Genève · AI-native consulting
We work from strategy to systems in production that your people use. Our teams are small and senior, and they use AI in their own work. Our fees are tied to results.
- Trade confirmations matched · 0 rekeyed fields
- KYC refresh files drafted · reviewer approves
- Board brief compiled in FR · DE · EN
- Eval suite passed · 412 / 412 cases
- Workflow live · week 4
- Data region · Switzerland
- Model swapped · no code change
- Exception routed · with full context
Network
Network figures cover Emineon and its global delivery network.
- Engineers across the network
- 130+
- Projects delivered
- 200+
- Cities, from Genève to Silicon Valley
- 7
- Days to a first workflow live
- 30
01 · The gap
The licences are bought. A few people use them well. Most of the work is still done by hand, as before.
When someone finds a better way to do a task, it often stays with them. It does not reach the systems where the team works.
We help firms put these tools into daily work, for all teams, with the right governance.
An ordinary Tuesday
WaitingConfirmation to reconcile
Sitting in a shared inbox
Client file to refresh
Due in three weeks, untouched
Report to compile
By hand, on Friday
Question to answer
Ask the one person who knows
22% → 34%
AI adoption among Swiss SMEs, over one year.
AXA SME study · kmu.admin.ch
02 · What we do
We help you decide where AI is worth it. Then we build the systems, run them safely and help your people use them. The same team advises and builds.
- 01
AI Strategy
Where AI is worth it in your firm, in what order and under which rules. Business cases with ROI, a policy and a roadmap for your board.
Explore - 02
Agentic Systems
AI agents that work across your systems. They read the document, check the record, update the ERP and route the exception in production.
Explore - 03
Data & AI Platforms
The base for your AI systems. Governed access to your data, retrieval that cites its sources and a platform your teams can build on safely.
Explore - 04
AI Operations
LLMOps, evaluation, monitoring and governance for AI in production. We keep systems accurate, affordable and auditable after launch.
Explore - 05
Adoption & Enablement
Training, change and AI policy so that all teams use AI well. We measure usage in daily work.
Explore
03 · How we work
You can take one step at a time. Each step has a fixed scope, a clear price and a result you can check before the next one.
- 01 · Assess
Find where AI is worth it
A fixed-fee diagnostic. It covers readiness, prioritised business cases with ROI, an AI policy and a roadmap for your board.
Fixed fee
2–4 weeks
- 02 · Build
Put systems into production
We build agentic workflows in your environment, on your data. Evals and controls are in place from day one. Priced per milestone or per workflow live.
Milestone or outcome-based
4–12 weeks per workflow
- 03 · Run
Operate, improve, hand over
We monitor and improve what is live and roll it out to other entities. We hand it over to your team when it is ready.
Retainer
Ongoing, with a planned exit
- Week 1
Scope
We look at where AI can help, agree the baseline and choose the first workflow.
- Week 2
Build
A senior pod builds it in your environment, on your data. We write the eval set first.
- Week 3
Harden
We handle edge cases, access controls and the audit trail. Your people test it on their own work.
- Week 4
Live
The workflow runs in production. Your team is trained and has a plan for the next one.
04 · Forward-deployed engineering
Our forward-deployed engineers work with your people in Geneva or on site. They learn how the work runs and build in your environment. They stay until your team uses it. Our global network adds capacity across time zones.
Embedded with your team
They join your stand-ups, read your exceptions queue and talk to the people who do the work. Requirements come from daily practice.
Backed by a global network
Engineers in Asia-Pacific and the Americas add build capacity. A small senior pod can then do the work of a larger team.
Sensitive data stays protected
Work on regulated data happens in Switzerland or in your own cloud. The network only works on what is safe to share.
One firm, seven cities
Headquartered in Genève. Engineering and delivery across Asia-Pacific and the United States.
- Cities
- 7
- Follow-the-sun delivery
- 24h
- Accountable team
- 1
Genève
Headquarters
Ho Chi Minh City
Asia-Pacific
Hanoi
Asia-Pacific
Tokyo
Asia-Pacific
Melbourne
Asia-Pacific
Silicon Valley
Americas
Texas
Americas
05 · Industries
Many documents, several languages, regular audits. These Geneva sectors gain a lot of time from AI, and need it done with care.
CoreCommodities Trading & Shipping
Confirmations, KYC and trade finance documents, handled by agents. Traders and operators stay in control.
Value streams
- Trade operations
- Trade finance documentation
- Compliance and KYC
- Market and desk intelligence
CorePrivate Banking & Wealth
KYC, client reporting and advisory preparation. FINMA-level controls, and client data stays in Switzerland.
Value streams
- Client lifecycle and KYC
- Advisory preparation
- Reporting and communication
- Operations and compliance
CoreInternational Organisations & NGOs
Multilingual documents, reporting and knowledge work, with governance suited to public mandates and donors.
Value streams
- Document production
- Institutional knowledge
- Programme and grant operations
AdjacentLife Sciences & MedTech
Regulatory writing, quality documentation and scientific knowledge. Built for GxP and validated where needed.
Value streams
- Regulatory affairs
- Quality and compliance
- Scientific knowledge
AdjacentFragrance, Flavour & Specialty Chemicals
Formulation knowledge, regulatory screening and customer briefs, connected so creators have more time to create.
Value streams
- Brief to formula
- Regulatory and product safety
- Commercial and customer service
AdjacentIndustry & Manufacturing
Maintenance, quality and supply chain knowledge put in the hands of the people on the shop floor.
Value streams
- Maintenance and reliability
- Quality
- Supply chain and procurement
06 · Selected work
Representative engagements, anonymised under NDA. Each one shows a system in production, its baseline and its result.
Commodities trading01A Geneva commodities trader
Trade confirmations matched by agents, exceptions routed to people
- Handling time per confirmation
- −62%
- Confirmations matched without manual rekeying
- 81%
Representative engagementRead the case
Private banking02A Swiss private bank (client withheld under NDA)
Periodic KYC reviews prepared in minutes, decided by people
- Time per periodic review
- −55%
- Average file quality score in compliance sampling (out of 5)
- 4.1 → 4.6
Representative engagementRead the case
International organisation03A Geneva-based international organisation
A searchable institutional memory in six languages
- Time to locate relevant precedent
- −70%
- Languages answered with source citations
- 6
Representative engagementRead the case
07 · Your data
We do not train models on it and we do not keep what the work does not need. Everything runs in an environment only your team can reach.
- 01
Swiss data residency
Hosted in Switzerland or in your own cloud tenant. You choose the region; we document it.
- 02
Built for nFADP, FINMA and GDPR
We follow the revised Swiss data protection act, in force since 1 September 2023. This covers FDPIC breach notification, the register of processing, and privacy by design and by default.
- 03
Prepared for the EU AI Act
We plan risk classification, documentation and human oversight from the start. Systems used in EU markets are then ready when obligations apply.
- 04
Model-agnostic, without lock-in
We pick the right model for each task and can swap it without rewriting the system. You own the code, the prompts and the evals.
Most mid-market clients qualify for the nFADP exemption from the register of processing for low-risk processing. We tell you whether you do.
08 · Insights
What we learn when we put AI into daily work. Written for the people who run it.
EngineeringEvals as the new QA: how to check an AI system is ready, and stays ready
Traditional testing checks that code does what it was written to do. AI systems also need a regular, measured answer to one question. Is it still good enough?
5 min read
DeliveryForward-deployed engineering: why we build AI systems next to the people who use them
Requirements written in a workshop miss much of what matters. Engineers who sit with the team find it within a week. This is how the model works and why it suits AI.
4 min read
CommercialOutcome-based pricing for AI work: how it works, when it fails, and what to ask for
AI is changing the economics of consulting. Hourly billing rewards effort, and buyers want to pay for results. This is how we structure that fairly for both sides.
4 min read
Next step
Tell us about one process that takes your team a lot of time. Happy to have a 30-minute call, no preparation needed.