Article by Damian Holmes – Founder & Editor of World Landscape Architecture.
In recent weeks, several ranking lists (top 10, top 50, or top 100) of landscape architecture firms have been posted on social media. The majority of the lists were all AI-generated, and many contained errors, such as spelling mistakes or duplicates of firms or duplicates under different names (James Corner, JCFO, Field Operations).
Ranking/Top 10 lists reached their peak in the mid-2010s, as search engines, Wikipedia, and social media began favouring them for their ability to drive clicks and engagement. In more recent times, websites have stopped publishing them, but I am often asked by firms, organisations and others who the top 10 are or where they rank in the world of landscape architecture. My response is that the lists or rankings would depend on the criteria – revenue, projects, awards, design quality, employees, number of offices, locations, sector, and so on.
World Landscape Architecture hasn’t published regular rankings or lists, as I didn’t think they provided a good indication and could create a false impression of the landscape architecture profession. Due to the recent increase in AI-generated lists and the ability to ask AI chatbots any question, I felt it was time to test the various AI chatbots. This is not a scientific test. From the outset, my conclusion was already predetermined – i.e. AIs lists are often flawed and biased, and my idea was to show that the lists can vary widely and sometimes hallucinate the answers.
Simple Prompt
I asked eight AI chatbots (free versions only), including ChatGPT, Gemini, Claude, Grok, DeepSeek, Dola, Qwen, and Mistral, the same question via a prompt: provide a list of the top 10 landscape architecture firms in the world. Some may say that this methodology is flawed because the prompt is too simple and does not provide criteria; however, I decided to use a simple prompt because I have seen examples of students, the public, firm employees, and managers, and others run the same or similar simple prompts as part of their research into landscape firms.
Differing Answers
The answers were interesting, as some included a disclaimer (noted as Disclaimer – see Fig 1 & 2) that there is no single list, ranking, or accepted method for ranking landscape architecture firms, yet provided a list anyway. Others provided a ranking based on reputation, scale of work, and global influence, and one (Mistral) provided a list of reasons for each. Interestingly, four models (ChatGPT, Gemini, Claude, and Grok) provided honourable mentions. Three Chatbots – ChatGPT, DeepSeek and Mistral provided the sources of their ranking, and three asked a follow-up question about whether I required more information about each firm or whether I wished to see a ranking by sector or geographic location.
Results
Conclusion
This shows that AI chatbots will provide information in response to your prompt, sometimes including a disclaimer or an option to seek clarification at the end. You can assume that some answers reflect cultural or language bias, or the number of times a firm appears in publications or on awards lists. Some chatbots used revenue sources (ENR listings), while others used top-10 lists from other websites.
To take this study beyond a simple prompt would involve adding various criteria to the prompts, conducting a reader survey, and further analysing the data. However, given the number of landscape architecture firms and the variance in size, sectors, and geography, it is very hard to determine a ranking list without compromise or by narrowing the ranking criteria, such as revenue, employees, or revenue per employee or number of design projects, etc.
In short, we all should be cautious about any AI ranking or list, and if your management or employees ask this question, it would be better to focus on your design quality and your clients’ (and the public’s) satisfaction with your projects rather than be concerned with your AI ranking.
Article by Damian Holmes – Founder & Editor of World Landscape Architecture.
DISCLAIMER: This article is for educational purposes only. The content is intended only to provide a summary and general overview of matters of interest. It’s not intended to be comprehensive nor to constitute advice. You should always obtain professional advice appropriate to your circumstances before acting or relying on any of that content. This advice is general in nature.
NOTICE: The data and names of landscape architecture firms listed were provided by the AI chatbot. The results may differ when undertaking the same tests with the same prompt due to changes in model versions, differing algorithms, cookies, information provided, prompts, etc.
Be the first to comment