Client story Accelerating Complex Configuration Migration
CLIENT STORY
Accelerating complex configuration migration with targeted AI tools
Client
A leading UK software and IT services provider needed to accelerate a highly complex configuration management migration for more than 170 customers. Its semi-automated, consultant-driven process was projected to take seven years, creating significant cost and resourcing pressures. The organisation worked with Insight AI to streamline the previously manual steps.
Industry: Software and IT services
Challenge
The client faced long timelines and high costs to migrate more than 170 customers to target configurations using a semi-automated, labour-intensive process. Each customer migration typically took around 16 days, and projections indicated the full programme could take seven years, which the client considered unacceptable. Its complex legacy configuration management system, built in-house, was difficult to analyse and automate, and existing tooling only partially supported the work.
Outcomes
Insight AI delivered a set of bespoke Windows and AI-driven tools, backed by a retrieval-augmented generation database, to streamline previously manual steps and make migration consultants’ work easier. The tools enable faster progression towards a more easily supported end state.
Insight AI was chosen for its analytical skills, AI expertise and a track record the client already trusted.
The client already knew Insight AI from previous work and trusted its reliability and capability. Insight AI demonstrated an ability to quickly ingest large amounts of information, make sense of complex legacy systems, and think creatively to suggest candidate solutions with a clear commercial rationale. Its capacity to deliver prototypes and proofs of concept, draw on wider specialist resources when needed, and operate as an independent consultancy made it well suited to an organisation-wide configuration migration challenge under regulatory constraints.
Our solution
Insight AI acted as a consultancy and discovery partner, operating as an independent delivery unit alongside the client’s migration consultants. It began by listening to the consultants performing the work, then analysed the configuration management landscape in depth, including the complex legacy and home-grown systems. Multiple candidate solutions were proposed for different parts of the migration process, triaged with the client, and prioritised by their potential return on investment in terms of time reduction.
Key highlights:
- Approach: A structured listen-analyse-propose-agree-build-iterate approach, rather than a formal named agile methodology
- Tools: Three bespoke Windows command line or Explorer-triggered tools and two AI-driven tools, designed as point solutions to plug previously manual steps in a multi-step, human-driven process
- Technology: Python and C#, TypeScript and other frontend technologies, PostgreSQL and AI/ML engineering, with a retrieval-augmented generation database backing a copilot for image and document processing
- Delivery: Deployment support and iteration with migration consultants to reach production quality, keeping the solution tailored to the client’s unique environment rather than relying on off-the-shelf products
Benefits and outcomes
The tools gave the client practical means to move faster on a migration programme that was projected to take seven years.
- Operational efficiency: Bespoke Windows and AI-driven tools replaced several previously manual steps, making consultants’ work easier and reducing labour intensity in the semi-automated process
- Strategic impact: The solution directly addressed a programme projected to take seven years, helping the client accelerate progress across more than 170 customers
- AI-enabled analysis: A RAG-backed copilot and AI research capabilities allowed rapid ingestion and understanding of an extremely complex legacy configuration landscape
- Innovation in tooling: Multiple proofs of concept, both AI and non-AI, demonstrated how LLMs and targeted utilities could deliver commercial value in configuration comparison and data management
By / 5 Nov 2026 / Topics: Application development
