Guide
Checking a company name with ChatGPT: why that alone is not enough
You have found a name for your new company. Before registering the domain and commissioning a logo, you ask ChatGPT: ‘Is this name still available in Switzerland?’
It is a natural question. ChatGPT helps with business ideas, writes copy and suggests names. Can it also check whether you may use a name?
ChatGPT can provide an initial orientation. A careful name check, however, requires searches in the relevant registers, a search for similar signs and an assessment of potential conflicts. A general chat response does not establish that these checks have been carried out.
Checkname.ch brings these steps together for the Swiss market. The difference matters particularly when a simple search finds no obvious matches.
In brief
- ChatGPT can help develop names and research information. What matters for a name check is which sources and search methods were actually used.
- No exact match does not establish availability. Similar trademarks and company names can also create conflicts.
- Checkname.ch combines Swiss registry data, specialised similarity searches and your intended use in a structured screening process. An additional AI risk assessment helps interpret relevant results.
What ChatGPT can do when choosing a name
ChatGPT is useful for developing name ideas, exploring linguistic associations and asking about the difference between a company name and a trademark. With web search, it can also research current information and link to sources. It would therefore be incorrect to claim that ChatGPT has no internet access. OpenAI explicitly describes this search capability.
For an availability check, the important questions are which data was actually searched and how potential conflicts were investigated.
Did the research cover Swiss trademarks and company names? Did it look for alternative spellings and similar-sounding names? Does the assessment reflect your planned goods and services?
These questions need answers before you turn research into a business decision.
‘Nothing found’ does not establish freedom to use
A web search can reveal existing companies, products and online activity. A name check also needs the relevant registers.
The Swiss Federal Institute of Intellectual Property (IPI) lists Swissreg for Swiss trademarks and international trademarks with protection in Switzerland, ZEFIX for registered companies, and domain-name searches among the relevant sources. It also warns that similar trademarks must be investigated even when an initial search has found nothing. IPI: possibilities and limitations of your own research, in German.
For a ChatGPT response, this means that an absence of reported conflicts leaves a question open: were there no relevant conflicts in the results, or did the research fail to capture them?
Particularly with a little-known company or a new trademark application, a language model’s familiarity with the name tells you little. What matters is whether the relevant record appears in the research.
Similar names are often the decisive issue
Searching for the exact spelling is a useful starting point. It is not enough to identify potential name conflicts.
The IPI lists ‘Alcon’ and ‘Alicon’, for example, as a pair of trademarks for which a conflict was confirmed in court. Names can be spelled differently and still be legally too close. IPI: examples of similar trademarks, in German.
A name check therefore raises several questions:
- Sound: Are there names with a similar pronunciation?
- Appearance: Are there similar spellings or distinctive word elements?
- Meaning: Do the signs convey a similar concept?
ChatGPT can explain these similarities and discuss names you provide. The earlier task still has to be completed: finding the relevant names to compare.
A convincing explanation cannot compensate for a missing registry result. OpenAI itself distinguishes between retrieving the right information and processing it correctly when discussing the quality of AI applications. Errors can occur at either step. OpenAI: optimising LLM accuracy.
The name alone is not enough to assess the risk
Whether a similar trademark is a problem also depends on what you intend to use your name for.
A name for software and the same name for a food product can raise different conflicts. The goods and services involved are particularly relevant. The Nice Classification helps organise them; different classes do not automatically rule out a conflict.
Checkname.ch considers this intended use in its trademark search. Information about your offering is therefore part of the process. The additional AI risk analysis then helps interpret relevant matches. How the Swiss trademark search works.
A general question such as ‘Is this name available?’ leaves essential information unspecified.
What Checkname.ch does differently
Checkname.ch is designed for early screening of company and trademark names in Switzerland. It combines several layers of research in one process:
- Company names: Searches in ZEFIX data for identical and similar company names.
- Trademarks: Searches in trademark data, including pending applications and international registrations with protection in Switzerland.
- Similarity: Comparison of spellings, name elements and phonetic features.
- Intended use: Consideration of relevant goods and services classes.
- Domains and the web: Additional information about domain availability and visible use of the name.
The results bring together specific matches. An additional AI risk assessment explains which ones warrant closer examination. Find out more about company-name screening and trademark searches.
This means you do not have to work out the individual research steps and search variants from a general chat response. The process is already designed around a Swiss naming decision.
Company names, trademarks and domains answer different questions
‘Available’ can mean different things when it comes to a name.
An available domain initially tells you something about an internet address. It does not establish unrestricted permission to use the name as a company name or trademark.
A commercial register entry does not settle every conflict question either. The register authorities examine legality and company-name identity. Disputes about confusingly similar company names may need to be decided by a court. The Federal Office of Justice also points to possible conflicts between company names and protected trademarks. EHRA: company-name searches and protection, in German.
In trademark registration proceedings, the IPI does not check for infringement of earlier rights, for example in trademarks, company names or domain names. It therefore recommends research before filing. IPI: investigating conflicts before a trademark application, in German.
A useful preliminary check needs to distinguish these layers and consider them together.
Why Checkname.ch also uses AI
Checkname.ch also uses AI. Its usefulness depends on the information it receives and the task it performs.
At Checkname.ch, AI complements the specialised search and helps explain the results found. The basis consists of researched company, trademark, domain and web results, together with your description of the planned offering.
The value lies in the combination of data, search methods and interpretation. For your decision, what matters is whether potential conflicts become visible and you understand which next steps make sense.
How to use ChatGPT and Checkname.ch effectively
Use ChatGPT to develop name ideas and prepare your shortlist. Check the remaining names before investing in design, packaging or marketing.
Checkname.ch supports this early screening. For significant investments or unclear results, you can then discuss the findings with a professional specialising in trademark and company-name law.
Even structured research has limits: it depends on the available data and the scope of the search. It does not reserve a name or guarantee registration or conflict-free use.
Its value is in making identifiable risks visible while a name idea is still relatively easy to change.
Check your proposed name before building on it.