B2B Software Development
Building an AI-native operating model for a B2B software company
Implementing an agentic infrastructure powered by Anthropic's Claude to integrate AI into a software company's development processes.
Industry
Financial and document workflow platforms, and paperless processes
Challenge
Transform the technology and data organisation to make AI central to the company culture
Result
70% reduction in the time needed to resolve complex engineering tickets
Expertise
Technical audit
LLM configuration
Agentic workflow architecture
MCP implementation
Architectural structuring and change management
Tech Stack
Claude
C#
.Net
Jira
MongoDB
The software company operates in a highly regulated sector where sensitive data must be managed securely. Its ambition was to become a fully AI-native organisation.
Our partnership began in 2023, giving us a shared understanding of the company's technical environment. This close collaboration allowed us to support its technology and data leadership teams in transforming their ways of working within just a few months.
Our dedicated AI architects structured the use of Anthropic's Claude to build a foundation for agentic development. The goal was to create a scalable infrastructure that could support AI adoption across the company's SaaS solutions.
Key challenges
As large language models (LLMs) become more accessible, companies must adapt quickly. The technology and data teams needed to move beyond isolated AI experiments and establish a consistent approach across the organisation.
- Widespread adoption: Turn individual initiatives into a standardised approach that all engineers can adopt.
- Code consistency: Ensure AI tools follow the company's existing coding standards and software architecture across its B2B products.
- Technical governance: Introduce a structured integration process to prevent tool fragmentation and reduce security risks.
Collaborative execution
The transformation plan starts with a core team of experts who test the approach on a specific scope before rolling it out across the organisation.
Novity’s AI experts (FDEs, Product Engineers/Builders and AI Engineers) connected Claude AI to existing work environments, including the codebase, technical documentation and coding standards, using Anthropic’s Model Context Protocol (MCP). This enables direct integration with MongoDB and Jira.
As a certified member of the Claude Partner Network (CPN), Novity developed custom skills tailored to the client’s architecture. The pilot team tests workflows in production, documents results and refines configurations before expanding the approach across the organisation.
Weekly reviews help continuously optimise configurations and ensure the reliability of this new technical foundation.
Measurable impact
The new agentic infrastructure is changing how engineering teams work.
- Faster development: Complex architectural issues that previously took two weeks to resolve now take four days. This has reduced ticket resolution time by approximately two-thirds.
- Standardised practices: The company now has a secure technical foundation across its platforms. This supports consistent LLM use throughout its codebase.
- Cultural transformation: The pilot team demonstrated that the approach works in production, creating a foundation for wider AI-native adoption.
Discover more case studies
Every project is unique, but our commitment remains the same: to co-deliver, de-risk, and move products forward—regardless of context or constraints.
