Starting a business: From farmer to AI tech founder

It’s not easy to disrupt an industry when starting a business, especially one that has changed very little over centuries such as farming, but our guest this week has done just that.

Our disruptor is Sam Watson-Jones, a fourth-generation farmer and founder of Small Robot Company.

Navigate the show notes to find out more:

Develop a thick skin – there will be challenges to face on your business journey

Reviewing business data and finances can highlight eye-opening insights

Technology can solve your business problems

Create an army of early adopters

If it suits your business, get innovation grants

Software and hardware challenges can be difficult to deal with

Robots need the right software to work

Robots and AI could fix some of your business problems

Be aware of your limitations

Invest in coaching and have a good team around you

Develop a thick skin – there will be challenges to face on your business journey

Bex Burn-Callander:

You are a fourth-generation farmer. It is a little bit more complicated than that because you are a farmer, but you also were a consultant at Accenture. You are a hybrid of the old world and the new world.

Sam Watson-Jones:

I never thought of myself as a farmer growing up. I lived on a farm and worked on it in the holidays, but I never thought, “This is what I want to do forever”.

I would not have considered myself to be an entrepreneur, either. That has happened slowly.

By the end of university, I was interested in the farm, which was the family business, and the only thing we did at the time.

But I thought of it as a family business that happened to be involved in farming if that makes sense, and it could ultimately become anything.

But I also knew that I did not want to go straight home to the farm. I wanted to experience living in the city and do some other things.

So, the first thing I did straight out of university was Teach First, an excellent programme. It is two years teaching in an inner-city school.

It has a fast-track programme where you cram your teacher training into about six weeks, and then you are on your own in a classroom with 30 kids. It was about four or five months after I finished university.

I was straight in at the deep end, and the reason I did that was not necessarily because I wanted to be a teacher but get me as far away as possible into something different and completely different life experience.

It certainly did not disappoint on that front. It was and remains, and I include starting Small Robot Company in this, the hardest thing I have ever done.

The thing it taught me most of all was the ability to get over disappointments. I would take my first lesson as an example. I had spent probably four hours planning this lesson for one hour with this Year Eight class.

Everyone, once they’d seen my timetable, had said, “That Year Eight class, that’s going to challenge you. That’s going to be pretty tough.” I went in, and the first 10 minutes was great. They were quiet but engaged, doing the activity and a little bit of writing.

Two kids were sitting next to each other. About two minutes in, one pulled the paper away from the other, and that kid wrote on the desk accidentally.

Furious, he turned around and threw his chair on the floor and whacked the other boy in the head. Then there were desks tipped over and a full-on brawl about 15 minutes into my teaching career.

I did not know how to get control of the situation, for one thing.

Also, I was devastated that I planned this great lesson that did not happen, and then you need to get over that and go onto your next class, which starts 30 seconds after. That happened repeatedly, and I did not know what I was doing, which was great!

Then I had a year at Accenture.

My job was supposed to start in September 2008, which was not a great time to create a new role, so they postponed my placement for a year.

It gave me a year to work with a farming charity called Jamie’s Farm, which took kids from Teach First type inner-city schools and gave them experience on a farm. I think that year was essential for me too because I started to think, “Actually, farming is something that I want to be involved in.”

That year back out of the city, if you will, although I was still living in London at the time, helped me fall in love again with farming, and think, “At some point, this is going to be something that I’m going to get involved with.”

And then with Accenture – it teaches you all sorts of things that were useful about working in a big company, about understanding, and implementing new technologies, sparking something that I had not thought about in much detail before.

Bex Burn-Callander:

You followed the perfect path for anyone who’s going to join a family business because they always say, “You never go straight into the family business. You have to get experience outside, ideally in other industries, and then bring all that home.”

That is amazing that you did that and found that valuable.

When you did come home, what was the big problem that you were trying to solve? Do you mind giving us a bit of background about the challenges that farming is facing, specifically the farm that you were trying to look at and change?

How did that become the seed of an idea for Small Robot Company?

Reviewing business data and finances can highlight eye-opening insights

Sam Watson-Jones:

When I went back home, I threw myself into being a farmer and tried to learn how the business operated. Part of that was understanding how it ran as a business, where our income streams came from, our costs, and trying to grapple with those figures.

But Dad, bless him, has only just realised that Microsoft Excel exists as a concept. He was very good at recording figures, and so was my grandfather before him and my great grandfather before him, but they did everything in pencil.

Our performance figures, such as costs per acre, revenues per acre, were meticulously recorded in pencil in these red books stacked up.

I thought, “There’s got to be something useful in those books that I can discover if I can find a way to analyse the data.”

I spent an afternoon looking at a 25-year period intending to take the bulk of the data out of these books and put them into a spreadsheet. This tedious exercise told me that our revenues had not increased in 25 years, and our yields had not increased for 25 years.

I was doing this, and I think the yield was the same in 2016 as it was in 1990, and the amount that we sold it for was pretty much the same in 1990.

That hit me like a bullet between the eyes, and I was thinking, “Well, we’re potentially in real trouble here,” because, over that period, costs have gone up massively.

I have since done some research on this, and the costs have gone up about 150% over that same period, but the revenues have not moved.

That was not necessarily the moment that pointed me towards Small Robot Company because there are ways you can wriggle out of that through revenue streams you can pull in. There are things you can do to improve the business or different systems you can adopt.

The first time I thought, “Right. There needs to be something significant that changes here,” was a conversation with an 80-year-old guy who used to work on the farm for my grandfather and had lived next to this field for 40 years.

I was just talking to him, stood there one day, and he casually remarked, “This field used to be about 20 inches higher than it is today.” I said, “What do you mean?” He said, “Well, the soil level was about 20 inches higher than it is today,” That was scary because that meant that the soil and field were disappearing.

I started to research this and found out this is soil erosion – a huge issue and something we created as farmers.

As most people did at the time, we had an intensive farming system, which was about cultivating the field and ploughing it most years to reset the soil. What that does is loosen the soil up.

Then what would also do is apply broad-spectrum herbicides across the whole field, so you kill everything in the field. There is nothing alive, or there is very little that is alive in that soil, making it vulnerable to water and wind erosion.

Little by little, the field was disappearing. I calculated on that field that we lose thousands of tonnes a year. This had been going on for probably longer than 40 years. But it has accelerated in the last 40 years because of the intensive way we have used tractors.

The second instance was the Beast from the East, I think in 2018, where there was a freezing spring in the UK. Usually, if snow falls, it disappears as quickly as it arrives, but some late snows came in this season, and they stayed for about three weeks.

I was driving around the farm and gradually noticed that these enormous snow banks on the roads’ side were slowly turning pink. I did not know what it was, so I jumped out of the car and had a look, and it was soil.

So, we have sandy, light soil blowing off our fields and being caught in this snowdrift in fine particles and being frozen there in place.

It was an indication of something that was happening all the time. Usually, you could not see it, but because of the snow’s particularities that stayed for a few weeks, you could see this soil disappearing.

And it got worse as the days went on.

You could see it day after day, and that was the moment when I thought, “Right. The farm is physically disappearing, and so regardless of what I’m able to work out in terms of diversifying income streams and things like that, there is going to be no farm.”

As I said, I am a fourth-generation farmer, which means a responsibility to the generations that have come before you weigh heavily. I thought, “Unless there is something significant that we do here to protect our soil and make it healthier, we’re in real trouble as a family farm.”

Technology can solve your business problems

Sam Watson-Jones:

Having had my time at Accenture. I thought, “Well, there has got to be a technology solution that can enable a better way of doing this.”

I think that was the first spark towards Small Robot Company. The next crucial step was talking to different people who had ideas.

One academic called Simon Blackmore had developed this concept by asking the question, “What if it wasn’t tractors just getting bigger in the future of farming?

“What if it was small robots instead? What would the future of farming look like?”

I thought this was a fascinating concept, but it had not developed into anything tangible.

Simon said, “I’m sure this is something that farmers want, but it’s not developed yet.”

He said a sentence that was important to me, which was, “What is needed is an entrepreneur to sit between the gap between technical possibility because all this stuff we’re developing is technically possible and what customers demand.”

That was a critical moment for me because I just suddenly thought, “That’s going to be me. That’s what I’m going to do.”

I believe that had been some time coming. It is not quite the sort of bolt from the blue that you imagine when you look back, but it was one critical moment.

Simon’s other significant contribution was introducing me to Ben, my co-founder, around four years ago. Ben was working at Ordnance Survey at the time, and we had a couple of phone conversations.

Ben was working in London at the time, so I got the train down to London, and we met, went for a coffee, walked around, had a bit of a chat, and went for lunch.

By the end of lunch, we decided we were going to form a company together. Honestly, I wish I could say there was more concerted thought than that, but it just felt right, and so we went for it.

Bex Burn-Callander:

I want to go back a step, though, because when you talk about meeting Simon, I imagine that many people listening to this will perhaps have an idea of a problem that needs solving and may also want to use technology to solve that problem.

How did you stumble across Simon, though? Were you looking at different universities and agricultural colleges to see what was going on? How did you make a beeline for him?

Sam Watson-Jones:

It was work that had been going on for a few years.

It is strange to think back on how you join the dots up in your head. Simon had done presentations at significant conferences and things like this.

He had the concepts of the things he was talking about, and although I had never met him, I had seen him speak.

I remember him always emphasising the damage that tractors did to the soil, which resonated with me. He had some interesting figures.

Soil erosion causes a billion pounds of economic loss in the UK every year, and 50% of global topsoil are disappearing, for example. All sorts of things.

I joined up the dots, and interestingly, went to Simon in the same way Ben went to him. I emailed him out of the blue and said, “I think what you’re talking about is a solution that I need on my farm. Who is developing it? Can I offer up my farm as a trial site?”

I was not thinking this was something I was going to start.

I was thinking, “Is there some way I can be involved? I don’t know what that looks like.”

But I remember being excited. I said, “I’m a local farmer and interested in what you’re doing,” and he got straight back.

I went to talk to him, and Simon said, “No one’s developing this.”

He had pitched it to all of the big machinery manufacturers, the people that you would think would build this thing, and they’d all said, “Nah. We are not interested. It is too different from what we do now.”

I did not go into that first conversation thinking, “Oh, this is going to be something I’m going to work on for potentially the next decade,” or however long I work on this. I went in thinking, “This is interesting. I’ll see what happens.”

As the conversation went on, there was something in my head that said, “This is what you want to do.”

Create an army of early adopters

Bex Burn-Callander:

What I love about this business is that it has been evidence-based from the beginning. Even when you developed the robots, you did not just go off and design something. That was all done in conjunction with farmers.

How did you create a community of early adopters – an active focus group that ultimately helped you build what loads of farmers were looking for?

Sam Watson-Jones:

Ben was great because he is an experienced entrepreneur who had been on the entrepreneurial journey four times previously. This is startup number five for him.

So, he said, “We’ve got to focus on the customers, and before we spend any money, we should go and talk to them.”

And so, for the first six months, we did not build any robots. Ben was still working on his job. I was still working on the farm, but I spent time driving around, speaking to about 100 farmers in six months,

I honestly did not know what I was talking about when I went into those first meetings. I had nothing.

I had Simon’s cartoon concepts and asked questions like, “We’re thinking about doing something around getting down to an individual plant level. Do you think that is interesting? What sort of things could a robot do that would be interesting?”

I constantly got verbally beaten up for about two months, but it was great because you went into one meeting where you pitch something as an idea, and they look at you as though you had five heads.

They say, “That is rubbish. That will never work,” and you go, “OK. Right. Chalk that one up to experience.”

And then you go into the next meeting, your pitch has improved, and the proposition and the way you talk about it have improved.

In these conversations, there were a series of moments where I thought, “We’re on to something here. There is a big idea here that we could work on.”

So, we first employed someone in mid-2017, which is Joe, our CTO (chief technology officer).

Ben and I were not making any money out of it. Joe was not taking much money himself, but we had money that we had to come up with, and we did not have many means of funding.

Ben agreed to put a bit in, my family farm put a bit in, but it would only keep us going for a few months. It was my wife who came up with the idea of, “Why don’t you try and get a group of farmers and take a deposit from them?” which I thought was a great idea.

And so, we came up with this concept of, “OK. We are building this thing. If you believe in this, put down a £5,000 deposit on our services, and in return, we will give you £10,000 worth of our service when it is ready.”

That was easily said but very difficult to do.

Six months from that point – mid-2017, we had done the initial interviews. Then from that point, we go, “OK, we’re going to try and form this group.”

We called it our Farmer Advisory Group.

We wanted it to be first customers. We were ideally looking for £100,000, which we knew would see us through nine months, so that was 20 farmers times five. We also wanted it to become a community and become a focus group for us, and that group still exists today.

The group has grown to just over 30 farmers and is so central to everything we do.

From the start, we were building the first robot and asked, “Well, this is roughly how we think the robot should look. What do you guys think as farmers?”

I had an opinion as a farmer, but I was just one farmer. So, I asked questions like:

• How much should it cost?
• What are the most important tasks that this robot should do?
• What are the things in your everyday working life that are the most painful?
• What are the things you would love to automate or have taken away or improved in some way?

All the product ideas that we have developed have been off the back of that group.

But it was hard to get that money out of them, frankly. It was a significant shift from, “Oh, here’s this guy driving around and talking about some crazy technology that’s never going to happen,” to, “OK, you and I,” talking to the farmer, “are going to do this together. We will do this together, and for that to happen, I need you to make a financial commitment.

“I have not got anything. I have not got a robot to show you. I have got a poor slide deck. I haven’t got a team, but we’ve got a clear idea of what this should be.”

A show of faith was a massive, massive confidence boost, but it took a lot of persistence. It took at least six months to get that. A lot of no’s, a lot of no’s that took hours to reach.

In a way, this is a bit of a startup cliché, but the reasons why we were getting the no’s was encouraging, because the no’s were from people who were saying, “If this existed, I would buy it”.

It was reasons like, “It’s just not appropriate for me to put £5,000 in,” or whatever it might be.”

It was all down to their lack of belief in the technology, but not in what the technology was going to do. The farmers were not making an informed decision about technology. It was, “Well, that sounds a bit far out.”

So, they were valuable no’s in that sense, but it was painful, and it required us to stick at it. But it has been the foundation, so it was a great thing to do. I am pleased we did it.

If it suits your business, get innovation grants

Bex Burn-Callander:

So, that £100,000 – how far does that get you towards building a prototype robot?

Sam Watson-Jones:

Not so far. We built our first prototype for around £5,000, but as Joe likes to say, it was essentially bits of pieces bought off eBay and tied together with elastic bands and Blu Tack.

The robot was not farming ready, but it got us into fields. It essentially looked like a remote-controlled car but with a camera.

It enabled us to start gathering data on the fields, understand what needs to happen and what the environment is like on a farm. It was all-important learning.

The real breakthrough moment where you think, “OK, this actually could happen,” was winning our first Innovate UK grant.

We spent a lot of time applying for Innovate UK.

Innovate UK offers UK government grants, where they fund 70% of your project costs for those who do not know. You submit a project proposal under a particular heading, and you get 70% of those costs paid, but they are paid quarterly and in arrears.

You go out, find the money somehow, spend it, and get 70% of those costs back. We won one worth about £560,000 over three years, which was to develop, and we will come on this, Wilma, our software package, which is about per plant recognition.

That was a real moment where you thought, “Wow, we’ve now got some funding in place and can start to see this happening.”

At that point, we could start to get a small team.

We were still building the robots, but there was progress. We had momentum and secured funding. That took us through to the end of 2018, so we were not in bad shape at that point.

Software and hardware challenges can be difficult to deal with

Bex Burn-Callander:

That’s an interesting point because when it comes to building a robot that works on a farm, it is not as simple as making a robot that works on a farm. You need to have the software and analysis behind it.

You need to have all the data and constantly refresh it, so there is a whole wealth of other stuff that needs to be in place before you even put the robot on the farm.

Talk me through that process and how complicated that might have been to set up as a pioneer in this field. Pardon my pun.

Sam Watson-Jones:

This was a real challenge; getting our head around what we were going to do because the advice is for something simple.

Go for the smallest thing you can do. And we were going for the smallest thing you could do, but the smallest thing we could do was complicated.

It was multilayered and something we have been grappling with for four years. It is part of the challenge of developing things for farming and being one of the first to do it.

We tried for a long time in that first year to buy an off-the-shelf robot that we could just get into the field and start gathering data, but we could not do it.

We could have done things like hanging a camera on a tractor or quad bike, but they were not suitable solutions for various reasons, mainly because of weight.

We could buy no robot, so we were forced into building the robot and software, and the technical challenges have been numerous.

This is a complex startup problem because there are so many challenges around physically engineering a robot that can operate effectively on a farmer’s field, such as different weather conditions, soil types and field topographies.

Simple things like dealing with mud are engineering challenges, but they are not insurmountable.

And then for the software.

Again, we tried hard to build on top of existing geospatial or farm software, but there was nothing accurate as we wanted it to be. The proposition had become per plant farming – “We’re going to gather data on every single plant in a field.”

The first crop we chose to focus on was wheat, one of the UK’s biggest crops. If you think of a wheat field, there are billions of wheat plants in the field, depending on the area’s size.

The challenge was gathering billions of data points and knowing exactly where each data point is, giving an exact geo-reference point.

Then it was about doing something different to one data point plant versus the data point plant right next to it. It was about identifying a wheat plant here and a weed right next to it and then how to kill the weed, but not the wheat plant.

For us, this was about digitising the field. It was about genuinely taking farming into a different space. I talked about at the time examples of other sectors transformed because digitisation had become possible.

One example would be the mobile phone, which has become more powerful and cheaper over the last ten years. It is because a series of digital technologies have been built on top of each other, to become computing platforms.

But farming over the same period has not become massively more powerful. The technology that we use has only become incrementally or slightly more powerful, and it certainly has not become cheaper.

We thought those things could become possible if we started with the first digitisation step, but we have only now solved the technical challenges as I stand here now in February 2021.

It has taken until now to solve what we talked about in 2017 naively as: “This is the first step. We are just going to get a per plant database. Done. And then we’ll go on to do those other things.”

The first step has proved to be tricky.

There were constant challenges around, “You’re trying to do too much. This is too complicated. This is too big a problem for you to try and solve.”

There may have been times when they were right, but I am pleased we have stuck to our guns because now we have got something valuable to build on.

And we know how difficult it is to get to this point, so it will be difficult for people to quickly copy what we are doing because of the realities of computers interacting with nature. I see that as being the next frontier for technology.

We have come so far over the last couple of decades, with computers interacting with computers.

There are all these graphs you can look at showing the increase in microchips’ processing speed over the last couple of decades, and it is an exponential growth curve.

But the next step is, “How do you get computers to interact with the natural world and to increase our understanding of nature, and improve the way that we’re able to interact with it?”

We have built the basis for that, but honestly, we are still right at the start.”

Bex Burn-Callander:

It feels like you’re building robots to help farmers.

You are making a platform that will ultimately be the underlying technology for countless other applications, so it is almost like all this groundwork could branch off into all these different areas.

Sam Watson-Jones:

That’s precisely the vision that we become the platform that transformed farming. But what you say is interesting about other technologies.

As I became aware of what robots could do and understood the potential for them, there were so many examples where I looked around the farming industry and thought, “Well, that’s a stupid way of doing things.”

There are sensor application technologies available that a human being can deploy, but this involves doing intensely dull things like going into a field and taking a soil sample.

It is on a handheld device, so after a while, the human goes, “Well, I’m a bit bored of this.”

You end up with a data set that is gappy. Maybe samples have not been taken in the same way. They do part of the field in a good way, and the rest of the field not very well at all.

You end up with results that you cannot trust. Technology like this is a perfect example of what could be done by a robot.

A robot does it 24 hours, the same way time after time, and what you then get is a data set that will transform the way you plan.

We are trying to turn framing into an exact science, to move it from being something where decisions are taken on gut instinct and actions are taken in a broad-brush way towards something done with surgical precision.

The sins of the last century, I suppose, really have been about focusing on speed and scale. Do not worry about the accuracy if you can cover the ground quickly because that is the best way to produce food.

The next wave of agricultural technology needs to be about doing things in a precise way.

Robots need the right software to work

Bex Burn-Callander:

This is probably a good point to introduce robots because we have not talked about what they do.

Can you speak briefly on Tom, Dick and Harry, and maybe a little bit more about Wilma and how she relates.

Sam Watson-Jones:

Our robots are called Tom, Dick and Harry. That was Ben’s idea, by the way, to come up with those names.

Tom is a scanning robot, and he goes out into the field and digitises the field by identifying all the plants and weeds in that field.

Tom sends the data to Wilma, who is the boss. She is the brains of the operation, so she is sucking in the data. She is using machine learning and AI algorithms to recognise the crop and the weeds.

In time, Wilma will also have detailed soil data to know what the soil variations are around the field.

She is also going to recognise the disease, pest damage, and nutrient issues. You have got all the things that could go wrong for a crop understood by this computer.

At various times throughout the year, you need to take actions where Dick and Harry come in.

So, the first use of Dick is that he is a non-chemical weeding robot, and this is what we have advanced with first.

Wilma identifies a weed in a field and gives it this exact geo-reference point. Dick goes out to that same location and physically touches the weed with an electrically charged arm, sending electricity through that weed, killing it without using chemicals.

That is a huge step forward for farming because herbicide resistance is a massive issue in global agriculture. We are trying to take the herbicides out of the equation.

There are all sorts of reasons why using fewer herbicides is good, including reducing pollution and improving soil health. Instead, we are going to use electricity to kill the weeds.

And then, Dick is also a broader crop care robot.

He is still applying chemicals and fertiliser, but he is doing it in a much more accurate way. The power of artificial intelligence is to start to be much more precise with your predictions around when a particular fertiliser or pesticide might be required.

You can begin to make assessments earlier than the human eye can detect what is needed. So, we are trying in those instances to move away from one of the big problems in farming, which is that these chemicals and fertilisers are just routinely applied.

Without necessarily identifying a pest or identifying a disease, farmers just go, “Well, I think there’s a bit of a risk, so I’ll apply it anyway.”

We are trying to get much more exact in terms of how we decide and what we apply.

And then, to finish off, Harry is a lightweight autonomous planting robot going out into the field and placing seeds individually into the ground in an exact way.

When we started looking at Harry, we discovered that they achieved yields approaching three times what you could achieve in a commercial environment in trial plots.

That was interesting to us because then you thought, “Well, it’s not the biology or the genetics that’s holding us back from producing much more food. It is something to do with the farming system.

“There’s something that is not optimum about the farming system.”

Harry can help in developing a way of managing a field that significantly increases that field’s productivity.

Robots and AI could fix some of your business problems

Sam Watson-Jones:

The robots are automating existing processes and doing things in a much more accurate way. But the power that comes in of what we are trying to develop is in two ways.

Firstly, we will enable an interconnected network of robots, learning machines and computers to enter the field. Farmers are fond of saying that you have got 40 harvests to learn about your farm and learn how to grow your crop.

You learn from experience.

You do something on your farm – it works, or it does not work – and then the following year, you try and improve it. That improvement takes a long time.

What robots and artificial intelligence enable is a massive acceleration of that learning cycle. What if a farmer did not have just 40 harvests but 40,000 or 40 million harvests to learn from because every harvest that happens is a learning experience for every robot.

Once you have got thousands of these robots globally learning, they are not only learning from the things that are happening on their farm – they are also learning from the things that are happening on other farms.

If a specific seed spacing for a particular soil type at a certain longitude or latitude is experimented with, the robot acts, and it works, or it does not work. However, that robot learns from the experience and then immediately shares that learning throughout the whole network.

One of the problems with farming now is farmers are isolated. What happens on that farm is down to the journals that the farmer happens to have read or the conferences they happen to have attended, or how switched on to better ways of doing things.

An interconnected network will enable you to make decisions much more quickly. You end up you developing a learning network.

The other thing we get excited about is the potential for developing a farming system that moves away from monocultures.

If you drive around the countryside today, the way the countryside looks has changed significantly over the last 100, 150 years, and that has been to make the machinery that we use in farming more efficient.

Miles and miles of hedges have been removed. Trees have been removed from fields. Fields are big, open, square units of production now, and in those fields, you will see a monoculture.

It will all be wheat, or it will all be corn, or whatever it might be. That is to make the machinery, spray and combine harvester operator more efficient.

But it could be possible to stick a seed in the ground, to watch that seed turn into a plant, and to treat that plant differently from the plant next to it throughout its lifecycle.

Ultimately, if you could harvest that plant individually so that this plant is ready to harvest, although the one next to it is not, you could have a mix of different species in the field.

You could have multiple different varieties of a particular crop. You could have a commercial crop next to a crop that is not commercial but is attracting pollinators of a specific sort.

And if you did that, you would immediately reduce the millions of tons of fertiliser worldwide required.

Now you have got a load of genetically identical things next to each other, which is an entirely unnatural creation. And because you have got that, of course, you exhaust the soil of a particular nutrient. Therefore, of course, you must replace that and replace that nutrient synthetically.

And the same thing with disease.

Because you have got genetically identical plants next to each other, one plant gets the disease, all of them getting the disease is very high.

So, therefore, you spray the whole field with a chemical. We think there is the potential to get away from that and bring biodiversity back into the field, doing things on the per plant level.

Be aware of your limitations

Bex Burn-Callander:

What you are trying to achieve and change is enormous and can impact farmers and our economy, and biodiversity. It feels like it is almost too big for two people.

Does it ever feel overwhelming that you are trying to do all these things? Is it like trying to push the needle on ten different massive industries at once almost?

Sam Watson-Jones:

Yes, but it has never felt too big because I am painfully aware of my limitations. I do not know how to put together any of this stuff practically.

I think the thing I realised early on was what we were talking about was technically possible. We were not waiting for the invention of new things that would enable it to happen. The skills that we need are out there in the world.

And you say, two people. There are more than 40 of us in the team now, and as that team has grown, my confidence in what we are doing has increased at every stage

I did not know how to go out, map a field, and get an exact location for each plant, but we found two or three people who, between them, had the skills to work that out, and we think we are the first people in the world to do that.

They had never done it before, but they had the requisite skills.

So, my job is about getting people excited about this as a concept and as something that they want to devote to the best years of their life.

It is not that it feels too big, but we should be realistic about the chunks we bite off. We move as quickly as we can, but we are realistic about the range of different things we can do in the first sweep.

But the things I have described to you around the end to monocultures, and a massive increase in biodiversity and reduction in pollution. I do not think those things are reliant on Small Robot Company.

I think those things will happen. They are an inevitable change using better technology to produce the food we eat.

I believe we are one of the first companies to see it, and this is a wave that we are trying to catch. The longer we surf that wave, the more we will find the best talent that knows how to solve these problems.

In a sense, Small Robot Company’s journey has been about staying alive until the next stage when you can get the following grant in or the next round of investment.

You can then hire the next level or the next rung of people to solve the following problem.

I will say in some senses that this has got easier over the four years. It felt like pushing water uphill to start. Now, I think it is an easier decision to join Small Robot Company and go, “Well, this is something that I want to focus on” than four years ago. If we stay alive for another four years, then it will be easier again.

It is about momentum and staying with it. It is not trying to solve everything initially. While we have a big vision, it is a multi-decade vision.

I think per plant farming will become the world’s dominant farming system over the next 20 years or so. We are trying to stay alive for those 20 years and become one of the leading players in that market.

Small Robot Company is the next two years and then get out. I want to work on it for a good chunk of the rest of my career.

And that point about not biting off more than you can chew. When you encounter a new issue, like, for example, we needed to have an AI built to analyse these billion data points per field, did you look to make it yourselves, or did you bring it in off the shelf?

When you encounter this barrier, what was your approach to get over that barrier?

A quote that I heard the chief executive of Waymo say the other day was, “Building autonomous cars is a team sport,” and I think that applies to this. This is a team sport, so the approach is to find good people.

That is expensive, so also partner wherever possible. For the AI, we did that through Innovate UK and partnered with a company, and for three years, they provided a lot of the AI solution.

As we have gone on, we decided that the AI was something we want to build in-house. That decision to partner with them took us way further forward than we would have been if we tried to create that all ourselves.

It worked out very nicely for both of us.

Intel acquired our partners, and part of what they were doing with us was a big part of their sell, fitting in exactly what they wanted to do.

That has prompted us to go, “This is something we’re going to build in-house,” because now we have the resources. We can afford skilled data scientists and AI specialists to build this thing.

I think we have always taken a partnership approach. If we can find a quick way to get there by tapping into things that other people have already figured out, we will do that.

There is a balance because you do want to create your IP.

You do not want to give away too much in terms of dilution. There is not a precise formula. We have always taken the approach that we will be pragmatic and try and partner wherever that is sensible.

Invest in coaching and have a good team around you

Bex Burn-Callander:

And just finally, what about how you learn and develop as a leader? Have you found any great mentors along the journey? Any people who have helped guide you to figure a lot of this stuff out?

Sam Watson-Jones:

I have learned a lot from Ben to start with, so that partnership has been crucial. It is my first time making this entrepreneurial journey, so leaning on people that have done it before is essential.

Along the way, I have met other peers and entrepreneurs, so that is important.

I have invested in coaching before and still do. I have tried group coaching, which is business coaching where you get together, and you talk about your businesses and try and solve those problems in groups.

I have also done one-to-one coaching, as well. These are one-to-ones with an experienced entrepreneur who had been through multiple funding rounds and exits and things like that. That is valuable.

The other thing that we did early on was building an excellent non-exec team – a group of advisors.

We did not initially give them any stock options, and for some of them, we did not pay them straight away, so I think that is something that you can do as a resource-strapped early startup.

You can get some good people with relevant experience to give you their time.

We have a couple of hours a month where we have a non-exec meeting, which does take some time to prepare, and then we have probably an hour a month on top of that.

We do this one-on-one, or both Ben and I will catch up with one of them, and we will do a deep dive on a particular issue.

When we thought about it, we thought, “We want someone who’s experienced in partnerships and developing excellent partnerships with big corporates.

“We want someone who’s done the whole entrepreneur’s journey.”

So, we got a guy who was employee number three into a company that eventually sold for a hundred million and went through multiple rounds.

Then we wanted people experienced in developing hardware and AI.

We initially thought, “OK, those are the big things where, if we’re honest, we don’t know what we’re talking about.”

And so, that took some time to find some good people. But it has been a great way of accessing mentorship.

Bex Burn-Callander:

It is about not trying to do everything yourself because I think that can feel like way too much when you start building a business, especially an ambitious business, because you initially believe that it is all on you.

But that partnership approach where you are both working towards a similar goal and your skills complement one another.

That is fantastic advice for anyone building a business.

Sam Watson-Jones:

I remember who, not how, which came from one of the coaching workshops I did. I just thought that was great because your immediate instinct as an entrepreneur is to go, “We’re going to build some robots. How do I do that?” Or whatever the problem might be. Your immediate instinct is to jump in there and try and solve those problems yourselves.

As I said, the skills are out there. It is just a matter of finding the right people who can help you do the how.

If you use that as a mindset, it is helpful because there is no problem that is too daunting. It never feels too big. It is just a case of finding the right who and finding the way that works for both of you for you to be able to access those skills.

It might be hiring on a part-time basis. You can get some amazing people to come and work for you for one or two days a week, and they will be way more than you could afford yourself, but they can go and work part-time.

It is a great win for everyone.

It can be a partnership or outsourcing. There are multiple ways to find a solution to that problem, but the who not how mindset is an important one.

Bex Burn-Callander:

Sam, thank you so much. I feel like I am kind of looking at the face of the man who is changing the way food goes from the farm to my plate, so this is quite a big moment.

I feel like I have learned so much, not just about building businesses, but about farming and about what goes in to growing a crop and keeping it healthy. It is fascinating, so thank you so much.

Like what you’ve read?

Subscribe to Sound Advice on Apple iTunes here.

We are also on Spotify, and anywhere else you get your podcasts.

We would love for you to join our community and share your insights and stories. Perhaps you would make a great future guest on Sound Advice. Let us know on Twitter @SageUK using the hashtag #SoundAdvicePodcast or in the comments below, and we will get back to you.

You can follow Sam on Twitter @samjwj