From layered searches to intelligent insights, AI is helping businesses discover, interpret, and act on data more effectively.
Customer
Our Customer is a UK-based construction market intelligence service provider, catering to a wide range of users in the Built Environment sector including building material suppliers, labour contractors, consultants, architects, designers, electricians, plumbers, and more.
The Challenge
The customer has traditionally delivered online market intelligence through multi-layered search queries. This process required significant user interaction and navigation through a complex interface, making it difficult for users to access the full value of the system.
Solution
The client aimed to enhance its services by becoming one of the first providers to offer AI-enabled market intelligence. The use cases identified were:
- Automated Data Presentation: Extract and present data from structured, semi-structured, scanned, and even handwritten business documents with minimal manual intervention.
- Natural Language Search: Enable less technically experienced users to perform context aware searches in natural language and receive accurate results quickly.
- Document Summarisation: Generate concise summaries of large tender and planning documents to help users make faster, informed decisions.
- Intelligent Communication: Quickly identify and trace business opportunities via email, supported by relevant context to improve productivity and communication.
The client operated multiple database instances with several service layers built over decades using diverse technology stacks. Addressing this requirement required multiple tools, techniques, and efficient integration between components. Cost was another important consideration, particularly for pay-as-you-use services and cloud subscriptions.
Arrk recommended an iterative PoC strategy for rapid evaluation. During PoCs, we primarily leveraged NLP-based OpenAI models. Different business requirements called for different tools and approaches. Ultimately, we selected OpenAI, Mistral, and models available on AWS Bedrock (including LLaMA and Claude). These foundation models were selected and configured to support specific use cases. The adopted technology stack included:
- Python Flask API deployed on AWS ECS.
- pgvector as the vector database.
- Marimo for dashboarding and visualisation.
During the project, we encountered several challenges that needed to be addressed:
- Data Extraction from Scanned & Handwritten Forms: Extracting data from scanned and handwritten application forms presented an early challenge. We integrated MistralOCR after evaluating multiple methods, significantly improving text recognition accuracy.
- Processing Large Document Volumes: We encountered gateway timeout issues during large scale document processing. To address this, we implemented an asynchronous ingestion architecture using AWS Batch, enabling reliable and scalable processing.
As with any AI-based solution, performance improved through ongoing refinement and training. We engaged in-house subject matter experts to test and use the system extensively, helping to improve overall outcomes.
Key Outcomes
- Reduced manual operational effort by automating repetitive, time-consuming tasks.
- Significantly enhanced user experience and interaction quality.
- Reduced deployment times, with new AI use cases now going live within days rather than weeks.
- Improved the consistency, accuracy and reliability of responses through contextual data grounding and workflow orchestration.
- Established a scalable AI platform capable of supporting future business requirements
Conclusion
In today’s digital landscape, the ability to adopt new technologies quickly can provide a significant advantage. Working with Arrk, this customer achieved their ambition of bringing AI-enabled market intelligence to market and established an enterprise grade AI platform to support future growth. This investment has improved engagement with the customer platform and is expected to contribute to higher subscription renewal rates and improved customer retention.
Similar Case Studies
Discover how Arrk Group used AI tools to modernise and redevelop a customer’s legacy systems.
Arrk helped the customer improve data sourcing, harmonisation and operational efficiency using AWS SageMaker.
Discover how Arrk Group used AI-assisted tools to modernise a Customer’s legacy application, supporting a faster and more efficient delivery process.
Arrk provides end-to-end support for complex digital environments, helping organisations maintain, modernise and optimise business-critical systems.
Arrk Group has extensive experience delivering digital solutions across complex supply chain environments.
A London local authority partnered with Arrk to design and deliver a new digital waste and recycling service, supporting more than 225,000 residents.
Arrk implemented an AI-powered recommender system to improve customer engagement and support revenue growth.
Our client aimed to increase public participation in charitable donations by covering transaction costs and contributing additional funds through its corporate trust.
Arrk helped the Customer improve the speed, accuracy and scalability of market data collection through API-driven data extraction.
A leading transatlantic airline partnered with us and our strategic delivery partner to modernise its core customer‑facing services, including check‑in, seat reservations, complaints handling, and call centre operations.
