Following the ‘digital transformation’ from discourse to performance and practice in the field

Faced with such prospects, it is mainly well-known international bodies such as the World Bank and the European Commission (EC) or philanthropic organizations like the Bill and Melinda Gates Foundation (BMGF) that have set the agenda dominating the funding landscape for the digitalization of (African) agriculture and thereby influence the nature of development programming itself (see Schurmann 2018). Throughout the last years, state agencies have responded to these calls and tailored different programmes such as digital strategies (USAID n.d.) or the SAIS-project (“Scaling digital agriculture innovations through start-ups”) (GIZ 2024). The call for the digitalization of agriculture has also found its way into the African Union’s Digital Transformation Strategy 2020–2030 (African Union n.d.: 45f.). FAO and ITU assessed the state of the digitalization of agriculture in 47 African countries in 2022 while negotiations over the establishment of an ‘International Digital Council for Food and Agriculture’ have taken place since 2019 (FAO 2020, FAO and ITU 2022).

Apart from that, a new generation of African entrepreneurs seeking IT-based solutions for agriculture has also entered the scene. These agripreneurs increasingly rely on capital and resources from the Global North to make their solutions come alive (AI, blockchain, machine learning, precision farming, etc.). Accordingly, Dalberg Advisors and CTA take this into account when stating that ‘new entrants in the D4Ag space – including “big tech” players like Microsoft, Google, IBM, Bosch and Alibaba, as well as ‘big agri’ incumbents like Bayer, Syngenta, Yara, John Deere and UPL – will change the sector’s scale and scope’ (Dalberg Advisors and CTA 2019: 21). Or to put it differently, ‘The entrance of big tech firms will advance the data revolution in new ways’ (ibid.: 145). And indeed, ‘big agri’ and ‘big tech’ are at the forefront of setting the scene for a qualitatively new digitalization wave to come in the African agricultural sector.

The Nigeria-based digital platform Hello Tractor can serve as a prime example of novel alliances between IT solutions ‘made in Africa’ collaborating with Western companies which provide capital, networks, and further resources. Hello Tractor is an IoT platform founded in 2014 on which farmers can rent tractors on demand to cultivate their fields. The idea of renting tractors has become popular because the purchase of costly machinery by individual farmers is often not economical. While the start-up was awarded a grant by Bosch including a three-month participation in its acceleration programme (Onaleye 2019), IBM provided blockchain technology and cloud services to the platform (Dalberg Advisors and CTA 2019: 150). In 2020, the start-up signed a partnership with John Deere – one of the largest global manufacturers in farm machinery and so far, hardly present in markets of the Global South – to test sensor technologies on 400 tractors in Ghana and Kenya (Onaleye 2019). In the attempt to establish a ‘D4Ag solutions landscape’ (Dalberg Advisors and CTA 2019: 32) with regard to agricultural pitfalls, the state, as the formerly central actor, gets increasingly ascribed the role of an ‘enabler’ that is meant to create the right political, fiscal, and regulatory conditions for a ‘digitalization ecosystem’ to thrive (Malabo Montpellier Panel 2019: iv, see also Baumüller and Addom 2020). In this regard, many governments have developed funds to promote agribusinesses. Due to limited financial resources, this largely depends on ‘an emerging community of impact investors. To attract these funds, ICT4Ag entrepreneurs must pitch their products or services in a highly effective and convincing manner’ (CTA 2017: 31), leading us directly to the large conventions we visited.

Nonetheless, all participants are seen as being part of the game, which even though allowing only for a limited number of winners, is supposed to be open, inclusive, and democratic, e.g. through digital live polls and voting mechanisms. Moreover, these events produce moments of (self-)identification. For example, when a forum thread in the online course asked participants to identify as ‘techie or aggie’, it prompted them to position themselves within a given spectrum of expertise that supposedly makes up the field of ICT in agriculture. Moreover, moderators at both conferences asked participants, ‘Hands up, for whom is this the first time at this conference? The second time? The third time?’ This can be read as identifying how experienced one is as a member of the community. By creating such moments, these events contribute to fostering a community of practice that supposedly works towards the same ends with shared basic assumptions and visions (see Johnson 2007 on communities of practices in development).

Through our ethnographic approach, we found that the visions promoted and negotiated by these actors generally reflect the technological solutionism that was also apparent in the written documents. Enthusiastic presentations, for example, about the ‘Ubers of mechanization’, a digital tractor service, celebrated big and successful technology firms. Furthermore, we identified three connected themes that are combined with and further differentiate the broader idea of technological solutionism at the events we visited.

First, there is a business rhetoric, expressed in a focus on the private sector and the purpose of generating predominantly economic results. As a keynote speaker at ICTforAg put it, ‘the challenge isn’t to develop great tech but to translate the tech revolution … into economic benefits’. This leads to a narrative in which all actors involved are seen as businesspeople, including small-scale farmers. Second, there are calls to do things faster, better, bigger, i.e. at a large scale. On the one hand, this applies to the technologies themselves: essential for big data, artificial intelligence, or the Internet of Things is the ability to compute large data sets. On the other hand, scale is understood in terms of reach to large numbers of users. As an interlude prior to the lunch break the moderator at ICTforAg, for example, asked the participants: ‘Hands up, to how many people have you scaled your innovation or product? Up to 1000? 10,000? 100,000? One million? More?’ (Moderator at ICTforAg, 2018).

And finally, these events are characterized by a vision of data-driven agriculture – a vision that is yet to be realized. As an expert in the online course noted discussing a decade-old contribution to a volume published by Microsoft Research (Bell 2009: xii), ‘this really jumped out at me of how [it was written in 2009 that] “in the 21st century the vast volume of scientific data […] is likely to reside forever in a live […] publicly available, curated state […] for continued analysis.” Most institutions […] are nowhere near that vision’ (Speaker at live presentation in online course, 2018).

In order to eventually achieve this vision of data-driven agriculture, the field has to rely on technology experts, data scientists and software developers. Hence, the new development experts are those ‘who have the key to the code’ (Speaker at Big Data for Agriculture Convention, Nairobi 2018) – for example, someone like Gervin, a Dutch technology expert we met at the event in Nairobi, who told us that he has been working on data-based IT solutions for African smallholder farmers for the last 20 years, but has never been to a village on that continent. This already indicates that the farmers not only remain largely absent from the stages of such events, they also hardly play a role in related project designs.

Some of these donor organisations also support the pilot project in the Mbeya region of southern Tanzania, which is part of our case studies. For many years, a local NGO has pursued the goal of providing agronomic support to smallholder farmers, playing a central role as a facilitator for the transfer of Western agricultural knowledge and technologies to trigger the ‘African Green Revolution’ (Shepherd 2006: 399). This included the dissemination of improved seeds, fertilisers, pesticides, and herbicides, as well as training in good agricultural practices. However, when the results still did not meet the targets set, this NGO agreed to test a new digital information service for the rice farmers in the area.

The technology introduced to the trial community basically represents a multi-perspective analysis of the rice fields. Every 10 days, a high-resolution satellite image is taken. The drone takes multispectral images of the rice fields, which had earlier been mapped with an app. Photos and manual descriptions of the crop status supplement the data as ground truth. In addition, there are USSD queries1 USSD, Unstructured Supplementary Service Data, also known as quick codes, is a GSM communications protocol to send text messages from the phone to another device, usually a network or server, e.g. for balance inquiries, mobile money services, different information services, and location-based content services. Due to a real-time connection, queries and answers are nearly instantaneous. for the farmers themselves, about their practices and timing in the fields. The data flows into a cloud and is processed to produce maps that show the status of rice, plant growth, and the water level in a visual form. The final product is an online platform that displays all the fields with colour codes indicating when something goes wrong regarding the growth of the plants. By clicking on the different rice fields, information about the farmer can be accessed if he or she has participated in the survey. In turn, in case the data about the fields indicates nutrient deficiency, disease or pest infestation, or inadequate water levels, the respective farmers are immediately informed via SMS.

The difference the digitization makes regarding the organizational structure of the project becomes visible in Figure 3.1. First, there are the philanthropic funders and facilitators that have also been present at the different conferences and fairs. To work with the technologies, a cluster of technology and data experts comes into the scene, along with SMS service providers. Most of these are not present in the Mbeya region, not even in Tanzania, but work in places providing the necessary data infrastructure (in this case, the Netherlands), complemented by software programmers and modellers sitting in co-working spaces or home offices somewhere in the world. As the technology expert in charge of programming, who is based in Israel, explains, ‘the NGO has all the contacts on the ground from previous projects like the banks, the agribusinesses, the processor and of course to the farmers. And then for the technological part it gets really messy. They are so many, and some are just ghost-companies’ (Interview with technology expert based in Israel, Dar es Salaam, 21.07.19).

This kind of analysis can only work in areas comprised of clearly defined and relatively large fields, a prerequisite for the algorithm to be able to distinguish rice from other plants. Consequently, the rice farmers involved in this project are considered to be wealthy and productive. According to the NGO, for a pilot one needs a trial community that is willing and motivated to learn, and this applies to people who ‘have no fundamental problems in life’ (informal conversation with project staff, May 2019). This allows for a focus, first and foremost, on the engagement between farmer and technology and not on the pursuit of ‘development’ or the provision of needed assistance. This balancing act between testing and development formed a basic element of the project and posed a considerable challenge in terms of coordination between the ‘techies’ and ‘aggies’ in the field as they tried to gain acceptance among the farmers. As the technology expert from Israel explained, ‘in Holland, the technology is actually used by a large agricultural company so that the farmer knows more about his fields and potatoes in real time, can optimize his processes and intervene quickly in case of drought or disease. That is what the technology is good for’ (Interview with technology expert based in Israel, Dar es Salaam, 21.07.19).

What the technology is good for and what it is supposed to achieve in a pilot project in Tanzania seems to differ considerably. The official goal is to provide 125,000 farmers with weather forecast information, agribusiness support, and plot-specific crop advice. On top of this, 400 agribusinesses are to be connected with farmers, with the idea that this would make the value chain more efficient and enable smallholders to make better decisions on the basis of digitally mediated knowledge and thus to become more productive and economical. However, in the Mbeya region, only very few of the small agribusinesses have laptops and Wi-Fi to load the data-heavy platform. Most farmers have no access to it at all. Therefore, SMS services and USSD queries are designed to circumvent these issues of online connectivity.

However, it soon becomes clear that there is a significant difference between what the technology can do and what is actually done with it. At first glance, farmers seem to appreciate the information that is sent to them. As one of the farmers states:

However, the same farmer, later in the interview, admitted that, despite following the advice, the harvest had been bad due to drought, so the money for the pesticides might have been better spent elsewhere. Others tell us that the information and weather forecast they receive via SMS often arrive too late, or that they do not receive them at all as they did not give their correct number to the NGO. As an NGO staff member tells us:

Again, others admit that they do not read the messages because they think they are spam. This is also noted by the extension officer who complains that ‘farmers do not answer the USSD-survey about their practices in their field, because they are afraid that those messages are gambling and gaming messages!’ (Extension Officer of trial community, Mbeya, 16.06.2019).

Being aware of a lack of trust between those bringing the technologies and those selected as the trial community, it is the technology itself that is supposed to solve the issue. As one of the dealers trading in seeds and fertilizers explains:

This reveals how data generated through the technologies is used for disciplining and control. More than ever before, in this new phase of agricultural development, farmers seem to be pushed into the background, serving mainly as data sources. The same holds true for most of the NGO staff. From time to time an IT specialist comes and teaches both NGO workers and farmers the necessary applications and services to collect data. For this purpose, he leaves lots of smartphones and tablets on site. But even if the technologies are promoted as fast and simple, in practice they are not. The data collection often proves tedious and time-consuming, as it involves mapping entire areas and entering repetitive and mundane data, sometimes twice or even three times if the cloud service is constantly interrupted by a poor internet connection. At best, at some point the data will be sufficient to refine the algorithm.

While, at least at this point, the pilot project mainly serves the development of the technologies themselves, for the farmers, pilots come and go and much appears to stay the same.