The visual for using ICTs in agriculture is becoming increasingly granular - any farmer (in particular smallholders) can have mobile access to detailed, precise information about their farm, and receive specific advise on how to maximize yields. This has the potential to increase yields and production comparable to what the green revolution did in many parts of the world.
With that vision, the obstacles to achieving it are also starting to become known. An important part of it is how to 'chop up' services and information that needs certain agricultural scale in order to be feasible. Creating business models that allow for smaller farms to benefit from high-tech, high investment agri services is a critical part of the puzzle.
If it's relatively easy to identify this as a problem, finding solutions that actually work is a lot harder. The agri world has been working for decades on ways of grouping small farmers together. The current state of most cooperatives does not suggest that this has been cracked yet. Another clear obstacle is finding ways to motivate small farmers to pay for yield-enhancing services. While some countries have made great strides in this area, significant parts of the developing world still work on a 'minimal cash outlay' farming system.
It goes too far - and is probably too early - to start promoting definitive solutions in this area. IFC is working on a number of topics in ICT for agri, which cover traceability, farmer extension, benchmarking and GIS/precision ag. As so often, the real question of success may be shaped more by who is most successful in getting their services adopted, rather than who has the technically most elegant solution.