AI tools

AI tools for business: what is out there and what you really need

The number of AI tools is growing faster than anyone can test them. For a company, the more important question is not “which one is the best?”, but “how many do we need, and who keeps track of data, contracts and costs?”. This overview sorts the tools by purpose – including areas where we are not the provider.

Sorted by purposeWith selection criteriaAs of October 2026

AI tools by purpose

Examples of well-known providers in each category, not an exhaustive list.

Category	Used for	Examples
AI chat and language models	Writing, summarising, analysing	ChatGPT, Claude, Gemini, Mistral Le Chat
AI platforms for businesses	Several models, your own knowledge, rights management, contract	novendix.AI and other platform providers
Office assistants	AI directly in Word, Excel, Outlook and Teams	Microsoft 365 Copilot, Gemini in Google Workspace
Translation	Translating texts and documents	DeepL, Google Translate
Research	Answers with sources	Perplexity, web search in AI platforms
Transcription and meetings	Recording and summarising meetings	Features in Teams and Zoom, meeting features in AI platforms
Images and design	Graphics, illustrations, drafts	Image generators in design tools and AI platforms
Automation	Connecting processes between systems	Zapier, Make, n8n, workflows in AI platforms
Programming	Writing, explaining and reviewing code	GitHub Copilot, coding features of the major models

General market overview, as of October 2026. Features and terms of the providers named change frequently.

The real problem: too many tools

Every tool a contract

Every tool that processes company data needs a data processing agreement (DPA), a check of the processing location and an entry in the record of processing activities.

Every tool a login

If you introduce ten tools, you manage ten user directories – and when an employee leaves, you forget at least one.

Every tool an invoice

Many individual licences quickly add up to more than one platform that covers several tasks.

Why employees use their own tools anyway is described in Shadow AI in the workplace.

Five criteria for choosing

Data protection first. DPA, processing location, training exclusion. The full checklist is on GDPR-compliant AI.
Coverage instead of point solutions. A tool that covers chat, documents, research, meetings and automation often replaces three to five individual tools.
Choice of models. If you commit to one vendor, you accept compromises for tasks outside its strengths. See LLM comparison.
Connection to your systems. An AI tool that does not know your ERP, CRM or filing system remains a writing tool.
Total cost. Licence, rollout, training, internal time. The calculation is set out in What AI costs in a company.

Which combination for which company

Few users, simple textsA business plan from a chat provider plus a translation tool

Everything in Microsoft 365Copilot for office work – see Copilot alternative for weighing it up

Mid-sized company, many departmentsAn AI platform with several models as the basic equipment for everyone

Own systems need to be connectedA platform with individual integration instead of point tools

Specialist tasksTargeted additions, such as a coding assistant for development

Common mistakes when introducing AI tools

Tool before task. Buying the tool first and then looking for a use case. The other way round works better.
Individual licences for individuals. Equipping only a few people creates two classes – and the others use private tools.
No rules. Without an AI policy, nobody knows which data may go in.
No training. Since February 2025, the EU AI Act has required employees to have sufficient AI literacy.
No follow-up. After 90 days, check which tools are being used and which only cost money.

For help deciding between a standard subscription, a platform and building your own, see Standard subscription, own platform or a mix?

Frequently asked questions about AI tools for business

Which AI tools does a mid-sized company need?

Usually fewer than expected: an AI platform or a business plan for writing, analysing and research for everyone, plus targeted specialist tools, for example for programming. What matters is that every tool is covered contractually and in terms of data protection.

Are free AI tools allowed in a company?

As a rule not for personal data and trade secrets, because there is no DPA and no guaranteed processing location. Details on Using ChatGPT in a GDPR-compliant way.

How much does AI cost a company?

Licences usually cost between €20 and €50 per user per month; on top come rollout, training and internal time. The full calculation is on What AI costs in a company.

Are there AI tools from Europe?

Yes, for example Mistral from France and platforms with processing in the EU. An overview is on European AI.