Agricultural Horizons: Research, Technology and Sustainable Development Vol. 5

Digital agricultural advisory services are frequently assessed through their capacity to distribute information, although farmers require support for decisions shaped by uncertain weather, heterogeneous production conditions and unequal access to resources. This critical narrative review examines when digital communication becomes effective, inclusive and accountable decision support in the Global South. Peer-reviewed experimental, observational, qualitative and review evidence was selected through iterative scholarly searching, supplemented by institutional implementation studies. The synthesis distinguishes communication reach, learning, practice change, agricultural performance and farmer agency rather than treating these outcomes as interchangeable.

Randomised evaluations demonstrate that digital services can improve knowledge and selected management practices, but production and economic benefits are less consistent. Positive pooled effects coexist with well-designed studies reporting no clear yield, profit or sale-price improvement. This variation reflects differences in the decisions addressed, counterfactual services, resource constraints, delivery arrangements and outcome measurement. Personalised recommendations can improve management without increasing total input use, while information on uncertainty and opportunities for dialogue may be as consequential as technical precision. Facilitated video, voice services and agent-mediated tools show that the relevant intervention is often a combined digital and organisational service rather than an application alone. Yet human mediation does not automatically ensure effectiveness or inclusion.

Farmer-centredness requires influence over advisory objectives, understandable alternatives, feasible actions, meaningful control of data and opportunities to question recommendations. Gender-responsive targeting, accessible interfaces and institutional capacity are therefore substantive components of decision support, not peripheral implementation concerns. Climate services and generative artificial intelligence extend the range of possible assistance but introduce additional requirements for calibration, local validation, professional oversight and accountability. A proposed decision-support framework connects goal definition, contextual diagnosis, comparison of alternatives, supported action and feedback. The strongest defensible conclusion is that digital advisory services should be evaluated by the quality and distribution of decisions they enable, with agricultural and livelihood benefits demonstrated rather than inferred from access, engagement or technological sophistication.