← Back to case studies
Live Service

Home Office MBTP: from product team to converged live service

Building Mastek’s application management capability for the Home Office — moving a development team into live service, then growing it into the service other Atlas products transitioned onto.

RoleService Architect, then Delivery Lead
EnvironmentHome Office Migration and Borders Technology Portfolio
FocusRun/Change operating model and live service
Cost26% savings over the first two years
Live serviceIncident backlog cut by 90%
Automation60% faster caseworking process

Context

Mastek built the Biometric Residence Permit (BRP) Fulfilment platform for the Home Office in line with the GDS Service Standard, from discovery to live. By the time I joined, the platform was live and handling millions of transactions a year, and more UKVI products were joining it. Mastek, though, was almost entirely an application development business with no application management experience. I was brought in to change that.

The challenge

A delivery model designed for active development was now also running a live service. The team worked across two worlds — the MBTP programme environment and DDaT Enterprise Services’ live environment — with no model joining them up. The result was friction, unclear accountability between Run and Change work, and a ceiling on service quality. The client wanted more for less, so the new model had to be commercially sound as well as operationally safe.

What I did

I moved the team, incrementally and without interrupting service, to a hybrid DevOps Run/Change model. Teams were reorganised across products, with shared DevOps and QAT engineer pools instead of product silos. DDaT’s ITIL-based incident, problem, change and release processes were adapted to the hybrid model rather than applied wholesale. Run moved onto ServiceNow while Change stayed on Jira, Confluence and Slack, and the Change Advisory Board became a backlog-grooming forum that put business, architecture, development and operations in one conversation.

As Delivery Lead, I then converged the team into a Level 3 support and continual-improvement service for the wider portfolio, onto which Atlas products from other suppliers transitioned. I stabilised Atlas early life support through problem management, and delivered the portfolio’s first robotic process automation solution, for the Asylum Accommodation Support Transition programme.

Outcome

The Run/Change model delivered 26% cost savings over the first two years and a 35% lower run-rate from Year 2. About 80% of incidents were resolved at an earlier tier, incident volumes fell by around half, and deployments dropped from 7–10 hours to 1–2. On Atlas, the incident backlog fell from 100+ to single digits (90%) and aged incidents from 18 months to one. That cleared the way to migrate the legacy ASYS system onto Atlas Asylum Support and decommission it, saving £3.7M a year from that one system alone, within the retirement of a far larger legacy estate that included CID. The RPA solution sped up the process by 60% and avoided recruiting and training 12 caseworkers. Customer satisfaction stayed green throughout.

Why it mattered

  • Built an application management capability the supplier didn’t have, and made it the base for seven years of account growth.
  • Joined up Agile delivery and ITIL live service in one model, with the economics to match.
  • Took new technology from idea into live government service.