Step 02 · Teardown

Two weeks to a costed exit.

A fixed-fee review of your cloud bill and architecture, run by your team with our skills and reviewed by us. You leave with a savings ledger, an exit route for every major workload and the skills to run it again, whether or not you hire us for the rest.

Book a Teardown See a sample ledger
Duration
10 days
Access, held by your team
Read-only
Fee
Quoted upfront
Your data
Stays with you
What we need from you
  • Billing exports. AWS Cost and Usage Report, Azure Cost Management export or Google Cloud billing export, covering at least 90 days. Your team exports them, and they stay in your estate.
  • Read-only roles for your team. An audit role or Reader assignment your engineers hold. The skills never need write permissions.
  • Your commitments. Enterprise agreements, savings plans, reservations and their end dates.
  • Your people's time. Your platform lead and an engineer for the working sessions, plus one session with finance or procurement. We agree the hours at kickoff.
  • An AI assistant. One that supports Agent Skills and can run Python. The skills run there, on your data.
What you get
  1. 01Savings ledger. Every major workload: what it costs today, what it would cost on each route, and how confident we are.
  2. 02Exit routes. Stay and optimise, move to another cloud, move to bare metal, or a hybrid, each with cost, effort and risk.
  3. 03Renewal brief. Your commitments, your switching rights under the Data Act, and the numbers to take into the negotiation.
  4. 04DORA note. For financial entities: where exit plans are missing or untested against Article 28(8).
  5. 05The Teardown skills. Yours to run again when the bill or the estate changes.

We sign the review. It states what we checked, against which inputs and skill version, and what we didn't.

Sample savings ledger

Workload by workload, in euro a month.

Illustrative figures for a mid-sized SaaS platform on AWS. Not client data.

Workload Today Route After Saving
Production Kubernetes (EKS) €48,200 Bare metal, k3s €14,900 −€33,300
PostgreSQL (RDS Multi-AZ) €21,600 Bare metal, Patroni €6,800 −€14,800
Data egress €12,700 Mostly removed by the move €1,200 −€11,500
Analytics (Redshift) €18,900 Stay; resize the commitment €13,400 −€5,500
Object storage (S3, 400 TB) €9,400 Stay; tiering and lifecycle rules €5,100 −€4,300
Total per month €110,800 €41,400 −€69,400

Bare-metal figures include hardware or rental, colocation, bandwidth and operations. Staying put is a valid route and shows up as one.

The ten working days
  1. Days 1–2

    Kickoff

    The skills installed in your assistant, read-only roles and billing exports in place, the platform session.

  2. Days 3–5

    Map the bill

    Your team ties every euro to a workload, an owner and a dependency with the skills, while we coach.

  3. Days 6–8

    Price the routes

    Real quotes for bare metal and colocation, which we obtain (the first condition of the people test), plus cross-cloud pricing, effort and risk.

  4. Days 9–10

    Readout

    Ledger, routes and renewal brief, presented to engineering and finance together, with our signed review.

Questions buyers ask

How is this different from the free estimate?
The Price your exit skill reads your bill and assumes typical utilisation. A Teardown reads your estate: it measures utilisation, maps dependencies, adds your commitments and renewal dates, and prices the routes with real quotes.
Will you tell us to leave AWS?
Only where the numbers say so. Plenty of workloads should stay, and a credible plan to leave the rest is still worth money at renewal.
Do you need access at all?
No. Your team runs the skills on read-only roles and exports you control, and we review what you share. If you would rather we ran it, ask us in writing: we then work with read-only access only, and delete the exports and inventory data when the engagement ends.
We're a bank. Does this help with DORA?
Yes. The Teardown flags which critical or important functions lack a tested exit plan, and an Exit Plan engagement produces one.
What happens to our data afterwards?
It stays in your estate. Outputs you share for review are deleted at the end of the engagement unless you ask us to keep them. You keep the ledger, the models and the skills.
We don't have time to run it ourselves.
That is the fourth condition of our people test: capacity, not capability. We can add bounded hands for a fixed term, and the skills still go to your team.
Can we trust AI-generated findings?
Skill output is indicative, and says so. Our signed review states what we checked, against which inputs and skill version, and what we didn't.

Price your exit before your provider prices your renewal.

Email hello@counterweight.ie