Let your people build with AI — safely.
Your team builds apps, dashboards and automations with AI coding agents in a secure Azure sandbox. When the business approves a proof of concept, Breathe IT takes over everything needed to run it for real.
Your people have discovered AI coding agents. The moment a proof of concept looks useful, three risks appear.
Shadow IT
Prototypes on personal machines and accounts — often with company data — and nobody accountable for security or operations.
The “now what?” cliff
A good proof of concept gets approved, then stalls for months because nobody owns deployment, compliance, scaling or support.
Governance drag
IT clamps down to stay safe, and AI adoption quietly dies before it ever delivers value.
Breathe AI Launchpad resolves this with a two-speed model: let people innovate at full speed in a safe sandbox, and let Breathe IT industrialise whatever the business chooses to take forward. Innovation never stops; risk never leaks.
Two zones, one approval gate.
Everything left of the gate is fast, free-form and owned by your people. Everything right of it is governed, repeatable and owned by Breathe IT. The business makes exactly one decision: which proofs of concept graduate.
Innovation zone
Employee + AI agents · fast, self-service
- Owner
- Employee + AI agents
- Speed
- Fast, self-service
- Data
- Synthetic / non-sensitive only
- Goal
- Prove the idea
Industrialise & run
Breathe IT · governed, repeatable
- Owner
- Breathe IT
- Speed
- Governed, repeatable
- Data
- Real data, access-controlled
- Goal
- Scale it to the organisation
From idea to production.
Request access
one-time self-service form
Your peopleBuild & test
AI agents + secure sandbox
Your peopleApproval gate
the business approves the PoC
Business decisionIndustrialise
compliance · security · pipelines
Breathe ITOperate & scale
monitoring · support · roll-out
Breathe ITFive phases, step by step.
Onboard the company
Azure landing zone, sandbox blueprint, governance baseline and enablement workshops.
Employee self-service
One request provisions a personal resource group wired to AI coding agents, pre-loaded with tooling.
The approval gate
The designated approver reviews the proof of concept. The single human decision point in the flow.
Industrialise
Compliance, code and security review, data architecture and productionisation.
Deploy & run
UAT + PROD, CI/CD pipelines, roll-out, monitoring, support SLA and lifecycle management.
The employee never has to think about industrialisation or operations. That is the entire point of the service.
Four things, handled end to end.
Safe sandbox
A governed Azure environment per employee, wired to AI coding agents — budget-capped and secure.
One approval gate
You decide which ideas go to production. Nothing else needs your attention.
Production, handled
Compliance, security, data, UAT + PROD, CI/CD, version control and support — all run by Breathe IT.
Predictable cost
A fixed monthly team fee by size, plus a flat fee per app we run in production.
Everything runs inside your own Azure tenant.
No customer data passes through Breathe. Innovation and production are separated into different subscriptions, with shared platform services governing both.
Customer Azure tenant · Entra ID · management groups · policy · cost controls
Innovation subscription
self-service sandbox
- Per-user resource group
- Budget caps and guardrails
- AI agent workspace wired to user identity
Managed subscription
operated by Breathe IT
- UAT and PROD environments
- Git repo + CI/CD
- Monitoring, backup and disaster recovery
Shared platform · identity · Azure Policy · Key Vault · monitoring · cost
Isolation by design
Sandbox work cannot reach production data or systems; promotion only happens through the approval gate and Breathe IT.
Data security
Real data is introduced only in the managed subscription, with encryption, Key Vault secrets and role/row-level access.
Auditability
All access, policy and cost events are logged centrally for compliance reporting.
Platform, every graduated proof of concept, and continuous operations.
Platform & governance
- Azure landing zone and per-user sandbox blueprint
- Azure Policy guardrails, budget caps, tagging and cost reporting
- Identity and access integration (Entra ID) and audit logging
For every graduated PoC
- Compliance and licensing report (APIs, data, libraries)
- Code review, security review and remediation
- Data model, access-control and secrets design
- UAT + PROD, Git repo, CI/CD and roll-out
Continuous operations
- Monitoring, alerting, backup/DR and patching
- Quarterly governance and cost review
- A maintained catalogue of every app in flight and in production
The app suite that runs the whole service.
Breathe IT operates Launchpad as a reusable platform, deployed for every client during onboarding. The same apps your people build in are the apps we govern and run with — which is what keeps delivery fast.
Console
A single portal for builders, approvers and Breathe IT.
Onboard
Self-service onboarding — provisions the sandbox, identity and AI agent workspace in minutes.
Gate
PoC request and approval workflow with a full audit trail; approval triggers industrialisation.
Assure
Automated compliance, licence and security scanning — the reports reviewers used to write by hand.
Foundry
The environment and pipeline factory: UAT, PROD, Git and CI/CD from blueprints.
Catalog
System of record: every app with owner, cost, status and data classification.
Every developer also gets a builder starter kit — opinionated, guard-railed repository templates — so proofs of concept arrive already production-shaped.
Keeping it running is included. New work is billed by the hour.
Once an app is managed, operation is included in the monthly fee. New development is billed by the hour — but because managed apps already live in the Launchpad framework, documented and with pipelines we know, each billed hour delivers considerably more than the same work greenfield. Every change is estimated and approved first.
Included in the monthly fee
- Operation of UAT + PROD
- CI/CD pipeline, Git repo and version control
- Monitoring, alerting and uptime
- Backups and disaster recovery
- Security patching and dependency updates
- Bug fixes — corrective maintenance
- Minor changes — config, content, tweaks
- Compliance, logging and access management
- Support desk with SLA and quarterly review
Billed by the hour
- New features
- New integrations
- Major rewrites
Simple and predictable.
Price follows two things: how many people will build, and how many apps we run in production. No per-user licence for the people who merely use the apps.
Fixed monthly fee by team size
Includes the secure sandbox, the Launchpad apps, governance and support. Scales in steps from a couple of builders to whole departments.
Flat fee per app in production
You pick the service level per app — from a one-time safe deploy to a fully managed, business-critical service.
One-time onboarding
Landing zone, sandbox blueprint and workshops. Scaled to team size, so a small team can start small.
Azure consumption and AI coding-agent licences are passed through at cost. Fully bundled pricing available on request.
Request the price list →From invisible risk to a governed portfolio.
This is the pattern at an IT manager in a mid-sized business where AI use has grown faster than governance.
apps built outside IT, with no owner or documentation
backlog before IT could deliver a new internal tool
visibility into what data those apps could reach
| Before | With Launchpad | |
|---|---|---|
| Visibility | None | Full — one Console |
| Who maintains the apps | Nobody / IT by default | Breathe IT |
| Where data lives | Unknown / external | The client's own Azure tenant |
| IT's role | Blocker / firefighter | Governor and enabler |
Representative scenario — a composite of patterns we see repeatedly in mid-market IT teams adopting AI. The figures are illustrative and do not describe a single named client.
Why this works
- AI adoption keeps accelerating — no clamp-down, because the sandbox is safe by design.
- One predictable gate — leadership controls what goes to production, not how.
- No “now what?” cliff — graduation to production is a known, managed path.
- Compliance and security are built in, not bolted on after an incident.
- Costs are visible and capped at every stage.
Does this sound familiar?
- People are building with AI faster than you can govern it
- You can't say where company data is going
- Your backlog grows no matter how you staff it
- You inherit apps nobody documented or owns
Half a day to scope. Builders live the same week.
Half-day workshop to confirm scope and governance.
Landing zone, sandbox blueprint and enablement — 3–5 weeks.
The first builders go live and start producing proofs of concept.
Let's map your situation in 30 minutes.
We'll walk through the AI activity you already have, find the apps that are worth governing, and show you exactly how Launchpad would run them — inside your Azure, on your terms.
Book a working session →