This page walks a Mactores engagement end to end. What happens inside each of the five phases, who does which half of the work, what you put your signature on at every gate, and what the engagement leaves behind for your auditor.
The short version: automation carries the repetitive bulk of the engagement, forward-deployed engineers carry every judgment call, and the ship date is written into the statement of work before build begins.
Thirty minutes with a forward-deployed engineer who runs this method on live engagements. Ask how a phase closes, what the automation covers, and where it has gone wrong before.
If a form is not how you work, write to lets-talk@mactores.com and ask for the method walkthrough. It reaches the same people.
Ask the awkward one first
How the fixed date survives a discovery surprise in week three.
Your question goes to a delivery lead, not a marketing list. We reply inside one business day.
Five figures, each of them checkable, and the reason the rest of this page can describe a date as a commitment rather than an ambition.
60–70%
of the engagement hours a traditional proposal of equivalent scope staffs for, absorbed by automation
12 wks
median time to production across Mactores engagements
21
public case studies with named customers and named outcomes
200+
AWS-certified engineers, with 7 competencies and 17 service validations behind them
2008
building on AWS since, which is eighteen years of production migrations
Start with the deadline
Bring the date you are working against and we will tell you what reaching it takes.
A renewal, an end-of-support cliff, an audit finding, or a programme that already stalled once.
Traditional consulting is priced as a pyramid. A large junior pool runs the analysis under a thin senior layer, and the commercial model depends on those hours being billed, which is why time and materials remains the industry default. Agentic delivery breaks that arithmetic in one specific place. The analysis hours collapse. Work that used to occupy a team of analysts for weeks now runs in an afternoon, and what remains is architecture, judgment and the customer relationship, which is precisely the work a pyramid was never shaped to do.
So we reshaped delivery around the work that is left. Automation takes the repetitive majority. Senior engineers take everything that requires a decision. The commitment then moves out of the pitch and into the contract, because a model with fewer billable hours in it can afford to carry a date.
The traditional shape
The shape we run
The 60 to 70 percent figure is measured against a traditional consulting proposal of equivalent scope, which is the only comparison that means anything. Those hours leave the estimate because the work is automated, and the scope in the statement of work stays the same.
The clause behind the number
The date carries a consequence, and the consequence is written down.
Read the overage terms before any conversation about scope. They govern every engagement we sign.
Both halves are named in the statement of work, and the split is identical on every engagement, so you always know who owns the thing in front of you.
Automated
Discovery, mapping, schema analysis, harness generation, documentation
Engineer judgment
Architecture, sequencing, risk calls, cutover, sign-off
Your review
Domain validation and acceptance
Scope and design
Phases 01 and 02
Build and validate
Phases 03 and 04
Cut over and retire
Phase 05
Automation
Repetitive at scale
Scope and design
Discovery across source systems, dependency graphs, schema and code complexity analysis
Build and validate
Test and validation harness generation, reconciliation runs, documentation and runbook drafts
Cut over and retire
Observability wiring, decommission inventory, evidence pack assembly
Forward-deployed engineers
Senior, embedded
Scope and design
Target architecture, trade-off calls, the ship date and the fee they are prepared to sign
Build and validate
Risk-weighted sequencing, exception handling, the calls the automation cannot make
Cut over and retire
Cutover command, rollback decision, production sign-off and handover
Your team
Reviewers and owners
Scope and design
Access, context on the systems only you know, and acceptance of the scope
Build and validate
Slice-by-slice review in your own environment
Cut over and retire
Go or no-go on the window, and the signature that closes the phase
Your team appears in this table because your effort is part of the plan. We would rather schedule it in week one than discover it in week nine.
Meet the person, not the org chart
The engineer who would own your engagement will take the first call.
Same person through scoping, build, cutover and the signature at the end.
Five phases. Each one has an entry condition, a named output, and a gate that only your signature opens. Every phase exit produces a document, and those documents are what the audit section of this page is about.
Automated discovery runs across your source systems while the forward-deployed engineer builds the target architecture on top of what it finds. You get a scope, a ship date and a fixed fee, all three written into the statement of work. This phase exists to produce a number your finance team can plan against instead of a range they have to hedge.
Gate
You sign the scope, the date and the fee before any build work begins.
The target design is documented and then the riskiest assumption inside it is built and tested ahead of everything else. If a plan is going to break, it should break in week three on a proof rather than in month five on a deadline. Whatever we learn here is reflected in the plan before the build is funded.
Gate
You accept the design and the result of the proof.
Work ships in slices that run end to end, each one demonstrable in your own environment. You review something working rather than a status percentage, which means progress is observable and a slipping slice is visible the week it slips instead of the month it lands.
Gate
You review each slice in your environment as it completes.
A generated validation harness runs the new system beside the old one on real data and reconciles the difference object by object and field by field. Where the two disagree, the engineer explains why before anyone signs. This is the phase that turns a migration from an act of faith into a reconciliation you can read.
Gate
You accept the reconciliation, including every explained variance.
Cutover follows a rehearsed runbook with rollback criteria agreed in advance, so the decision on the night is a check against criteria rather than a judgment made under pressure. Then the old environment is switched off. Retirement is written in as a deliverable, because a modernization that leaves the legacy system running has moved cost rather than removed it.
Gate
You sign production acceptance, and the decommission record closes the engagement.
The commercial half of this method
The date, the fee and the overage terms are contract language, not positioning.
Our commitment page sets out exactly what goes into the statement of work at each of these five gates.
Twenty-one case studies are published with named customers. The three below each stress a different part of this method: the phase gate that asks for evidence, the rehearsed cutover, and the split between what automation runs and what a person decides. Every figure is the customer's own, published on their case study.
AI agents · Healthcare
KlearTrust
12 wks
to production, after eighteen months across two earlier attempts
First
internal audit review cleared, with no remediation round
A claims-review agent on Amazon Bedrock, HIPAA-aligned and eval-harnessed. The earlier pilots reached a demo and stopped where the evidence was asked for, which is the phase gate this method makes unavoidable.
Read the case studyData platform · Marketplace
Poshmark
Zero
downtime through cutover on a live consumer marketplace
70%
reduction in operational overhead once the legacy platform was retired
The runbook and the rollback criteria were agreed before the window opened, so the night itself ran as a sequence of checks. Retiring the old platform was a deliverable inside the statement of work rather than a follow-on project.
Read the case studyAI agents · Legal services
Sterne Kessler
10 min
to classify a filing, down from around two hours
85%+
accuracy across all forty-five Nice Classes
The volume work runs on Amazon Bedrock while the attorneys keep the judgment calls, which is the same division of labour this page describes, applied to a legal workflow rather than a migration.
Read the case studyFull detail on each is published on the case studies index. Named accounts and audited figures are shared under NDA during commercial discussions.
Ask for the closest reference
We will name the engagement nearest your stack, not the category.
If nothing in our history is close enough, that is a useful answer too, and you will get it on the first call.
These are the tools our engineers build agents with most weeks, which is why estimating work on them is a matter of record rather than research. The 17 services below each carry an AWS Service Delivery validation — AWS grants it only after watching that service run in production for a named customer — so discovery starts from what we already know rather than from a blank page.
Validation is retrospective. AWS grants it for work already in production, so every badge above is a service one of our engineers has run under real load, at a named customer, against a real date. What came out of those engagements — the mapping rules, the failure modes, the reconciliation queries — is what our automation replays during discovery. It is why week one hands you a map instead of a questionnaire.
The services we assemble agents from every week. AWS issues no Service Delivery validation for these, so they are named in type rather than shown as badges.
Build stack
The source estate sets the sequence, so the list below matters as much as the target. Each of these is a system our engineers have moved to production more than once, which is why the migration plan arrives with an order of operations already in it.
Your estate, not a reference architecture
Send us the estate you actually have, including the parts nobody documented.
Discovery starts from the real dependency graph, which is usually where the surprises live.
The competencies behind the delivery
A competency is AWS reading our customer references and agreeing we have done this at production scale. On this page it answers a narrower question. When your design review lands in week two, which of our people has already built your kind of workload — and what did it look like when they were finished?
Legacy estate on the clock
The engagements that start with something which has to be switched off — a licence renewal, an end-of-support date, a warehouse nobody runs new workloads on. We have taken this route often enough that the cutover rehearsal is a checklist rather than an invention. It is also the credential AWS requires before an engagement can draw on MAP funding.
Oracle · SQL Server · Hadoop · legacy warehouses
The foundation everything else reads
Most programmes that arrive here are a data problem wearing another label. This covers collection, storage, governance and query at the sizes where a naive design quietly stops working — and it is why a platform we hand over still performs a year after the engagement closes.
100+ platforms delivered and handed over
Agents against real operations
AWS's assessment track for AI delivery, and the one that shapes how we staff. It is judged on systems running against live data with people depending on the output — a different bar from a demo that answers questions about a document.
Bedrock under load, not in a sandbox
Models that have to keep running
Covers the unglamorous half: feature pipelines, retraining, drift, and the handover to whoever operates the thing on Monday morning. We hold it because the models we ship are expected to survive a production release cycle without us.
70+ customers, models still in service
How the work reaches production
Less a service line than a property of every engagement. Infrastructure as code, pipelines, monitoring and logging ship alongside whatever we built, because a system nobody on your team can deploy or observe is not a finished system.
Ships by default, never as a change request
When the data is someone's record
An industry credential, and AWS assesses it in a different order — what you do about privacy, access and audit trail comes before what you built. It matters when a reviewer will want to know who could see which record, and when.
HIPAA claims platform, audit cleared first pass
Plant floor to query
The second industry credential. Operational and fab data arrives fast, dirty, and in formats nobody chose, so the engineering problem is usually ingestion rather than analysis. Earned against work of exactly that shape.
Fab data lake, 40% more throughput
No trust required
All 7 sit on our AWS Partner Solutions Finder entry under Mactores Cognition Inc. If a credential is not on that entry, it is not claimed on this page.
Open the AWS listing
Fifteen certification types across the team, which is the practical reason the engineer assigned to your phase two is already qualified on your services rather than reading ahead of you. We staff from the people who hold the ticket.
Used at handover rather than at kickoff. Where your team will operate what we built, we run AWS's workshop format against your actual system, so knowledge transfer is a working session and not a slide deck.
Meet the engineer, not the logo
Ask which of our people would own your design review, and what they built last.
We will name the engagement nearest your stack and the engineer who signed it.
Regulated modernization has to produce a defensible record as well as a working system. In this method the record is a byproduct of the work rather than a workstream you fund afterwards, because the validation runs and the phase signatures already generate what a reviewer asks to see. Designing against a framework is a different thing from certifying you to it, and the certification stays yours and your auditor's.
What each framework wants, and where the method already produces it
Wants evidence of
Cardholder data boundaries, change control, and evidence that scope did not quietly expand
Produced by
Dependency graph from phase 01, plus the signed scope that fixes the boundary
Wants evidence of
Where protected health information lives, who reached it, and what happened to it on decommission
Produced by
Field-level lineage from phase 04 and the decommission record from phase 05
Wants evidence of
Traceability from source of record to the output a person or an agent acted on
Produced by
Reconciliation output from phase 04, retained with its variance explanations
Wants evidence of
That the numbers in a filing can be traced back to the system that produced them
Produced by
The same phase 04 reconciliation, retained with its lineage
Wants evidence of
That changes were authorised by someone with the authority to authorise them
Produced by
The customer-signed acceptance record that closes every phase gate
Wants evidence of
Lawful handling of personal data through the migration, and proof of disposal afterwards
Produced by
Lineage from phase 04 and the disposal detail inside the phase 05 decommission record
Marks appear as published by each body and are not recoloured. PCI DSS and SOC 2 are named in type throughout: PCI SSC issues no compliance logo and prohibits marks implying compliance, and the AICPA licenses its mark only to organisations holding their own unqualified report. Naming a standard indicates what the delivery work is designed against, and not endorsement or certification by that body.
Bring your examiner's last finding
We will show you which phase exit would have produced the evidence for it.
The full audit evidence pack is available under NDA during commercial discussions.
AWS reviewed our delivery record, our certification depth and our customer references before granting Premier Tier, which is the highest tier in the partner network. Everything below is listed publicly, so procurement can confirm it before the first call rather than after it.
Solutions Architect – Professional
DevOps Engineer – Professional
Generative AI Developer – Professional
Solutions Architect – Associate
Developer – Associate
SysOps Administrator – Associate
Data Engineer – Associate
Machine Learning Engineer – Associate
Advanced Networking – Specialty
Data Analytics – Specialty
Machine Learning – Specialty
Security – Specialty
AI Practitioner – Foundational
Cloud Practitioner
AI Business Strategist – BusinessThe Agentic AI Specialization is named in words because we hold the credential and not the badge artwork. We would rather name a credential you can verify than display a mark we cannot source. Full detail is on the partners page.
Verified by AWS, not by us
Every credential here is granted by AWS, so you can check it without us.
The tier, the specialization and each competency are listed on our partnership page and in the Partner Solutions Finder.
A senior engineer who embeds with your team for the engagement, owns the architecture and the cutover, and signs for the delivery date personally. The person who scopes the work is the person who lands it, which removes the handoff between the team that sells and the team that delivers.
It carries the repetitive work at scale: discovery across source systems, dependency graphs, schema and code complexity analysis, validation harness generation, reconciliation runs, documentation and runbook drafts, and assembly of the evidence pack. Choosing the target architecture, sequencing the work against risk, handling exceptions, calling a rollback and signing for production all stay with a named engineer.
Discovery finishes inside phase 01, before the scope, the date and the fee are signed, so most surprises land while all three are still open. Phase 02 then builds and tests the riskiest assumption in the design ahead of everything else, which is where the remaining ones tend to show up. Anything found after signature is handled under the change and overage terms on our commitment page, and delays we cause are ours to absorb.
The artifacts a reviewer asks for are the artifacts the delivery work already produces: the phase 01 dependency graph and signed scope, the phase 04 reconciliation with its variance explanations and field-level lineage, the customer-signed acceptance record at every gate, and the phase 05 decommission record. They are retained as the work runs, so the evidence pack is assembled rather than reconstructed. Designing against a framework is a different thing from certifying you to it, and the certification stays yours and your auditor's.
The scope, the ship date and the fee, all three written into the statement of work at the close of phase 01, together with the five phase gates and the overage terms that say who carries the cost if the date moves. Legacy retirement is written in as a deliverable. Our commitment page sets out the contract language in full.
Ask us directly
Holding a question this page did not answer? That is the one worth a call.
Thirty minutes with the forward-deployed engineer who would run the engagement.
Signed, then shipped.
Thirty minutes with the forward-deployed engineer who would run the engagement, and you leave with a real view on the date and the shape of the work.
