Enhancing Contract Management through Natural Language Processing(NLP): A Case Study of Three African Countries

Abstract

This study explores how Natural Language Processing (NLP) has transformed contract management, enabling businesses to interpret government contracts more accurately and quickly. The research involved extracting data from African public procurement portals in three African countries, which are Morocco, Ghana, and Benin, and applying NLP techniques for language translation and keyword extraction. The results demonstrate the potential of NLP in transforming contract management, providing valuable information for informed decision-making, and promoting transparency and accountability in government spending. This study serves as a practical demonstration of NLP's application in contract management and suggests future research opportunities for expanding its use in other African countries

Full text

Enhancing Contract Management through Natural
Language Processing(NLP): A Case Study of Three
African Countries
Ibukun David Babatunde*, Arinze Ezirim, Bid Oscar Hountondji
School of Collective Intelligence (SCI)
University of Mohammed VI Polytechnic (UM6P), Rabat, Morocco.
david.babatunde@um6p.ma, arinze.ezirim@um6p.ma, oscar.hountondji@um6p.ma
Abstract
This study explores how Natural Language Processing (NLP) has transformed
contract management, enabling businesses to interpret government contracts more
accurately and quickly. The research involved extracting data from African public
procurement portals of three African countries which are Morocco, Ghana, and
Benin, and applying NLP techniques for language translation and keyword ex-
traction. The results demonstrate the potential of NLP in transforming contract
management, providing valuable information for informed decision-making, and
promoting transparency and accountability with government spending. This study
serves as a practical demonstration of NLP’s application in contract management
and suggests future research opportunities for expanding its use in other African
countries
1 Introduction
1.1 Background and significance of contract management
Contract management plays a vital role in various sectors, including government procurement. It
involves the administration, negotiation, and monitoring of contracts to ensure compliance with
terms and conditions, mitigate risks, and optimize outcomes. Government contracts, in particular,
are critical as they involve taxpayer funds and impact the delivery of public services and projects.
Government contract management involves complex processes and challenges. Government agencies
must navigate stringent regulations, adhere to transparency and accountability standards, and manage
large volumes of contracts with diverse stakeholders
1.2
Natural Language Processing’s (NLP) Increasing Contribution to Contract Management
NLP can be integrated with advanced technologies to process unstructured data and improve human-
computer interaction, leading to meaningful outcomes that enhance decision-making and improve
operational efficiency in various industries (Bahja, 2020). Indeed, NLP methods have proven
to be effective for managing unstructured text data, such as maintenance requests in buildings
(Bouabdallaoui et al., 2020). Furthermore, in different sectors like healthcare, NLP has shown its
prowess as an effective technique for detecting a broad range of adverse events in text documents,
outperforming traditional and other automated methods (Melton Hripcsak, 2005). Additionally, the
advent of machine learning techniques within NLP can analyze qualitative data almost instantaneously,
proving valuable for qualitative researchers, guiding their creation of codebooks (Leeson et al., 2019).
Natural Language Processing (NLP) has grown significantly in importance in contract management
5th Deep Learning Indaba Conference (DLI 2023).

over the past few years. Advanced contract analysis and insight extraction capabilities are provided
by NLP technologies, which also streamline contract interpretation and automate time-consuming
human tasks
1.3 Research objective and scope
Mtasigazya (2018) indicated that key problems to contract management and enforcements are
corruption, collusion between local government officials and private companies, and poor monitoring
of private companies. Notably, Heald (2012) argued that the structure of transparency mechanisms
profoundly influences their impact on public policy, affecting not just efficiency, but also equity and
democratic accountability. In this context, this study aims to explore the potential of NLP technologies
in contract management, focusing on government contracts in Morocco, Ghana, and Benin. The
central objective is to enhance accuracy, efficiency, and transparency in managing public contracts
through NLP analysis of data from African public procurement portals. Specifically, the study
endeavors to demonstrate how NLP can facilitate transparency in government spending, possibly
reshaping the efficiency and equity in public policy as suggested by Heald.
2 Methodology
The research methodology was organized systematically, starting with data extraction from public
procurement portals of Morocco, Ghana, and Benin using tailored Python scripts that integrated
web scraping tools like Beautiful Soup and Selenium Webdriver. Once the raw data, consisting of
tenders, was collected, a normalization step was conducted to ensure consistency. Given the bilingual
nature of the dataset, with some tenders being in French, specific Natural Language Processing (NLP)
modules were employed to translate the unstructured contract data into English, emphasizing the use
of sophisticated techniques like sequence-to-sequence models and attention mechanisms. To distill
the essence of these tenders, the Gensim library was deployed, leveraging its capabilities for keyword
extraction from extensive texts, with methods like TF-IDF and Latent Dirichlet Allocation. Following
the extraction process, a series of exploratory data analysis techniques were applied to present the raw
and extracted data into structured insights, including various visualizations that elucidated trends and
patterns in the data. This comprehensive and meticulous approach was designed to ensure both the
accuracy and efficiency of data processing, aligning the research with its core objective of exploring
the transformative potential of NLP in contract management.
3 Results and Findings
In the following subsections are some of the insights generated from the study
3.1
Some analysis of the extracted data from the Moroccan public procurement portal in 2022
Figure 1 shows that 41% of the contract awarded were for Services while Public Work and Furnitures
account for 39% and 20% respectively. While figure 2 shows Rabat ,Casablanca and Marrakech
accounts for the top three locations with contracts awarded.While Province d’Al Hoceima and
Province de Sidi Kacem are least awarded contracts
Figure 1: Entry title Figure 2: Location
2

3.2
Some analysis of the extracted data from the Ghanaian public procurement portal in 2022
Figure 3 shows Bekwai Municipal Hospital has won the most contracts in the health sector in Ghana
while figure 4 shows that Essential medicines accounts for 16.8% of all contracts awarded in Ghana
and it can be seen that the contract entries for Ghana is very competitive hence the lower percentages.
Figure 3: Procuring Entry Figure 4: Entry titles
3.3 Some analysis of the extracted data from the Benin public procurement portal in 2022
Figure 5 shows offers closed in 2022, 89 closed in November, 68 in both July and October.Among
the 9 offers that will be closed in 2023, three will be in May and two in March while figure 6 shows
the main authority having launched the calls for tenders is SBEE, the electricity company in Benin.
In the top 10 we also found the Port and the agency in charge of agricultural mechanization
Figure 5: End month/year Figure 6: Top Contracting authority
3.4 General insights
The offers are mainly launched by medical structures in Ghana, by electrical structures in Benin.In
Morocco, the offers are mostly services. The offers are often launched in the middle of the year: not
at the very beginning nor completely at the end. In Morocco, offers last an average of 139 days. Very
few of the offers launched in 2022 will be closed in 2023
4 Implications and Benefits
The implications and benefits of applying Natural Language Processing (NLP) in contract manage-
ment are manifold. Firstly, NLP has the potential to promote transparency and accountability in
government spending by providing efficient contract analysis and monitoring mechanisms. This
ensures that public funds are utilized judiciously, fostering public trust in government activities.
Secondly, NLP enhances decision-making for both businesses and governments by extracting key
contract information, such as pricing and deliverables, enabling more informed choices. Finally, NLP
reduces information asymmetry, enabling transparent decision-making processes and empowering
stakeholders with comprehensive insights into contract details. These advantages collectively un-
derscore the transformative impact of NLP in contract management, facilitating efficient resource
allocation and enhancing overall contract management practices.
3

5 Practical Applications and Future Directions
The practical applications of Natural Language Processing (NLP) in contract management within
the three African countries (Morocco, Ghana, and Benin) offer valuable insights. NLP can stream-
line contract interpretation, enhance accuracy, and automate key processes, benefiting government
agencies and businesses alike. Moreover, these successful applications pave the way for expanding
NLP adoption in other African countries and diverse sectors. Future research directions may ex-
plore fine-tuning NLP models for specific contract types, optimizing multilingual NLP for diverse
African languages, and integrating NLP with other emerging technologies for comprehensive contract
management solutions. However, potential challenges, such as data privacy concerns, language
complexities, and ensuring model interpretability, require careful consideration to harness NLP’s full
potential in transforming contract management practices across the African continent.
6 Conclusions
In conclusion, this research on the transformative potential of Natural Language Processing (NLP) in
contract management has yielded significant findings and valuable contributions. The study demon-
strated the efficacy of NLP in enhancing contract accuracy, efficiency, and transparency, benefiting
government agencies, businesses, and the public. By automating contract analysis, extracting crucial
information, and enabling informed decision-making, NLP emerges as a powerful tool for improving
contract management practices. The implications of this research extend beyond cost savings and
operational efficiency, emphasizing the promotion of transparency and accountability in government
spending. NLP’s transformative potential in contract management underscores its value in driving
efficient resource allocation and fostering trust in public institutions. As future research explores
advanced NLP applications and addresses potential challenges, the transformative impact of NLP in
contract management is poised to revolutionize governance, business practices, and service delivery
across industries and continents.
References
[1] Bahja, M. (2020). Natural Language Processing Applications in Business. E-Business - Higher
Education and Intelligence Applications. https://doi.org/10.5772/INTECHOPEN.92203.
[2]Bouabdallaoui, Y., Lafhaj, Z., Yim, P., Ducoulombier, L., Bennadji, B. (2020). Natu-
ral Language Processing Model for Managing Maintenance Requests in Buildings. Buildings.
https://doi.org/10.3390/buildings10090160
[3]Melton, G., Hripcsak, G. (2005). Research Paper: Automated Detection of Adverse Events Using
Natural Language Processing of Discharge Summaries. Journal of the American Medical Informatics
Association : JAMIA, 12 4, 448-57. https://doi.org/10.1197/JAMIA.M1794.
[4]Leeson, W., Resnick, A., Alexander, D., Rovers, J. (2019). Natural Language Processing (NLP)
in Qualitative Public Health Research: A Proof of Concept Study. International Journal of Qualitative
Methods, 18. https://doi.org/10.1177/1609406919887021
[5]Mtasigazya, P. (2018). Contract Management in Outsourced Revenue Collection in Local Govern-
ments in Tanzania: Experience from Selected Local Government Councils. Journal of the Institute
for African Studies. https://doi.org/10.31132/2412-5717-2018-45-4-32-51.
[6]Heald, D. (2012). Why is transparency about public expenditure so elusive?. International Review
of Administrative Sciences, 78, 30 - 49. https://doi.org/10.1177/0020852311429931.
4