The real questions to ask before launching an AI project
Frequently asked questions about AI agents and AI consulting
Data & security
Where is our data hosted if we work with you?
This is the first point we clarify together. Depending on your regulatory constraints and internal policies, we define the most suitable architecture: hosting in Switzerland, on your own infrastructure, or in a secure international cloud.
Will our data be used to train AI models?
No. vbch Sàrl does not use your data to train AI models. Any other use is defined contractually before the project starts.
How do you protect the confidentiality of our strategic information?
We are used to signing a confidentiality agreement before any discussion. The list of documents and data accessible to the AI agent is then defined and approved with your teams.
Can we use your solutions without relying on foreign servers?
Yes. We can deploy solutions based on open-source models or hosted in Switzerland, completely independent of foreign servers.
Cost & budget
How much does an AI project with vbch cost?
Every project is different. We offer flexible and transparent pricing. The initial strategic consultation is free and requires no commitment.
Do we need a large budget to get started?
No. A first targeted pilot project can be set up without disproportionate investment, to validate AI value before a broader deployment.
Are there recurring costs after deployment?
Yes. During the pilot, we host the solution. After that, hosting and maintenance are either included in a service package or managed by you or another provider.
Timeline & process
How long does a typical AI project take?
A first pilot project generally takes between 4 and 8 weeks, from scoping to the commissioning of an operational AI agent on a targeted scope. It can take several months when integrating the definition of a usage charter, the validation of a strategy and the creation of internal champions through workshops around use cases.
What happens during the first strategic meeting?
In 30 minutes, we review your challenges, data and constraints. You leave with ideas for reflection based on your use cases, budgets, constraints and existing internal skills.
Do we need to mobilize many internal resources?
Internal resources are essential because generative AI changes the way you work, far more than standard software technology. Mobilizing resources is therefore one of the most important topics we address with transparency and seriousness. The goal is to offer you solutions that can actually be implemented and used.
Do you continue to support us after deployment?
Yes, if you want us to, and as long as we're confident we still add value. Our DNA is consulting, training and support: for some clients, the journey takes a few weeks; for others, AI or not, we have been partners for many years. We define this together when the time comes.
Model choice
Which AI model do you use?
We are not tied to any model. We select the solution best suited to your constraints: OpenAI, Anthropic or Google for performance, Llama or Mistral in open source, or solutions hosted in Switzerland. We can also show you a site that references current models, and you may be surprised by the very broad choice of models already operational today. The model is not the issue; its integration and use within your organization are the key questions.
Are we tied to a provider if we work with you?
No. No contract ties vbch Sàrl to any provider. We remain neutral and independent, very Swiss, in short.
How do you choose the right model?
We rely on four criteria: confidentiality, performance, budget and technological independence. We then provide an objective comparison of the available solutions.
Ownership & governance
Who owns the AI agent once it has been developed?
The AI agent belongs entirely to you. The data, document base and configuration remain under your control. We retain no ownership rights.
What happens if we want to stop working with vbch?
You retain full ownership of the solution. We can also train your teams or another service provider to ensure continuity and guarantee your autonomy.
How should AI use be governed in our organization?
We define a clear governance framework with you: access, usage scopes, update rules and audit procedures. Everything is documented and can be presented to your control bodies.
Adoption & teams
Will our teams actually use these AI agents?
We develop AI agents in collaboration with their future users. The interface is simple and uses natural language. Initial testing is conducted with field teams before any widespread deployment.
Do we need technical skills to use the AI solutions you deliver?
No. No technical skills are required as a standard user. They can be used like ChatGPT. That said, we offer advanced training on prompting and analysis to better converse with these AI agents, whatever the models. We are available to train your champions.
What if the answers are not satisfactory?
Continue building and testing. We systematically plan a testing and adjustment phase. If an answer is inaccurate or unsatisfactory, we identify the cause and correct it before deployment.
