Harnessing AI for enhanced churn prediction in agriculture – 2024 perspective

Implementation

Manufacturing

USA

Dominik Zuk, CFOO & Co-Founder
Dominik Zuk

Co-Founder at Innsyte

Author’s historical experience before founding Innsyte
Global leader in agriculture with over +50 bn USD in yearly revenue conducted commercial analytics transformation achieving significant precision in hard and soft churn prediction – Dom was working on the EU side of business.
Data accuracy: 98%
AI precision (AUC): 72%
MVP available: in 5 months
Business impact:
Improved client retention

Problemusing AI to retain clients


The opportunity to reduce churn is immense. McKinsey reports that adopting an omnichannel strategy backed by data analytics can dramatically lower churn rates. For example, an agricultural inputs distributor saw over 60% of customer revenues shift to its digital portal within six months, reducing churn by threefold for digitally engaged customers compared to non-engaged ones. This effort generated over a billion dollars in revenue through the digital platform.
AI is also proving to be a powerful tool in churn prevention. Companies like Johnson Controls have leveraged AI to protect over $100 million in annual revenue by identifying at-risk customers and implementing targeted retention strategies.
The project Dom was part of began with a critical realisation: the client was losing around 10% of annual sales revenue in specific locations due to customer churn. This highlighted the urgent need for a strategic solution, as well as the untapped potential of using AI to foster business continuity with key B2B clients. Key challenges:
  • Lack of visibility into cross-sell opportunities and segmentation of B2B clients
  • Inability to predict churn effectively, leading to consistent revenue losses
  • Technical debt that slowed AI adoption and hindered progress
  • Fragmented data landscape, with information spread across multiple systems, complicating product management and analytics
Despite these challenges, we laid the groundwork for success through a standardized approach to model building, robust MLOps practices, and a commitment to building trust in AI among account executives. Expected solution? A scalable analytics solution and the foundation for significant revenue growth through smarter, AI-powered customer retention strategies​​​.

Solutiondeveloping churn prediction capabilities


The core of the solution centered on deploying a robust cloud infrastructure powered by technologies like Azure, Azure Synapse, Azure Databricks, Machine Learning, Tableau, Salesforce, and SAP. This infrastructure enabled the development of precise, country-specific, and industry-specific AI models capable of predicting client churn with high accuracy. Key solution components:
  • Salesforce integration: seamlessly integrated with Salesforce CRM, enabling a user-friendly decision-making process.
  • Advanced ML models: deployed gradient-based machine learning models for accurate churn prediction.
  • Business calibration & pilots: conducted pilot projects and business calibration across seven countries to fine-tune the solution.
  • Azure-native ML solution: The AI models, pulling data from SAP, Salesforce, customer call centers, and flat files, marked a major step forward. Initially developed using Jupyter Notebooks and Python, the solution was later enhanced through Azure Databricks and MLOps for improved efficiency and scalability.
This approach ensured the solution was not only powerful but also adaptable across regions and industries​​.

Impact: proactive, AI-powered clients experience


The implementation of this solution has set a new benchmark for proactive client management in the agriculture sector. By accurately predicting churn, the company can now act earlier, providing tailored solutions to at-risk clients and significantly improving the client experience. This not only strengthens client retention but also uncovers cross-sell opportunities, driving revenue growth. Key areas of impact:
  • Proactive client retention: by accurately predicting churn, the solution allows early intervention with tailored strategies, reducing client attrition and unlocking new cross-sell opportunities, ultimately driving significant revenue growth.
  • Scalable data-driven framework: the project highlights the power of a scalable, well-integrated data ecosystem, capable of being adapted across different regions and business models, setting a new standard for AI-driven client management in the agriculture sector.

With AI-powered churn prediction, companies can move beyond conventional strategies to deliver a personalised, proactive approach to client retention. This commitment to innovation ensures that every client interaction is informed by real-time insights, enhancing their experience and solidifying long-term loyalty.
Dominik Zuk, Co-Founder at Innsyte
Dominik Zuk, CFOO & Co-Founder

Lessons learnt from churn prediction with AI


In conclusion, the journey of churn prediction from identifying a critical business challenge to implementing a sophisticated AI-driven solution exemplifies the transformative potential of technology in the agriculture sector. As we look to the future, the lessons learned and the success achieved in this project serve as a beacon for other companies facing similar challenges, demonstrating that with the right approach, AI can indeed be a powerful tool in not only predicting client churn but in reshaping the entire landscape of client relationship management.
High qualify of data is key to success
A significant portion of the engagement was dedicated to understanding the business, explaining the data, and integrating it effectively. High-quality data proved essential to the project’s success, laying the foundation for accurate insights and impactful outcomes.
Always plan pilot phase in the engagement
The solution had a profound impact on the day-to-day operations of account executives. Their active involvement was crucial in building trust and ensuring the solution aligned with both business needs and operational workflows.
Involve Sales Operations team from Day 1
The definition of churn varies between companies, making it essential to involve ‘churn champions’ to establish a shared understanding between consultants and the client. This collaboration ensured clarity and alignment on critical metrics.
Start small and expand as value come
The proof-of-value stage was pivotal in the AI adoption journey. After demonstrating the solution’s value and running successful pilot sessions with account executives, the organization gained the confidence to fully adopt the MVP, paving the way for broader implementation.

Check how this project could look like in your company.

In the meeting, we will ask targeted questions about your specific business situation to assess how our solutions align with your objectives.

Line-access-ai

Read similar case studies to this one