writing the queries that produced the data the business thought it already had. **How is this different from n8n, Make, Zapier?** Those tools are great for moving data between boxes. We build when the work requires judgment, not plumbing. **What happens if it breaks?** Every system ships with acceptance tests and monitoring. Failures route to a human with full context, never silently. Run-phase SLAs are defined up front. **GDPR, security, where does the data live?** EU-hosted by default. Standard DPAs and sub-processor agreements. For sensitive work we run on-prem or in your own cloud account. The code is yours. **Who actually builds it?** The people you meet in the Scope Sprint. No handoff, no juniors, no offshore. Small team, senior operators, end to end. --- ## Links - Get started: https://skillmix.nl/get-started - Booking (free 30-min scan): https://app.reclaim.ai/m/skillmix/exploration-call - Other practices: [/ai-agents.md](https://skillmix.nl/ai-agents.md) · [/modeling.md](https://skillmix.nl/modeling.md) - Service terms: [/service-terms.md](https://skillmix.nl/service-terms.md) --- ## Autonomous AI Agents · SkillMix Source: https://skillmix.nl/ai-agents # Autonomous AI Agents — SkillMix > Not one agent. A team of them. A swarm of specialised AI agents working together inside your stack — researchers, drafters, reviewers, dispatchers — each one doing one thing well, each one swappable as the frontier moves. Fixed price, fixed date, no-cure-no-pay. --- ## The observation — most teams buy one big agent. Then the frontier moves. The 2024 pattern was a single large model wired to your tools, told to do everything. It mostly worked. Then a better model shipped. Then another. The thing you paid for last summer looks like a toy now. ### The old way — one agent, one model, one god-prompt - Breaks when the model changes. - Bad at any one thing in particular. - Full rebuild every 6 months. ### The SkillMix way — a team of specialists, each one swappable - One agent per job, best model for each. - Swap one agent when something better ships. - System stays. Engine underneath keeps improving. You don't buy a frontier agent. You buy the posture that keeps it on the frontier. --- ## Anatomy of a swarm — seven roles in conversation Not every swarm uses every role. Simple jobs run with three. Complex ones run with fifteen. But almost every swarm we build is composed from this shortlist of archetypes. ### 01 · Scout Fetches. Searches the web, your knowledge base, the CRM, anywhere the answer might live. Doesn't judge what it finds; just returns it. *Tool-heavy · parallel · cheap model.* ### 02 · Archivist Remembers. Maintains the swarm's shared memory: past decisions, client context, canonical facts. Asked before any new research runs. *Long-context · retrieval.* ### 03 · Drafter Writes. Proposals, replies, summaries, reports. Uses the Scout's findings and the Archivist's memory to produce the first version. *Frontier model · style-tuned.* ### 04 · Critic Reviews. Fact-checks every claim, stress-tests the logic, flags what doesn't hold up. The swarm's skeptic. Never ships anything unreviewed. *Different model · lower temp.* ### 05 · Router Dispatches. Decides which role handles what, what runs in parallel, what escalates. The swarm's traffic controller. *Small model · fast · deterministic.* ### 06 · Executor Acts. Writes to the CRM, sends the email, books the meeting, files the ticket. The only role that touches your real systems. Narrow and audited. *Scoped permissions · audited.* ### 07 · Escalator Knows when to call a human. Watches for edge cases, confidence drops, policy boundaries. Hands off with full context attached. *Rules + judgment · Slack / email.* ### 08 · A role that doesn't exist yet Every serious swarm we've built introduced at least one role we hadn't named before. Yours will too. --- ## The promise — same swarm, different engines underneath Here's what you're actually buying: the system stays stable. The models underneath get better. - **Nov '25 — baseline.** The swarm we shipped late last year. Solid for its time. Routing ran on a frontier model because small models weren't good enough yet. Archivist leaned on a vector DB. Same five core roles as today. - **Jan '26 — model swap.** Router → Haiku 4.5: 40× cheaper dispatch. Scout → Gemini 3 Flash: native web grounding with citations out of the box, three scaffolds deleted. System untouched. - **Feb '26 — new role.** Operator (computer-use, Anthropic) shipped on first pilot swarm — for legacy systems without APIs. - **Mar '26 — retired.** Three vector databases retired in favour of native long-context (1M tokens). - **Apr '26 — model swap.** Drafter → GPT-5.5 (better long-form voice consistency). Critic → Opus 4.7 (different provider on purpose). Same org chart. Newer engines. - **Watching:** open-weight, local-first swarms for clients with hard data-residency constraints. --- ## What ships, what's coming, what we'll say no to We only ship what works. Here's the honest current map. ### Ships today - Document drafting with review loops - Research synthesis across 50+ sources - Structured data extraction from messy inputs - Inbox triage with judgment, not rules - Workflows with clear escalation points - Narrow desktop automation (computer use) ### Close, watching - Multi-hour autonomous execution - Negotiation with real money on the line - On-device swarms for regulated data - Reliable voice agents for complex intake - Cross-session learning without retraining ### We'll tell you no - Irreversible decisions without human sign-off - Legal or medical advice as the system of record - Anything where a 1% error rate is catastrophic - Replacing judgment your team is paid for - "Agentic" for its own sake. Just use a script. --- ## How it goes — four phases, plus one that never ends ### Phase 01 · Scope Sprint — Paid · 1–2 weeks · €1,500 – €2,500 We watch the work being done by the humans who do it now. Then we write the swarm's org chart: which roles it needs, which models power each one, where it escalates, what "done" looks like. - Role inventory · model selection (per role) · escalation policy · written acceptance criteria. - Worst case: you walk away with an org chart for a team you might never hire. Still useful. ### Phase 02 · Prototype — No cure, no pay · ~2 weeks · ~€12,000 A working swarm, on real data, doing real work in front of you. Rough edges visible on purpose. If it doesn't clear the bar we agreed on, you don't pay. - Live swarm, end-to-end · your actual documents, your actual inbox · measurable outcome vs. human baseline · go / no-go decision at the end. - If the swarm can't keep up with the humans, we stop here. ### Phase 03 · Harden — Fixed price · 4–8 weeks · €15,000 – €25,000 Prototype becomes production. Guardrails, audit logs, approval queues, budget caps, rollback. Weekly demos. Every role model-swappable from day one. - Full audit log, every message, every tool call · approval queues + kill switch per agent · budget caps and blast-radius limits · acceptance gate before go-live. ### Phase 04 · Run & Upgrade — Frontier retainer · monthly This is the phase the rest of the industry doesn't offer. We track the frontier, swap models when something meaningfully better ships, and re-run your acceptance tests before any change goes live. No rebuild fees. - Continuous frontier tracking · model swaps on a cadence, not a contract · acceptance tests re-run on every swap · monthly changelog scoped to your swarm. --- ## Pricing summary | Phase | Pricing | No-cure-no-pay | |---|---|---| | Scope Sprint | €1,500 – €2,500 | No | | Prototype | ~€12,000 | Yes | | Harden (build) | €15,000 – €25,000 | Yes | | Frontier retainer | monthly, scoped per swarm | — | Agent work costs more than pure automation because the swarm runs on frontier inference; the retainer covers continuous model tracking and swaps on your behalf. --- ## Frequently asked **Aren't AI agents still too unreliable?** A single large agent — often, yes. A swarm of specialised agents with a Critic role and human escalation — routinely no. The reliability story of 2023 is not the reliability story of 2026. We only ship what clears the acceptance gate we wrote with you. **What happens when a better model ships a week after you deliver?** That's what the frontier retainer is for. We track model releases, swap the affected role, re-run your acceptance tests, and only promote the change if it clears them. No rebuild fees. If you don't want the retainer, you still get a fully model-swappable architecture — you just do the swaps yourself. **Who owns the IP and the code?** You do. Code, prompts, role definitions, memory, audit logs. All yours, handed over, documented. You can cancel the retainer and keep running. We don't lock swarms to our infra. **Which models do you actually use?** Whichever is best for the role right now. Today that usually means a mix: a frontier model for the Drafter, a different frontier model for the Critic (different provider on purpose, so failure modes don't rhyme), a small fast model for the Router, and open-weight or on-prem for anything touching regulated data. This mix will be different in six months. **How is this different from Automation?** Automation handles work where the shape is known and the rules are stable. Agents handle work where every case is different and the answer requires judgment. If you can describe the steps, you want Automation. If you can only describe the outcome, you want a swarm. **What about safety and data?** Every tool call is audited. Every agent has scoped permissions. Every action that touches money, customers, or irreversible state routes through an approval queue by default. EU-hosted standard. On-prem and local-swarm for regulated data. Kill switch per agent. --- ## Links - Get started: https://skillmix.nl/get-started - Booking: https://app.reclaim.ai/m/skillmix/exploration-call - Other practices: [/automation.md](https://skillmix.nl/automation.md) · [/modeling.md](https://skillmix.nl/modeling.md) --- ## Data & Modeling · SkillMix Source: https://skillmix.nl/modeling # Data & Modeling — SkillMix > Other firms sell you a project. We sell you the thing. Custom software and ML models on a contract you'd recognise as sane: fixed price, fixed date, and if we don't deliver, you don't pay. --- ## The contract — three terms, nothing clever Most firms won't commit to this because they can't. The padding in their quote is there to cover the uncertainty they know is coming. We can commit to it because we've removed most of the uncertainty on purpose. ### 01 · Fixed price You know the number before the work starts. No T&M. No discovery-phase billing. No change orders without a new, fixed number. *Why we can commit:* small, sharp team — everyone on the build ships. Opinionated stack we've shipped before, not a snowflake rebuild each time. Scope defined on paper in the Scope Sprint, not negotiated during build. ### 02 · Fixed date You know when you'll have it, and it's not measured in quarters. Typical build is six to twelve weeks, end to end. *Why we can commit:* small team, no handoff theatre — the people who scope are the people who ship. AI-accelerated where it helps (glue code, scaffolding, test coverage), not where it doesn't. No steerco. Weekly demo, decisions by end of call. ### 03 · No delivery, no pay If we miss the date or the acceptance criteria we signed, you don't pay. Not a discount. Not partial credit. Zero. *Why we can commit:* we scope small enough to be confident — if we're not, we won't quote. Acceptance criteria written in plain English before work starts. We eat the overrun, not you. Has happened. Will happen again. Priced in. The thing you're paying for, in the end, is the alignment of incentives — not the code. --- ## What you're avoiding — the typical enterprise build Every exec has been through this. The shape is always the same. | Month | Phase | What happens | |---|---|---| | M0 | Kickoff deck | Slides. Optimism. A steerco is formed. | | M2 | Discovery | Workshops. Personas. A Miro board no-one reopens. | | M5 | First change order | "We found integrations we didn't price for." | | M9 | Scope creep | Three features added. One quietly dropped. | | M13 | Demo, sort of | It runs on a laptop. Not in your stack. | | M18 | Re-scope or cancel | Budget spent. Need has moved. Decision deferred. | A perfectly ordinary enterprise build. No-one is cartoonishly incompetent. This is just what the shape looks like. What you actually want to buy: a delivered product, on a date you can plan around, at a price you already know. --- ## What we build — three shapes of work Not a catalogue, an orientation. Every engagement is scoped from scratch, but most of them look like one of these. If yours looks like two, that's fine — we'll price it as one. ### 01 · Operational apps The tool someone uses on Monday morning to get the job done. Schedulers, dispatch consoles, sales workbenches, internal portals. Full stack, real users, lives inside your systems. - AI scheduler for a multi-site operator - In-house sales tool for a BESS integrator - Internal workbench replacing four spreadsheets ### 02 · Forecasting & optimisation models The maths that turns signals into decisions. Forecasting, classification, operations research. Shipped as a service inside your stack, with retraining and drift monitors baked in. - Day-ahead electricity price forecasting - Demand forecasting, crew routing, shift optimisation - EMS signal layers for battery fleets ### 03 · Messy-data plumbing Rarely a product on its own, almost always needed. Getting the data clean, consistent, and somewhere the models and apps can reach it. We do this when it's in the way, and only then. - Pipeline from five legacy systems into one warehouse - Entity resolution across CRM, ERP, and support data - Cleanup before agents or models can run on top If your problem doesn't fit any of the three, tell us anyway. Half our engagements started as "this probably isn't your thing." --- ## How it goes — four acts We don't run multi-month discoveries. We run four short acts, in order, and the contract is signed before act two. By the time a traditional build is still writing its kickoff deck, we're usually done. ### 01 · Scope Sprint — Paid · ~1 week · €1,500 – €2,500 We sit with whoever does the work today. We write down what the thing needs to do, what counts as done, and what it would cost to build. You walk away with a fixed-price quote and written acceptance criteria — whether or not you continue with us. ### 02 · Build — Signed · fixed price · 6–12 weeks Two to three people, full-time on your project until it's done. We start with the riskiest part first, not the easiest, so if something is going to break the plan, it breaks the plan early enough to talk about. Weekly demo · decisions by end of call · no steerco. ### 03 · Hand-over — Acceptance Against the criteria you signed in act one. Not against vibes. If we clear the bar, we invoice. If we don't, we keep going on our time until we do — or we don't invoice at all. Code, infra, models, docs — all yours. No lock-in. ### 04 · Keep it running — Optional retainer Most clients take a small monthly retainer so someone is on call when a model drifts, an upstream API changes, or a new feature is wanted. You can skip it and run the thing yourselves. Plenty do. Monitoring · small changes · model refresh. --- ## Pricing — typical engagements | Engagement | Range | |---|---| | Scope Sprint | €1,500 – €2,500 | | Build (small) | €20,000 – €40,000 | | Build (typical) | €40,000 – €70,000 | | Build (large) | €70,000 – €90,000 | | Above €90k | We'll tell you to split it | Engagements should be forecastable. Past a certain size they stop being that. --- ## Selected case studies - **Battery Energy Storage operator** — ML forecasting model for energy prices. Higher average profits, less manual work via automated forecasts and simulation-based strategy testing. - **Logistics scaleup** — Optimisation model for distributing 360,000 tons/year from France across Europe. Geospatial demand analysis and route optimisation across regulatory regimes. - **Construction / energy** — EMS for a top-tier corporate, charging/discharging factory batteries automatically against real-time prices. - **Public sector (EU)** — GDPR-safe traffic optimisation: computer vision + an optimisation model that cut wait times at intersections. --- ## Frequently asked **What does "no delivery, no pay" actually mean? Partial credit?** Zero. If we miss the date or the acceptance criteria we signed with you, we don't invoice the build. We eat the overrun. This has happened to us. It will happen again. It is priced into how we scope and how many engagements we take at once. **What if *we* change scope halfway through?** Then we write a new fixed price for the new scope and you decide whether to take it. What you don't get is a surprise invoice at the end. The original contract holds or is explicitly renegotiated — never drifted. **How are you 5× cheaper than the other quotes?** Three things. First, no account managers, no delivery managers, no juniors — just senior people who scope and build. Second, we use AI heavily for the parts of engineering it's good at (boilerplate, test scaffolding, integration glue) so two people do the work that used to need eight. Third, we refuse engagements where we don't already know the shape, so we don't bill you to learn. **Is this just outsourcing with a different haircut?** No. Outsourcing sells you hours. We sell you an outcome on a date at a price. The incentive structures are opposite: an outsourcing partner makes more money the longer it takes; we make money by shipping and moving to the next one. **Who owns the code after?** You do. Code, infrastructure, documentation, models, training data. Yours. Handed over, written down. You can walk away from us and keep running. We don't lock builds to our infra. **What's your stack?** Typescript + React on the front. Python or Go on the back. Postgres by default. Torch or scikit-learn for models, frontier LLMs where judgment is needed. EU-hosted. Every choice is reversible, but we default to boring so the code is boring to inherit. **How fast can you start?** Scope Sprint usually starts within two weeks of the first call. Build kicks off within one to two weeks of the Sprint signing off. We cap concurrent engagements on purpose, so sometimes there's a queue — and we'll tell you. **Why don't big firms offer this contract?** Because the economics don't work at scale. A firm with ten layers of management and two hundred people on the bench has to bill hours to stay alive. A small senior team that says no to most requests can afford to sell outcomes. This is a feature of our size, not a trick we've figured out that they haven't. --- ## Links - Get started: https://skillmix.nl/get-started - Booking: https://app.reclaim.ai/m/skillmix/exploration-call - Other practices: [/automation.md](https://skillmix.nl/automation.md) · [/ai-agents.md](https://skillmix.nl/ai-agents.md) --- ## Get Started | SkillMix Source: https://skillmix.nl/get-started # Get Started — SkillMix > Start an engagement with SkillMix in three steps: free exploration call, paid Scope Sprint, then no-cure-no-pay implementation. The whole sequence is built to give you certainty before you commit real money — and an honest exit at every stage if it isn't a fit. --- ## Step 1 — Free exploration call (30 min, no commitment) Booked via [reclaim.ai/m/skillmix/exploration-call](https://app.reclaim.ai/m/skillmix/exploration-call). We use the call to: - Understand your context and the problem you actually want solved. - Qualify fit — sometimes we'll tell you the work isn't right for us, or that you don't need us at all. - Outline whether the work belongs in **Automation**, **AI Agents**, **Modeling**, or the **AI Exploration Retainer**. - Discuss likely scope, timeline, and the rough range of cost. You leave the call with a clear next step. No deck, no follow-up sales sequence. --- ## Step 2 — Paid Scope Sprint (mandatory) A short, fixed-fee discovery and design sprint. Typically 1–2 weeks. Pricing: **€1,000 – €1,500** for automation work, **€1,500 – €2,500** for agents and modeling. We sit with whoever does the work today. We map the workflow, shadow the people doing it, and write the acceptance criteria. You leave with: - **A precise scope.** Exactly what we'll build, what we won't, and what's out. - **Measurable success criteria.** Written in plain English. The "definition of done" we'll be tested against. - **A fixed price for implementation.** No T&M, no discovery-phase billing, no change orders without a new fixed number. - **A delivery date.** Not a quarter — a date. - **A risk register.** What could go wrong, where, and how we'd handle it. The Scope Sprint is mandatory before any no-cure-no-pay work. It protects both sides: clear deliverables, no surprises. If we conclude the work isn't worth doing, we say so. Worst case: you walk away with an honest diagnosis and a written plan you can hand to whoever you hire next. --- ## Step 3 — Implementation (no-cure-no-pay) Fixed-scope build against the criteria from the Scope Sprint. Standard engagements use a **50/50 payment structure**: half on kickoff, half on outcome acceptance. The implementation fee is owed only when agreed outcomes are met. Depending on the practice, this looks like: - **Automation:** ~2-week prototype (~€8,000, no-cure-no-pay) → 4–8 week build (€10K–€15K, fixed price). - **Agents:** ~2-week prototype (~€12,000, no-cure-no-pay) → 4–8 week harden phase (€15K–€25K, fixed price) → optional monthly frontier retainer. - **Modeling:** 6–12 week build (€20K–€90K, fixed price), no-cure-no-pay against acceptance criteria. Weekly demos throughout. Decisions by end of call. No steerco. If we miss the date or the acceptance criteria, you don't pay — not partial credit, zero. --- ## Step 4 — Run (optional) After go-live, most clients take a small monthly retainer so someone is on call when a model drifts, an upstream API changes, or a new feature is wanted. You can skip it and run the thing yourselves. Plenty do. For agent swarms, the **Frontier Retainer** specifically tracks model releases, swaps roles when something better ships, and re-runs your acceptance tests before any change goes live. No rebuild fees. You own everything: code, prompts, infra, models, audit logs. No lock-in. --- --- ## Contact - **Email:** info@skillmix.eu - **CEO direct:** xander@skillmix.nl - **Booking:** https://app.reclaim.ai/m/skillmix/exploration-call - **Contact form / start a conversation:** https://skillmix.nl/get-started --- ## Related - [/automation.md](https://skillmix.nl/automation.md) · [/ai-agents.md](https://skillmix.nl/ai-agents.md) · [/modeling.md](https://skillmix.nl/modeling.md) - [/service-terms.md](https://skillmix.nl/service-terms.md) · [/Collaboration-Framework.md](https://skillmix.nl/Collaboration-Framework.md) --- ## Join SkillMix | Careers Source: https://skillmix.nl/join # Join SkillMix — The Seat > SkillMix doesn't hire for headcount. We hire for the seat. 9 of 423 applicants got the seat in 2025. Decision in 21 days. No recruiters, no panel rounds, no ghosting. The projects end. The phone numbers don't. --- ## What "the seat" means We don't hire roles. We open seats. Each one fills slowly, on our timing. The work itself is project-based — a few weeks or months at a time. What you actually leave with after a few years here is a small list of operators who keep picking up your call for the next twenty years. - **2025 intake:** 9 of 423 applications. - **Decision:** ~21 days from application to offer or clean no. - **Engagement:** project-based. - **Reply rate:** every application gets a reply. - **Recruiters in the loop:** zero. --- ## The toolbox — filled by us, yours forever What you walk out with after a few years here isn't a job title. It's a working kit — technical, operational, and human — that compounds for the rest of your career. Real production projects with measurable outcomes, exposure to frontier AI work and senior consulting practice, and a strong network across the Benelux. --- ## The bar Every dot is an application we read in 2025. Nine got the seat. We read every application. Most of the people we say no to are sharp. The ones we say yes to are the ones we'd want to work with for ten years. We look for: - Strong fundamentals in software, ML, or quantitative methods. - Comfort owning outcomes end-to-end (not just tasks). - Pragmatism — we ship. - The kind of judgment that doesn't need to be managed task-by-task. --- ## The network — a launchpad, not a job Engagements are project-based — a few weeks, a few months. What's different is what happens *after*. The people who launch from SkillMix don't drift off. The line back stays attached. - **Project leads, before the job board** — engagements that didn't fit our loop, but fit you. Passed alumnus to alumnus, in a Slack DM, before they ever become a posting. - **Warm intros that open doors** — operators, founders, clients. A short message from someone they trust. - **References with weight** — written by people who actually shipped with you, signed by someone whose word the reader takes seriously. - **The first call when you start something** — a company, a team-lead role, a project you're staffing — you'll know exactly who to call first. - **Operators on speed dial** — something breaks at 11pm. There's a small list of people who will pick up, who already know how you think. - **Mentorship** — a handful of operators who watched you learn the kit. They'll keep telling you the truth, for the next ten years. --- ## How the seat works — engagement shapes We staff per project. None of these are "junior tracks." Every engagement, regardless of shape, goes through the same bar. - **Single engagement** — one scoped deliverable, 4–12 weeks, fixed deadline. The default. - **Part-time, alongside studies** — ~8–20 hours a week, on a single project at a time. For people who want to ship without putting their degree on pause. - **Sprint engagements (holidays)** — full-time during winter and summer breaks. Tight, well-scoped projects where two weeks of focus changes the trajectory. - **Repeat engagements** — the people we work with tend to come back. After a first project goes well, the next one is booked before the current one ships. - **The bench** — even when there's no project this quarter, we add a small number of people. We engage when the right project lands. --- ## Selection — five steps, no recruiters Submit a CV and (optional) a short letter telling us what you'd want to build with us. From there it's screening, an exercise, a founder interview, and a decision. Most candidates know where they stand within three weeks. | Day | Step | What happens | |---|---|---| | Day 0 | Apply | CV + optional letter | | Day 3 | Screen | AI-led screening | | Day 10 | Build | Take-home exercise | | Day 17 | Talk | Founder interview | | Day 21 | Decide | Offer or a clean no | No panel rounds. No ghosting. About three weeks, end to end. --- ## What we expect once you're in - Real ownership of deliverables — not just tasks. - Direct work with clients. - Hitting deadlines and communicating professionally. - Senior-level autonomy from day one. We hire people who can run, not people who need to be managed. --- ## Compensation & support - **Paid.** Compensation is competitive, set per engagement based on role, scope, and seniority. - **Senior project supervision.** You're never the most senior person in the room. - **Internal tooling & AI infrastructure.** Direct access from day one. - **Structured internal training.** Onboarding designed to make you operate at a high standard immediately. International applicants are welcome, provided they're authorised to work in the EU and/or able to operate as a freelancer. --- ## Apply - **Application page:** https://skillmix.nl/join - **General contact:** info@skillmix.eu The seat is open. If this page reads like you, the next move is yours. One application. One letter. The rest is a conversation. --- ## Service Terms | SkillMix Source: https://skillmix.nl/service-terms # Service Terms — SkillMix > Plain-language summary of SkillMix service terms. The authoritative, binding version is the page at https://skillmix.nl/service-terms. This markdown mirror exists for LLM context and is not the legal document. --- ## Provider - **Legal entity:** SkillMix BV (Belgian BV). - **VAT:** BE1029020540. - **PEPPOL ID:** 0208:1029020540. - **Registered address:** Abel Janssenslaan 10, 2845 Niel, Belgium. - **Jurisdiction:** Antwerp, Belgium. - **Governing law:** Belgian law applies to all engagements unless explicitly varied in writing. --- ## Engagement structure Every engagement follows the same three-step structure: 1. **Free exploration call.** No commitment, no fee. Used to qualify fit and outline scope. 2. **Paid Scope Sprint (mandatory).** A short, fixed-fee discovery sprint that produces a precise scope, written acceptance criteria, a fixed implementation price, and a delivery date. 3. **Fixed-scope implementation.** Built against the acceptance criteria from the Scope Sprint, under no-cure-no-pay (see below). No implementation work begins without a completed and paid Scope Sprint. This is not optional — it is the basis on which the no-cure-no-pay guarantee can be offered fairly to both sides. --- ## Pricing & payment - **Standard payment structure:** 50/50 — half invoiced on kickoff, half on outcome acceptance. - **Quoting model:** all SkillMix engagements are fixed-price. There is no time-and-materials billing on implementation work. - **Direct external costs** (e.g. cloud compute, third-party APIs, paid datasets) may be invoiced separately at cost, with prior written notice. - **Pro-rata invoicing** applies where work spans multiple phases or where a deliverable is partially accepted, per the Collaboration Framework. --- ## No-cure-no-pay — what it does and doesn't mean - Applies **only to implementation**, **only after** a paid Scope Sprint, and **only against the success criteria** defined in that Sprint. - If those criteria are not met, the implementation fee is not owed. Not a discount, not partial credit. - The Scope Sprint itself is always paid, regardless of whether you continue with implementation. - Direct external costs (e.g. compute) may still be invoiced even where the implementation fee is waived. - Material scope changes during the build trigger a short re-scoping pass and a new fixed price; the original no-cure-no-pay terms travel only with the originally scoped deliverable. --- ## Confidentiality & IP - All client information is treated as confidential. NDAs are standard and signed on request before any sensitive work begins. - On full payment of the implementation fee, the client owns the delivered code, prompts, models, infrastructure-as-code, documentation, and audit logs. No lock-in. - SkillMix retains the right to reuse generic methods, patterns, and unmodified open-source components. --- ## Liability - SkillMix's liability is capped at fees actually paid for the engagement giving rise to the claim. - Indirect, consequential, and reputational damages are excluded to the extent permitted by Belgian law. --- ## Disputes Any dispute is governed by Belgian law and falls under the exclusive jurisdiction of the courts of Antwerp, Belgium. --- ## Related - Collaboration Framework: https://skillmix.nl/Collaboration-Framework — see also [/Collaboration-Framework.md](https://skillmix.nl/Collaboration-Framework.md) - Privacy Policy: https://skillmix.nl/privacy — see also [/privacy.md](https://skillmix.nl/privacy.md) - Authoritative service terms page: https://skillmix.nl/service-terms --- ## Collaboration Framework | SkillMix Source: https://skillmix.nl/Collaboration-Framework # Collaboration Framework — SkillMix > Plain-language summary of how SkillMix runs client engagements: scoping, billing, change handling, acceptance, and hand-over. The authoritative, binding version is at https://skillmix.nl/Collaboration-Framework. This markdown mirror exists for LLM context and is not the legal document. --- ## Core principles - **Scope Sprint dependency.** No implementation work begins without a completed, paid Scope Sprint defining scope and written success criteria. - **No-cure-no-pay applies to implementation only**, measured strictly against Scope Sprint success criteria. The Scope Sprint itself is always paid. - **Fixed price, fixed date.** Implementation engagements are quoted at a fixed price with a fixed delivery date. No time-and-materials billing. - **Pro-rata invoicing** applies for partial acceptance or phased delivery — fees are invoiced proportionally to the accepted work. - **You own the output.** On full payment, the client owns the delivered code, prompts, models, infrastructure, and documentation. No lock-in. --- ## Phases of an engagement 1. **Free exploration call** — 20–30 minutes, no commitment, no fee. Used to qualify fit and outline scope. 2. **Paid Scope Sprint** — 1–2 weeks. Produces: - Written scope and out-of-scope list - Written, measurable acceptance criteria - A fixed implementation price - A delivery date - A risk register 3. **Implementation** — fixed price, fixed date, weekly demos, no steerco. - Automation: prototype (~2 weeks, no-cure-no-pay) → build (4–8 weeks, fixed price). - Agents: prototype (~2 weeks, no-cure-no-pay) → harden (4–8 weeks, fixed price) → optional frontier retainer. - Modeling: build (6–12 weeks, fixed price), no-cure-no-pay against acceptance criteria. 4. **Acceptance & hand-over** — implementation is tested against the written acceptance criteria. If it clears the bar, we invoice. If not, we keep going on our time until it does — or we don't invoice at all. 5. **Optional run / retainer** — month-to-month monitoring, small changes, model refresh, or (for swarms) frontier model swaps. --- ## Change handling - **Material scope changes** trigger a short re-scoping pass. We write a new fixed price for the new scope and the client decides whether to take it. Original no-cure-no-pay terms apply only to the originally scoped deliverable. - **Minor clarifications** during the build are handled within the existing scope at no extra cost. - **Discovery-phase billing is never used.** All scope work is concentrated in the Scope Sprint. --- ## Cadence & decision-making - **Weekly working demos** during the build phase. - **Decisions by end of call** wherever possible. No steerco, no escalation theatre. - **Riskiest part first.** Builds start with the riskiest part of the scope, so plan-breaking issues surface early. --- ## Confidentiality & IP - All client information is treated as confidential. NDAs are standard and signed on request. - On full payment, the client owns the code, prompts, models, training data, infrastructure-as-code, audit logs, and documentation. - SkillMix retains the right to reuse generic methods, patterns, and unmodified open-source components. --- ## Hosting, data, and security - **EU-hosted by default.** All managed infrastructure runs in the EU. - **On-prem and local deployment** available for regulated data and clients with hard data-residency constraints. - **Standard DPAs and sub-processor agreements** are signed before any personal data is processed. - Every agent / automation that touches money, customers, or irreversible state routes through an approval queue by default. Audit logs cover every tool call. --- ## Related - Service Terms: https://skillmix.nl/service-terms — see also [/service-terms.md](https://skillmix.nl/service-terms.md) - Privacy Policy: https://skillmix.nl/privacy — see also [/privacy.md](https://skillmix.nl/privacy.md) - Authoritative collaboration framework: https://skillmix.nl/Collaboration-Framework --- ## Privacy Policy | SkillMix Source: https://skillmix.nl/privacy # Privacy Policy — SkillMix > Plain-language summary of the SkillMix privacy policy. The authoritative, binding version is the page at https://skillmix.nl/privacy. This markdown mirror exists for LLM context and is not the legal document. --- ## Controller - **Legal entity:** SkillMix BV (Belgian BV). - **Address:** Abel Janssenslaan 10, 2845 Niel, Belgium. - **VAT:** BE1029020540. - **Privacy contact:** info@skillmix.eu. - **Regulation:** GDPR (Regulation (EU) 2016/679) and Belgian implementing law. --- ## Data we collect - **Contact form submissions** — name, email, company, message content. - **Job applications** — CV file, written motivation, contact details, optional links. - **Booking submissions** — calendar booking metadata via Reclaim.ai. - **Authentication & client portal data** — email and credentials for clients with portal access. - **Standard server / analytics data** — IP address (truncated where possible), user agent, referrer, page paths, language preference, geo-region (country level for language detection). - **Cookies** — strictly necessary cookies always on; non-essential analytics cookies only after explicit consent via the cookie banner. --- ## Why we process it (purposes & legal bases) - **Responding to inquiries** — legitimate interest / pre-contract step. - **Evaluating applications** — pre-contract step / consent for retention beyond the immediate decision. - **Operating client engagements** — performance of contract. - **Operating and improving the website** — legitimate interest, with non-essential analytics on opt-in only. - **Compliance with legal obligations** — accounting, tax, and where applicable B2B record-keeping. --- ## Processors & sub-processors We use vetted third-party processors under written DPAs. Current processors include: - **Hosting / backend (Lovable Cloud / Supabase)** — hosts the site, the database, and edge functions. EU region. - **Slack** — internal routing of contact-form inquiries and job applications to the team. - **Reclaim.ai** — exploration call booking. - **Email infrastructure** — transactional and notification email delivery. - **Analytics** — opt-in only, after consent via the banner. A current sub-processor list is available on request. --- ## Retention - **Contact data:** kept for the duration of the engagement and a reasonable period thereafter for legitimate business and legal reasons (typically up to 7 years for accounting where applicable). - **Job application uploads (CV, motivation):** retained for up to 365 days unless the candidate requests earlier deletion. - **Server / analytics data:** standard retention windows; aggregated where possible. - **Authentication data:** retained while the account is active. --- ## Your rights under GDPR You can at any time: - Access the personal data we hold about you. - Request rectification of inaccurate data. - Request erasure ("right to be forgotten"). - Request restriction of processing. - Request portability of the data you provided. - Object to processing based on legitimate interest. - Withdraw consent for any processing based on consent (e.g. analytics cookies). - Lodge a complaint with the Belgian Data Protection Authority (Gegevensbeschermingsautoriteit). To exercise any right, email **info@skillmix.eu**. We respond within the GDPR statutory timeframe. --- ## International transfers We host EU-first. Where a sub-processor processes data outside the EEA, transfers rely on standard contractual clauses or adequacy decisions, and processing is limited to what is strictly necessary for the service. --- ## Cookies A consent banner is shown on first visit. Strictly necessary cookies (session, language preference, cookie-consent state) are always on. Non-essential cookies (analytics, marketing pixels) are only set after explicit opt-in. You can change or withdraw your choice at any time via the cookie controls on the site. --- ## Updates This summary is updated alongside the authoritative policy. Material changes are reflected on the live page first; the markdown mirror is updated shortly after. For the full, binding text see https://skillmix.nl/privacy.