AI agents, Next.js and GEO/AEO for Swiss corporate finance & tech
French AI studio based in Toulouse. We work with companies in Zurich and German-speaking Switzerland — corporate banking, insurance and reinsurance, tech companies and spin-offs, Crypto Valley — remotely and on site when the project calls for it. FINMA, nLPD and DLT Act framework taken into account; Swiss hosting possible.
Nos clients types
Questions & réponses
Frequently asked questions about working with Nehos from Zurich
Nehos Groupe is a French company (KBC SOLUTIONS, based in Toulouse). We have no office in Zurich. Our engineering team works from Toulouse, and we travel to Zurich for scoping workshops, steering committees and reviews when the project calls for it. Our prices are published and identical wherever you are.
Architecturally. For Zurich projects we can deploy on Swiss providers (for example Exoscale or Infomaniak) and, for AI inference on sensitive data, self-host open-weights models (Mistral) on GPU infrastructure chosen with you, with no call to an external API on sensitive workloads. Every architecture decision is documented so that it can be reviewed before deployment.
Yes, it is a segment we can serve. Zug is about 26 minutes by direct train from Zurich main station. For FINMA-regulated crypto companies, AI agents can help with KYC/AML automation, FATF Travel Rule compliance and customer service scaling. We design them with data segregation aligned with the DLT Act and FINMA guidance in mind. We do not cite client names here; details of our method are shared in a first conversation.
They are published and the same as for any other client. AI express audit (1 day): 990 € excl. VAT, deducted from the project if launched within three months. Full AI audit (5 days): from 2,125 € excl. VAT. AI agent MVP: from 11,000 € excl. VAT. Multi-agent platform: from 50,000 € excl. VAT. Development on a time-and-materials basis: 275 € excl. VAT per day. Software maintenance: 750 € per month. Invoicing in EUR. Travel costs are stated in the quote.
Yes, it is one of the clearest use cases for a controlled stack. Spin-offs often have intellectual property that cannot be processed on US hyperscalers for IP or contractual reasons. We can build fine-tuning and inference pipelines on Swiss infrastructure, with contractual data residency, to train domain-specific models on proprietary corpora without data leaving the agreed environment.
Different regulatory instruments, same principle: control over data and models. For banks, FINMA's rules on operational risk and outsourcing, and model governance documentation. For insurers, the supervisory framework for insurance and the Swiss Solvency Test for quantitative risk models. We prepare the documentation with the client's compliance and actuarial teams; the regulatory sign-off remains with the institution.
Project delivery and technical documentation are in English, the default business language for enterprise AI in Zurich. For procurement processes or board presentations requiring German, we can work with a native German-speaking consultant. Contracts with Swiss entities can be executed in German or English.
The market structures differ. Geneva: private banking, international organisations, watchmaking, commodity trading — French-speaking, relationship-first, very sensitive on sovereignty. Zurich: corporate banking and asset management, insurance and reinsurance, tech companies and spin-offs, Crypto Valley — English-first, more process-formal. The same studio, the same prices, with attention to each market's expectations.
Exploration associée
#Nehos and Zurich — AI for Swiss corporate finance & tech
Nehos Groupe, a French AI studio based in Toulouse, works with companies in Zurich and German-speaking Switzerland. Corporate finance, insurance, tech companies and spin-offs, crypto/fintech: FINMA and nLPD framework taken into account.
Zurich concentrates an economic ecosystem unlike any other in Europe: banks and insurers, a university spin-off pipeline, large engineering sites, and the Crypto Valley corridor down to Zug. Companies here do not need convincing about AI — they need partners who ship production systems that pass a risk review.
That is how Nehos Groupe works. Every AI or data project is scoped against an outcome we measure during the pilot, before you decide to go further. We do not promise a figure in advance.
#How we work with Zurich companies
Nehos has no office in Zurich. With Zurich-based clients, we:
- Run workshops in person when the project calls for it — scoping sessions, sprint reviews and steering committees at your offices
- Work by video call for day-to-day follow-up, which shortens decision loops with your business teams
- Keep data confidential by design — hosting on a Swiss or European provider chosen with you, self-hosted models when your compliance team requires it
#What we deliver
- AI agents in production — connected to your core systems, with the observability and guardrails a regulated environment demands
- RAG on your document base — retrieval-augmented generation for legal, compliance and client-facing teams
- GEO/AEO visibility work — being cited by ChatGPT, Perplexity and AI Overviews when Swiss decision-makers ask for solutions in English or German
- Next.js platforms — fast, accessible, multilingual corporate and product sites
Every engagement has a dedicated lead, backed by the engineering team in Toulouse — more than 40 projects delivered.
#How an engagement starts
- Audit — an AI express audit in one day (990 € excl. VAT, deducted from the project if launched within three months) or a full audit over five days (from 2,125 € excl. VAT): we map your processes, identify the AI use cases that matter and size their potential
- Pilot — a working prototype on your real data to validate feasibility (an AI agent MVP starts from 11,000 € excl. VAT)
- Production rollout — progressive deployment with monitoring and continuous evaluation
- Skills transfer — training your teams so autonomy is the end state, not vendor dependence
The questions below come up in most first conversations with Zurich clients.
#Why a local AI presence in Zurich?
Proximity changes how an AI project runs. Working with Zurich-based clients, we:
- Know the local economic fabric — banking and insurance constraints, pharma and industrial supply chains, the standards Swiss counterparties expect
- Run workshops in person — scoping sessions, sprint reviews and steering committees at your offices
- Shorten decision loops through direct coordination with your business teams
- Keep data confidential by design — sovereign hosting on OVHcloud, self-hosted models when your compliance team requires it
#What we deliver from Zurich
- AI agents in production — connected to your core systems, with the observability and guardrails a regulated environment demands
- RAG on your document base — sovereign retrieval-augmented generation for legal, compliance and client-facing teams
- GEO/AEO visibility work — being cited by ChatGPT, Perplexity and AI Overviews when Swiss decision-makers ask for solutions in English or German
- Next.js platforms — fast, accessible, multilingual corporate and product sites
Every engagement has a dedicated lead who knows your ecosystem, backed by the full delivery team in Toulouse — more than 200 companies accompanied since 2014.
#How an engagement starts
- Free audit — we map your processes, identify AI quick wins and estimate the potential return
- Rapid POC (4–6 weeks) — a working prototype on your real data to validate feasibility
- Production rollout — progressive deployment with monitoring and continuous evaluation
- Skills transfer — training your teams so autonomy is the end state, not vendor dependence
The questions below come up in most first conversations with Zurich clients — proof that the local market has specific expectations of AI partners.