Kiwi entrepreneur Fiona Turner and her AI-powered technology are taking the guesswork out of crop yield forecasting, helping berry growers add to their bottom line.
Turner, who is based on her vineyard in Tasmania with her family, is the founder and CEO of Bitwise Agronomy, which harnesses AI and computer vision tech to enable accurate crop forecasting around the globe.
Growing up on the family sheep and beef farm in Whanganui, Turner still has family on the land and says she’s as Kiwi as they come. With a background in technology, working for Fortune 500 companies and banks, Turner felt the pull of her rural roots. When she and her husband purchased their vineyard in Tasmania, working remotely meant she travelled extensively.
“Because I was away so much, my vineyard was out of control. I thought I could use computer vision and AI in the vineyard to get eyes on the ground, even when I couldn’t be there, so I decided to have a play.”
Turner experimented with all sorts of tech in the vineyard from leaf moisture sensors to a swarm of drones that could scare birds. Eventually, she hit on the idea of using simple technology, a GoPro, to gather data.
“I was using software to get really good information and because I built systems for industrial processes I thought there was an opportunity to use GoPro cameras to gather data about vineyards and gain operational insight.”
Soon, the berry industry came knocking, and it all took off from there. “I knew everyone could have a GoPro and get data from around the world to give insight into farming operations.”
In a nutshell, Bitwise Agronomy provides yield forecasting for the berry industry, letting growers know how many berries they can expect to pick on any given day. This is important information, Turner explains, as berry growers are required to provide yield forecasts to their buyers every week, sometimes up to four weeks in advance.
Over or under supply can sting growers in the back pocket. Undersupply may mean having to pay the shortfall or purchase fruit from other growers to fulfil the order. Oversupply could result in a reduced price for extra berries, or the grower may be asked to pick them up or find a new buyer for the extra fruit. Defects can also result in rejection, or loss of value.
“They forecast with us, they make more money. Farmers want to maximise their return, and the more confidence the buyer has, the better the price will be for the farmer.”
Previously, crop forecasting was a manual process. Big operations in South America or Europe might employ people to count fruit, but human error is high.
Bitwise Agronomy uses a combination of AI and computer vision, GoPro footage (all that is needed is five minutes of video per hectare once a month), weather data and past pick information. All of this is converted into growing days and a localised ripening curve is produced.
Accuracy at a block level is about 90% and across a whole farm this averages out at about 95% accuracy. “Analysis throughout the season normally ends up within a couple of percent when compared to what actually happened for the season.”
Turner purposely chose not to build her own hardware, but to piggyback on existing, easily accessible and affordable options, like the GoPro or a drone. Instead, they invested in engineering the AI.
The beauty of Bitwise Agronomy is that anyone can use it. Farm workers simply need to be able to take a video and upload the footage. “The value of the forecasting is so important to them that they are very willing to go and collect data. A lot will attach it to their tractor and do their whole farm in one go.”
Turner worked with the Farm Innovation Network (FIN) to pilot the technology in New Zealand and, although no growers took them up on the pilot, that in itself gave her valuable insight on the market opportunity here.
“Part of the value has come in the connections we have made with other Australian companies who have been vetted through the FIN programme. For example, Nanobubble, who we now collaborate and share contacts with. Connecting with other agtech innovators who we know are offering value has been hugely beneficial.”
Turner says the pilot process helped Bitwise Agronomy to validate the tools they are designing for the New Zealand and Australian markets. “It helped us to understand where our product market fit is, even for growth areas like the UK, Europe and South America, it helped us in our decision to focus in on those markets.”
While the pilot told them New Zealand growers did not have a great need for a yield forecasting tool, they did have a need for tools that would help assess quality.
Turner has soft launched Pick Point, which assesses size and quality of fruit, in the UK and Australia. She hopes it will be released in New Zealand for the coming season.
“For example, sweet batch strawberries attract a premium price, but they need to be at a certain sugar level and size. We have developed a camera and, using AI, images will give you information about size, brix and sugar levels, and defects.”
Being a remote company has allowed Turner to build a truly global business, with employees spread around the world. She is thrilled to welcome a fellow Kiwi, Lotte Bayly, on board as the new head of global revenue, working out of New Zealand.
“We are always looking for talent, and we don’t care where they are in the world. We just find the best person for the job, but it’s super exciting to have another Kiwi on the team.”