Using AI to fix fractured value chains: a path best tackled facing forward


Matthew Smith, Microsoft ResearchMatthew Smith was an ecologist before he joined Microsoft to lead the development of new predictive models for environmental systems, he spoke to us a couple of years ago and in that time the digitisation of agriculture has gathered momentum.

We caught up with him ahead of the Pollinator “The AI’m of machine learning” to ask him how he sees things developing.

Q. What do you think have been the most important developments in the ‘digitisation of agriculture’ since you last spoke at an Agri-Tech East meeting?

Probably the establishment of the UK centres of excellence in agri-tech. I think we’ll look back at those and see them as hugely important early incubators for much of the improvements we see in the field and further up the supply chain. Beyond that, the roll out of the multiple flavours of agri-tech services as genuine businesses or applications, rather than just ideas. Thinking here about agri robots, aerial image analysis and connected livestock.

Q The market seems to be consolidating – Do you think there is still an opportunity for young innovative companies to create businesses or is it best for them to concentrate on interoperability with the dominant platforms?

I think there are more innovation opportunities than there ever were. The prevailing platforms should provide more opportunities for innovators; and that would be the sign of a good platform. Innovators should be more liberated now, as they will need to build less from scratch and can assume that many of the building blocks exist. Instead they should focus on creating new business capabilities that will deliver value.

Q What are you working on at the moment that is exciting you?

I’ve always been excited by information services for the food supply chain. Finding ways through which the right information can get to the right people, at the right time, in the right way. Right now that information supply chain still doesn’t exist, as it should. It is fragmented and inefficient. That is having a real impact on the health of our agri-food supply chains in terms of waste, price and sustainability.

Wonderfully, technology is approaching a maturity to make the information supply chains the world needs; harnessing things like cloud computing, IoT, AI and block chain. I have the privilege to be working on how to make such systems work, from both business and technical perspectives. It gives me a great sense of purpose.

Q Do you think the future is ‘farming as a service’ – with sophisticated contractors managing much of the work remotely – or do you think the technology will improve the decision making of farmers and enable them to offer new services/ use land more efficiently?

I think both will feature heavily in future and they will interact with each other in complex ways making it impossible to know how this will play out (no one knows!). “As a service” should give more time back to farmers and landowners to develop their business and their landscapes in direction they need to go. There are many aspects of farming and wider land management that don’t get the attention they should do right now.

We need to restore our soils, we need well-balanced landscape management, we need a broader societal appreciation of agriculture and rural landscapes. That said, the minute we develop our agricultural practices in a more sustainable, more well rounded, more diverse direction than can be automated the model for “as a service” changes – there are many more things to be accounted for.

What’s for sure: this is an unknown path we are all on, so it’s best tackled facing forwards.

Matthew Smith, Director of Business Development, Microsoft, will be presenting at the Agri-Tech East Pollinator “The AI’m of machine learning in agriculture” on 20 February 2018 at The Sainsbury Laboratory, 47 Bateman Street, Cambridge, CB2 1LR.

Agri-Tech Archives