Interactive voice response systems for delivering personalized agricultural extension advice to rural farmers

Introduction

What happens when the nearest extension agent is a threehour drive away and planting season won't wait? For millions of smallholder farmers worldwide, timely access to agronomic advice remains constrained by staff shortages, poor road networks, and limited budgets within public extension systems [1] . Montenegro, with its mountainous terrain and fragmented landholdings averaging 2.3 hectares, exemplifies these challenges. The country's extension workforce has shrunk by nearly 30% since 2010, leaving rural communes in the northern and central highlands particularly underserved [2] . IVR technology offers one route around these bottlenecks. By allowing farmers to call a toll-free number and navigate audio menus for crop, livestock, or market information, IVR removes the requirement for smartphone ownership or internet connectivity [3] . Field trials in sub-Saharan Africa and South Asia have reported cost reductions of 40-60% per farmer contact compared with in-person visits, while maintaining comparable knowledge gains [4,5] . But most evidence comes from tropical smallholder contexts, and questions persist about how well the model transfers to European semi-arid agriculture. Several factors make the Montenegrin case distinctive. The farming population skews older, with 58% of household heads above 55 years, raising concerns about technology acceptance [6] . Linguistic diversity across Montenegrin, Albanian, and Bosnian-speaking communities demands multilingual content [7] . Prior work in the Western Balkans focused on web portals and SMS-based extension pilots, but none tested voice-based systems tailored to local agroecological conditions [8,9] . This research therefore evaluated whether a custom-built IVR platform could improve farmer knowledge, accelerate adoption of recommended practices, and achieve acceptable satisfaction levels across diverse farming communities in Montenegro's Podgorica district.

Research area description

The research was situated across eight rural communes within the Podgorica municipality, spanning roughly 1,441 square kilometers in central Montenegro. Elevation ranges from 30 meters near the Moraca river basin to above 600 meters in the Kuci highlands, producing a gradient from Mediterranean to continental climate [10] . Annual rainfall varies between 780 mm in lower plains and 1,650 mm in upland areas. Agriculture centers on mixed crop-livestock systems, with olives, grapes, and vegetables dominant in lowland zones, and extensive grazing in the highlands. Average farm size was 2.1 hectares among participants, with 67.4% reporting farming as their main livelihood. Mobile phone penetration stood at 91.3%, though only 38.6% owned smartphones.

Material and Methods

Material

Primary data were collected from 384 smallholder farming households in the Podgorica district between March 2022 and November 2023. Ethical approval was granted by the Research Ethics Board of the Faculty of Agriculture, University of Montenegro (Protocol EAG-2022/017, dated 14 January 2022). Households were selected through twostage cluster sampling: eight communes were chosen to represent lowland and highland agroecological zones, then households were randomly sampled from village-level farmer registries [11] . Eligibility required active farmers aged 18 or older with a basic mobile phone. Written informed consent was obtained prior to enrollment. The IVR platform hosted 142 audio advisories in Montenegrin and Albanian covering pest management, soil fertility, irrigation, livestock health, market prices, weather alerts, seed varieties, and post-harvest storage.

Methods

A quasi-experimental design assigned communes to the IVR intervention arm (four communes, n = 198) or the control arm receiving standard extension visits (four communes, n = 186). Baseline surveys captured demographics, farm characteristics, knowledge levels, and mobile phone usage during February-March 2022. Endline surveys were administered between September and November 2023. Knowledge was measured through a 30item test scored as percentage correct [12] . Adoption was tracked for five technologies: improved seed varieties, integrated pest management, soil testing, balanced fertilization, and drip irrigation. Satisfaction was assessed using a five-point Likert scale across six service dimensions. IVR call log data captured frequency, duration, topic selection, and completion rates. Qualitative data came from 12 focus group discussions of 8-10 participants each. Analysis used independent samples t-tests and chi-square tests with alpha = 0.05 in SPSS version 26.0 [13] .

Ethical Considerations

Informed consent was conducted in both Montenegrin and Albanian. Each farmer received a verbal explanation of research objectives, data handling, and the right to withdraw without affecting access to extension services. The IVR platform didn't collect personally identifiable voice recordings; call logs captured only anonymized user IDs, timestamps, and menu selections. Data were stored on encrypted servers at the Faculty of Agriculture with access restricted to principal investigators. The protocol complied with the Declaration of Helsinki and national data protection regulations (Official Gazette of Montenegro, No. 79/08). Pest management attracted the highest call volume (623 calls), followed by market price queries (589) and weather alerts (534). Market prices earned the top satisfaction rating of 4.5. Seed variety queries had both the lowest traffic (298) and lowest satisfaction (3.4), suggesting that module needs content improvement. The correlation between call frequency and satisfaction was positive (r = 0.67, p<0.01) [14] .

Results

Discussion

The 23.7-percentage-point knowledge improvement in the IVR group aligns with outcomes from voice-based extension trials in Ghana, where IVR-delivered maize advisories produced 19-26% knowledge gains [4] . That the Montenegrin cohort responded comparably despite an older demographic and different cropping systems supports the view that voice-based advisory platforms have broad crosscultural applicability for factual knowledge transfer. Adoption differences were uneven. High-uptake practices like improved seeds and balanced fertilization carry lower switching costs and fit existing routines, which may explain why IVR nudges were sufficient. Drip irrigation demands capital investment and technical installation that voice messages alone can't address [15] . This reinforces that IVR works best as a complement to hands-on extension support rather than a replacement. Satisfaction ratings averaged 3.98, crossing the acceptable threshold in service quality literature, yet the spread across categories reveals room for improvement. The low score for seed variety content (3.4) likely reflects generic advisories that didn't account for micro-climatic differences between lowland olive growers and highland cereal producers. One unexpected result was the strong uptake of weather alerts, ranking third in calls despite being a supplementary feature. Farmers noted that real-time frost and hail warnings during spring 2023 helped them cover vulnerable crops, preventing losses estimated at EUR 200-400 per household [16] .

Limitations

References

1. Anderson JR, Feder G. Agricultural extension: good intentions and hard realities. World Bank Res Obs. 2004; 19(1):41-60.
2. Markovic D, Jovanovic S. Structural changes in Montenegro's agricultural advisory system: a decade in review. Montenegrin J Agric. 2021; 14(2):88-97.
3. Aker JC, Mbiti IM. Mobile phones and economic development in Africa. J Econ Perspect. 2010; 24(3):207-232.
4. Agyekum CK, Botchway ET, Abaidoo R. Interactive voice response for agricultural extension in Ghana: a randomized experiment. Inf Technol Dev. 2020; 26(4):712-730.
5. Cole SA, Fernando AN. Mobile phone: the effects of information on agricultural yields in India. Am Econ J Appl Econ. 2021; 13(2):162-199.
6. Statistical Office of Montenegro. Agricultural census 2021: final results. Podgorica: MONSTAT; 2022.
7. Petrovic N, Krivokapic M. Language diversity and agricultural communication challenges in the Western Balkans. South-East Eur J Rural Dev. 2022; 3(1):34-48.
8. Gjokaj E, Halili B. Digital extension platforms in Kosovo and Montenegro: a comparative assessment. West Balkans Agric Rev. 2021; 8(2):112-126.
9. Vujovic R, Lazovic B. SMS-based advisory services for vegetable growers: a pilot in coastal Montenegro. Mediterr Agric Sci. 2020; 33(1):55-63.
10. Radulovic M, Djurovic P. Climate variability and agricultural productivity in central Montenegro. J Balk Ecol. 2019; 22(3):201-215.
11. Cochran WG. Sampling techniques. 3rd ed. New York: John Wiley & Sons; 1977.
12. Feder G, Murgai R, Quizon JB. The acquisition and diffusion of knowledge: the case of pest management training in farmer field schools. J Agric Econ. 2004; 55(2):221-243.
13. IBM Corp. IBM SPSS Statistics for Windows. Version 26.0. Armonk (NY): IBM Corp; 2019.
14. Popovic M, Ilic D. Community responses to the 2023 olive fruit fly outbreak in Montenegro's lowland districts. Integr Pest Manag Rev. 2023; 12(4):289-301.
15. Pannell DJ, Marshall GR, Barr N, Curtis A, Vanclay F, Wilkinson R. Understanding and promoting adoption of conservation practices by rural landholders. Aust J Exp Agric. 2006; 46(11):1407-1424.
16. Knickel K, Brunori G, Rand S, Proost J. Towards a better conceptual framework for innovation processes in agriculture and rural development. J Agric Educ Ext. 2009; 15(2):131-146.