Thimm Zwiener – Using chat GPT and the best regen advisors to create a regenerative hotline for all

A conversation with Thimm Zwiener, co-founder of FarmOn, a regenerative hotline for independent farming advice, about how to give everyone of the 600 million farmers in the world access to a trustworthy and regenerative field experience when they have questions, the role of technology, especially the newly hyped large language models like ChatGPT, the risks, and much more.

  • FarmOn is looking for investors that share the company’s passion for the mission, see the potential of the project and want to embark on this journey with them
  • FarmOn is also looking for partners that work with farmers on implementing regenerative agriculture practices and want to scale their work.


It is safe to say we will never get there with our current approach, having amazing consultants and advisors who are usually overworked and can never reach enough farmers to really have a large impact. How do we make the knowledge accessible to all while creating a self-sustaining business and not falling into the trap of gatekeeping, plus on top of that rewarding the people who have assembled this knowledge in the first place?


Thimm argues that the biggest bottleneck right now is to truly understand how to create value on a farm.

”Just giving an answer might not be enough, and so I think truly figuring out how we can create a service that continuously on every day of your farming season, how we can provide value there? And I think that is the biggest bottleneck right now. Because we see that farmers are interested, farmers are asking a question or two. And then they also say that they really liked the response that we gave, but many of them are not coming back yet. And so what that means to us is, ‘Okay, people are generally interested. But maybe it’s too difficult, for example, to always come up with a well-polished question’; it takes me effort. It’s very much like people who already have a good idea and are then able to formulate a question. And I think we need to find ways to ease this process.” – Thimm Zwiener


The ideal stage would be having somebody who has implemented something successfully and has really learned to be able to answer other people’s questions.

”The practitioners who are answering the questions on our platform are actually going further than they’ve been while answering the questions that people are having. And so we see it really as a collective intelligence system. I think that’s the name that we have been phrasing it, where knowledge gets part of it and also gets applied. And then access to new knowledge will be developed from the people who are interested in doing so. But yeah, it should definitely be an attractive way for someone to make an income. And I think it’s also very nice about it, and we can maybe go into that in a bit more detail, but we’ve given a lot of tools that simplify the process of answering.” – Thimm Zwiener


Thimm says, No matter what’s happening, there’s always a human who basically looks over it.’ At FarmOn they have to do quite a lot of work at the moment, but over time, this may be reduced. They are seeing business models for experts too, so when they contribute knowledge, which is becoming part of the entire intelligence system, and it’s been reused more often, then they can compensate you for that.

”Building basically like a copilot, like a tool kit for people who have been trying out things and are also willing to share that, is what the system actually needs. And we don’t need another FAQ, knowledge database, or so. Now what we actually need is applicable knowledge that is context-specific. And this is where experts and pioneers can really play a role. You can, of course, read everything online, etc. But making it applicable, making it like an intelligent action that you’re taking, I think that is what’s really missing.” – Thimm Zwiener

”We give these tools to the experts to get additional context on the farm. […] And then we use large language models and ChatGPT to generate answers based on the previous answers people have been giving. And we let them modify these answers. And so right now, let’s say, in 30% of the cases or, I think, it’s even lower the cases in which ChatGPT, or large language models are able to actually produce an answer that is good. And the other 80% are kind of taken on by the experts right now.” – Thimm Zwiener


The number one concern people have when working together with them is giving away their acquired knowledge, which they have developed for the last 10 or 30 years.

”Gatekeeping is the evil of all scalability. And I think we really want to make regenerative agriculture the norm. That’s the first thing. I think sharing knowledge is, at its essence, the most important thing. […]. And we are looking for business models and intellectual property models that allow sharing revenues based on the knowledge that somebody has gathered over these years. And I think, honestly, we are talking with people exactly around that topic, because that’s the number one thing that is coming up, but I truly believe that there are ways to share all the knowledge that you have ever gathered and still make a profit out of it. And without gatekeeping your knowledge, I think we need to find a future in which this is possible. It can’t be that there are a few practitioners and experts who then do webinars or something for a small group of people, because how many webinars a year can you do? 200? We need to transition 600 million farmers in the world, right?” – Thimm Zwiener


Looking at things from an engineering point of view, Thimm came to realize that the KPIs, the things that you measure in agriculture, have been wrong.

”We measured yield in kilograms, and all of a sudden, the entire system optimizes towards yield in kilograms. And what we end up with is watery tomatoes. And I think, as an investor, I wouldn’t focus too much on measuring generalizable KPIs or anything, but really let farmers and local people set objectives that fit their contexts, right? And then measure these, and that’s fine. But don’t try to find ways to measure regenerative agriculture because it’s probably not possible on a generalizable, large scale. But what is possible is that locals come up with their own experiments with their own objectives that they may try to hit, and trust them that they’re making the right choice. And, I think our system actually really embraces this idea that it’s about the FarmOn context and that it cannot be just like an FAQ or like a YouTube video that is answering all your questions. But it always has to be curated in a way that’s context-specific.” – Thimm Zwiener


Koen and Thimm also talked about:

  • Regenerative agriculture and its potential to become the norm
  • Farming advice and the need for context-specific knowledge
  • Regenerative agriculture and soil health with a focus on tech and business




Feedback, comments, suggestions? Reach me via Twitter @KoenvanSeijen, in the comments below or through Get in Touch on this website.

Join the Investing in Regenerative Agriculture and Food newsletter on

The above references an opinion and is for information and educational purposes only. It is not intended to be investment advice. Seek a duly licensed professional for investment advice.

Leave a Reply

Your email address will not be published. Required fields are marked *