Agentic SDLC for software groups & holdings
A group with several software companies has the same methodology problem in each one, and a second problem on top: getting one consistent way of working without running the same programme from scratch five times.
Prove it once, then template it
Land in a single company or team on one real product, lock a baseline, and let its trained leads carry the method. Once it holds, the same playbook rolls out to the next entity, faster and cheaper each time, because you are reusing a proven setup rather than rediscovering it.
Trying to transform every company at once is how group programmes stall. One lighthouse, then expansion, keeps the risk small and the evidence in front of you.
One standard, not five dialects
Shared tooling choices, eval gates, guardrails and KPIs sit in a governed central layer. The glossary, repo config and local conventions stay with each team. No single team quietly redirects group-wide standards by editing a local file, and no entity is forced into a workflow that belongs to another team.
A company already on Cursor keeps Cursor; one on Claude Code keeps that. The methodology is harness-agnostic, so the standard travels without a tooling fight.
Group-level visibility
A consistent before-and-after measure across entities, complexity-adjusted velocity alongside the DORA keys, quality and adoption, gives the group CTO one read of where each company stands and where the next investment pays off. Comparable numbers, not a pile of incompatible dashboards.
How it runs
A diagnostic in the lighthouse company, a Code Factory Build to prove it, then a per-entity rollout priced per entity. Local champions trained in each one, so the method travels without me on a plane to every site.
Common questions
Do all our companies have to use the same tools?
No. The standards that matter (eval gates, review, the spec flow, KPIs) are shared and governed centrally; tool choice and local conventions stay with each team. The methodology is harness-agnostic, so a company on Cursor and one on Claude Code both fit the same standard.
How fast can we roll it across the group?
Start with one lighthouse company and prove it, then expand entity by entity. Each rollout reuses the proven playbook, so cost and time per entity drop. Transforming every company at once is how group programmes stall.
What does it cost across a group?
The first company runs as a diagnostic plus a Code Factory Build (from CHF 40k, scoped by head-count). Each additional entity is a Group Rollout from CHF 30k per entity, lower than the first because the playbook is templated. A one-day strategy session to scope the group approach starts at CHF 3k.
Start with one call, then a diagnostic or a scoping.
We go through where your R&D org stands, which codebases are ready, and whether the programme or a single module is the right first step.