Say hello Visit website Get your website checkedWebsite CheckWith report, from CHF 600

AI Agency
Basel

AI agency for SMEs in Basel and Baselland: automation, AI employees and custom AI applications

Scroll

workflow · enquiry-inbox
  1. mail.inboxEnquiry from client@company.ch
  2. ai.classifyQuote · urgency high
  3. crm.contactContact created · #4821
  4. document.quotePDF created · 2 pages
  5. mail.replyDraft ready for approval
  6. time.saved38 minutes

AI for SMEs in Basel

We are the AI agency for SMEs in Basel and Baselland that does not stop at the presentation. We build what runs in everyday work: automations that pass data between your programs, AI employees that handle routine tasks in Slack, Teams or by e-mail, and custom applications with AI at their core.

We have been developing software for the web in Sissach since 2008: websites, portals, configurators and shops. That is why AI from us comes not as a stack of slides but as a finished tool, connected to your e-mail, CRM, accounting and website, tested with real cases and introduced together with your team.

You have one fixed point of contact who knows your processes. For the kick-off we come to you in Basel, Liestal or Pratteln; the rest we handle by phone, e-mail or video call.

Discuss your project

What AI takes on in your business

  • Enquiries and inbox

    Incoming e-mails are sorted, summarised and given a draft reply. You read, adjust and send.

  • Quotes and documents

    An enquiry becomes a draft quote, a form becomes a finished PDF. The figures come from your data, not from the model’s imagination.

  • Data without retyping

    Invoices, delivery notes and orders are read and end up in the right system, from accounting to the CRM.

  • Knowledge on demand

    An assistant that knows your manuals, contracts and minutes and answers questions with a source reference.

  • AI employee with a role

    An agent with a clear task and clear permissions, reachable in Slack, Teams or by e-mail, around the clock.

  • Answers for your customers

    A chat on your website that answers from your content and hands over to your team when things need a personal touch.

From our own practice

Screen from the project ImmoCensio

Our own product · AI application and automation

ImmoCensio

ImmoCensio rates the property location of every Swiss municipality. A monthly data run fetches the official figures, recalculates the scores for 2,115 municipalities and updates the map. The assessment for each municipality is written by an AI from the key figures, in three languages. Developed and operated by us.

Visit website

Screen from the project Fox CRM

Our own tool · Automated bookkeeping and AI drafts

Fox CRM

Our own CRM, which we work with every day: customer file, quotes, invoices, bookkeeping and an automatic check of our clients’ websites. The system fetches the bank transactions itself every day, matches payments to open invoices and pre-books recurring items using rules it learns from every confirmed booking. When the website check finds something, an AI writes the email draft; it takes the effort from a fixed table. Every automatic booking waits for confirmation, and no email goes out without approval.

01

What an AI agency should deliver for an SME

AI is not a project in its own right. It pays off where someone today retypes, searches, sorts or writes the same reply again and again. A good AI agency therefore does not start with the model but with your process: what comes in, who does what with it, where does it get stuck?

Only then do we decide what the task needs. Often a clean automation without any AI is enough. Sometimes it takes a language model that understands text or writes drafts. And sometimes the honest answer is that the effort does not pay off yet. We tell you that too.

02

Developers, not middlemen

Many providers sell licences and connect two tools with a third. We write the code ourselves. That matters as soon as your case does not fit the mould: an interface to your industry software, a rule that only applies to you, a form on your website that triggers a process directly.

We work with n8n, Make and the interfaces of the major models every day. Where an off-the-shelf tool is enough, we use it. Where it reaches its limits, we write the missing part. That keeps the solution lean and you do not pay for features you never need.

03

Your data stays under control

Before a model reads a single line, we clarify which data it may see and where it runs. There are three routes. The direct one: the frontier models from Anthropic, OpenAI or Mistral via their business offerings, which do not use inputs for training. The shielded one: large models such as Claude via Amazon Bedrock, in your own AWS account and in European data centres. There, the model provider has no access to your prompts and responses. And the local one: open models on your own server or with a Swiss provider, where nothing leaves the building.

Which route fits depends on the data. For drafts, summaries and information without sensitive content, the direct route is usually enough. For customer data and contracts we often recommend Bedrock, because you get the performance of the large models and still determine where processing takes place. For personnel files or patient data, the local model is the safest choice.

That includes clear permissions and a human who approves. With us, an AI may prepare, sort and draft. What gets sent, booked or deleted is your decision, until you trust the process enough to let it run on its own.

04

Start small, measure, expand

We start with a single process that costs you time today. After a few weeks it runs in everyday work and you see from real cases what it delivers. Only then is the next one tackled.

That keeps the risk small and the benefit measurable: in hours per week, in shorter response times, in fewer errors. If you like, we keep monitoring operations, because models, interfaces and processes change.

Prices

Starting prices and the process are all on one page, so you know what to expect before our first conversation.

from CHF 2,200
AI potential analysis, 1 day plus evaluation
from CHF 5,000
Process automation, 2 to 4 weeks
from CHF 4,800
AI Employees, 2 to 4 weeks
from CHF 7,500
Local AI, 3 to 6 weeks
from CHF 12,000
AI application, from 6 weeks

Packages and prices

How we work

  1. 01

    Analysis & Potential

    We find the processes where automation or AI delivers the greatest benefit.

  2. 02

    Concept & Selection

    We choose the right tools and models, matched to data protection and budget.

  3. 03

    Build & Integration

    We build the solution and integrate it cleanly into your existing systems.

  4. 04

    Operation & Expansion

    We monitor operation, measure the benefit and expand step by step.

Questions and answers

What does an AI agency do?

An AI agency finds the processes where automation or artificial intelligence saves time, and implements them. At Fox Computers that means: analysis, choosing the right tools and models, programming, integration with your existing systems, and operation. You do not get a concept paper but a solution that runs in everyday work.

How much does an AI project cost for an SME?

An AI potential analysis costs from CHF 2,200, an AI employee for one role from CHF 4,800, a process automation from CHF 5,000, local AI from CHF 7,500 and a custom AI application from CHF 12,000. Depending on the solution, there are also running costs for models and operation, which we quantify in advance.

Will you come to Basel for a meeting?

Yes. Our office is in Sissach, around 20 minutes from Basel by train. For the initial conversation, workshops and the roll-out we come to you, whether in Basel, Liestal, Pratteln or Reinach. The 30-minute initial conversation is free of charge.

Is my data safe in an AI project?

Before we start, we define which data a model may see and where it runs. For non-critical tasks we use business offerings that do not use inputs for training. For customer data and contracts we rely on Amazon Bedrock: the models run in your own AWS account in European data centres, and the model provider has no access to your prompts. For particularly sensitive data we set up local language models that run on your server or with a Swiss provider, so nothing leaves the building.

Which models and tools do you work with?

We choose by the case, not by the vendor: frontier models from Anthropic (Claude), OpenAI or Mistral directly from the provider, large models such as Claude with stronger data protection via Amazon Bedrock (AWS) in European data centres, and open models that run fully locally on your own hardware. Add automation tools such as n8n or Make. Whatever is missing, we program ourselves.

How long does it take until the first solution is running?

An automation or an AI employee for a clearly defined task is usually running after two to four weeks. Local AI takes three to six weeks, a custom AI application six weeks or more. How fast it goes depends mainly on how well the process is described today.

Do you use AI solutions of your own?

Yes. With ImmoCensio we run our own platform on which an AI writes the assessments of 2,115 Swiss municipalities in three languages, fed by an automatic data run. And our own CRM reconciles bank transactions automatically, pre-books recurring items using rules it has learned, and has an AI write email drafts from the findings of the automatic website check. What we recommend to you, we use ourselves every day.

Do you also work for clients outside the Basel region?

Yes, all over Switzerland. AI projects are easy to deliver remotely because the work takes place in your systems. For the kick-off and workshops we are still happy to come by.

Challenge
us.

Tell us what you have in mind.
You’ll get an honest assessment instead of a sales pitch.

Say hello