Cloud or your own server – which is better?
Both are possible. The cloud is faster to start and cheaper to run. Your own server makes sense when internal policies or regulations require it – but then maintenance and availability are on you.
Answers about our products, the process, cost and operations.
Both are possible. The cloud is faster to start and cheaper to run. Your own server makes sense when internal policies or regulations require it – but then maintenance and availability are on you.
Yes – from Excel, Outlook, a database or a legacy system. The effort rarely lies in the import itself but in cleaning up: duplicates and inconsistent spellings should be resolved first, otherwise the mess moves along with the data.
Yes. Through interfaces we connect mailboxes, calendars, accounting, ERP systems and web shops, so data only has to be entered once.
Ouhud CRM brings contacts, sales processes, tasks, appointments and reporting together in one platform. It is tailored to each company instead of shipped as a finished product. Available at crm.ouhud.com.
Buying is faster at the beginning. Custom development becomes valuable when your processes strongly differ from standard CRM workflows or require deep integrations.
When recurring workarounds consume time, reduce data quality, or strategic processes cannot be mapped cleanly.
Yes. With clean data migration and a phased transfer of core functions, a controlled transition is very feasible.
Depending on process depth and integrations, projects often range from EUR 15,000 to EUR 120,000.
Yes. We extend existing systems through modules, API bridges, or dedicated portals without replacing everything.
Pipeline transparency, task logic, clean contact history, automation, and reliable reporting are usually the key drivers.
Yes, if the analytics are set up properly. Scanning produces an IP address, which is generally treated as personal data. Plain counting without recognition can rest on legitimate interest; recognising devices or building profiles requires consent.
With dynamic codes, yes: number of scans, time, approximate location and device type. Static codes cannot be measured at all because no server is involved.
With dynamic codes, yes, at any time. With static codes, no – and a printed static code cannot be turned into a dynamic one later. That decision has to be made before printing.
A static code carries the destination inside the pattern – once printed it cannot be changed and cannot be measured. A dynamic code redirects through a server: the target can be changed anytime, scans can be counted, and the pattern stays coarse and easy to scan.
A dynamic QR code does not contain the destination itself but a short redirect address. The actual target sits on a server and can be changed at any time without reprinting the code.
There is no general obligation for business-to-business transactions in Switzerland. Electronic invoicing is only mandatory for suppliers to the federal administration above a contract value of CHF 5,000. The QR-bill is a separate matter: it is a payment instrument, not an e-invoice.
XRechnung is a pure XML file with no visual form and is the usual route for public authorities. ZUGFeRD is a PDF with embedded XML – readable by people and software alike. Both meet the requirement, but ZUGFeRD only from profile EN 16931 upwards.
No. A PDF is readable for people but for software it is just an image of letters. Without structured data it does not qualify as an e-invoice – not even when sent by email.
Every domestic company has had to be able to receive e-invoices since 1 January 2025 – with no transition period. For issuing: from 1 January 2027 for companies with more than EUR 800,000 turnover in the previous year, and from 1 January 2028 for all domestic B2B transactions.
An e-invoice is an invoice in a structured format that machines can read and that complies with the EN 16931 standard. A scanned document or an ordinary PDF does not meet that requirement.
For medical practices, private practices and health centres that want patient administration, appointments, documentation and team workflows in one system instead of several side by side.
Appointment reminders, follow-ups, document assignment, status notifications and recurring reports run automatically. The principle behind it: automate what repeats and involves no judgement. Anything requiring clinical assessment stays with the care team.
AI can structure free text, prepare findings and letters, and summarise long documents with a reference to the source passage. It produces suggestions, not entries: every text is released by a person. That separation is not optional – it is what keeps the software from making a medical decision.
Clinical responsibility always remains with the care team. That is exactly why HealthFusion is built so that every AI suggestion is visibly marked as a suggestion and only enters the record after approval. Software that makes medical decisions itself would be a medical device and would need a conformity procedure.
HealthFusion is designed as a cloud application – no server in the practice, with updates and backups handled in operation. Where internal policies require it, running on your own infrastructure is possible in principle; maintenance and availability are then on your side.
For regular practice operation in Germany, a connection to the telematics infrastructure is required. It is part of our development goal but depends on an approval by gematik, which is not in place. As long as that is missing, HealthFusion does not replace an approved practice management system.
Not completely at this stage – see the question on the telematics infrastructure. What HealthFusion is useful for today is everything alongside the practice management system: appointment organisation, internal workflows, document management and team communication.
Health data falls under the special categories of personal data in Art. 9 GDPR. Encryption, role-based access restrictions and complete audit logging are therefore mandatory – as is medical confidentiality under § 203 of the German Criminal Code, which also binds service providers.
No. HealthFusion is a research and development project and is not offered as a diagnostic or therapeutic system. As long as no conformity assessment under EU Regulation 2017/745 exists, its use is limited to administration and documentation. We state this explicitly because this line is often blurred in our industry.
HealthFusion is our practice software for patient administration, appointments, documentation and team workflows – cloud-based, with a role and permission model.
Classic automation follows fixed rules: "if subject X, then folder Y." AI comes in where the input is unstructured – a freely written email, a scanned delivery note, a set of minutes. The model provides an assessment, not a certainty. Everything around it stays conventional software with clear rules and a review step.
It depends on the task and the data: cloud models such as OpenAI GPT or Anthropic Claude when quality matters and the data is uncritical – or open models such as Llama and Qwen on your own hardware when data must not leave the building. The choice follows the use case, not a preference.
That is your decision. With a cloud provider, their terms apply and a data processing agreement is required. With a model on your own or rented hardware, no request leaves your environment. A middle path is often sensible: pre-process and redact sensitive parts locally, send the rest externally.
When you regularly process data that must not leave the building, when volume is high enough for per-request cost to matter, or when a regulation rules out processing outside the EU. Against it: low volume, very high quality requirements, and no one to run it – an AI server is not a device you set up and forget.
Language models occasionally invent statements that sound plausible. No promise helps against that – only architecture: answers are backed by sources from your own documents, the model states how confident it is, and anything below a defined threshold goes to a person for review. Every decision is logged.
Wherever a wrong assessment costs money, legal standing or trust: creditworthiness, terminations, medical questions, legal advice. Here a model may assist but not decide. For fully automated decisions with significant effect, Art. 22 GDPR also applies – a human review step is then generally mandatory.
From 2 August 2026 the transparency obligations under Article 50 apply: chatbots must identify themselves as such, and artificially generated content must be labelled. The broader obligations for high-risk applications were postponed in July 2026 – to December 2027 and August 2028 respectively. For most companies, only the labelling is relevant for now.
With the step that occurs most often and costs the most time – not the technically most exciting one. A process that fires ten times a day is worth more than five that run once a quarter. Then measure whether it actually saves time, and only then expand.
Mainly repetitive, rule-based tasks with clear inputs: document processing, FAQ support, classification, routing, and pre-checks.
A first pilot is often possible in 2 to 6 weeks. After that, scaling happens step by step based on real data.
With human-in-the-loop, monitoring, thresholds, and clear fallback processes, automation remains reliable and controllable.
Three key levers: automated core processes, a reliable data foundation, and modular architecture for rapid extension.
Typical outcomes are shorter cycle times, fewer errors, better data quality, and teams that scale more efficiently.
Yes. We support you from the idea to a professional system – including planning, development, integration and training.
Analysis → Prototype → Development → Testing → Delivery → Ongoing support.
Depending on scope, projects usually range from €5,000 to €80,000. Small portals start lower, while complex platforms with roles, integrations and automation are significantly higher.
An MVP is often realistic in 6 to 12 weeks. Larger systems with multiple integrations and role models usually take 3 to 9 months.
Yes. We can analyze, stabilize, and gradually further develop existing applications without rebuilding everything from scratch.
A fixed price works well for clearly defined scope. For dynamic requirements, time-and-material with clear prioritization is often more economical.
Use a clearly scoped MVP, gather early user feedback, and prioritize core features consistently.
No. We build fully custom software, tailored to your company – without templates or website builders.
We create a UI/UX concept and analyze your workflows to deliver a clear, intuitive and modern interface.
Yes, all our systems are built to grow with your business.
Custom software fits your processes, reduces workarounds, integrates deeper into your stack, and stays flexible as your business evolves.
WordPress is suitable for content-heavy websites. For B2B systems with roles, APIs, and custom business logic, Laravel is usually more robust and scalable.
An API is an interface that lets systems exchange data in a structured way. Example: your shop reads inventory from ERP without direct database access.
We begin with a free consultation. Afterwards, you receive an initial design or prototype – completely non-binding and without advance payment.
Yes. Every system (CRM, web app, API, automation) includes a full functionality guarantee. Errors are fixed free of charge.
Absolutely. Our systems are fully modular and can be expanded at any time.
Look for proven references, a clear proposal with deliverables, and transparent post-launch operating costs.
Very fast. Critical issues are handled immediately with highest priority.
Yes. We provide updates, security monitoring, optimizations and continuous improvement.
We do not have an answer for that yet. Write to us – we reply personally.
ContactWe do not have an answer for that yet. Write to us – we reply personally.
Free initial consultation - no sales pressure, just clear answers.