White paper · A federal enterprise
From the Idea to the Outcome.
A federal enterprise with thousands of employees runs its technology work through ten portfolios of products. Ideas arrived faster than anyone could judge them. Plans were approved that nobody had truly validated. And the distance between the plan and the work became visible after the quarter was gone. This paper describes the operating system built to close that distance: validation that takes an idea to ready, and a live view that keeps the plan honest while reality moves.
The customer is described, not named, by agreement. The story is told in full.
Executive summary
Two disciplines do the work. The first is validation: every idea passes through a framework matched to its kind, and produces the artifact set the enterprise’s own governance requires, so an idea becomes a funded demand only when its problem has been proven, not just proposed. The second is plan versus reality: with ten portfolios in one view, leadership watches resources, commitments, and outcomes together, sees drift while it is still cheap to correct, prices the opportunity cost of every change before approving it, and keeps an audit trail that runs from the plan to the delivered impact.
White paper PDF
Get the PDF
The full paper, formatted for sharing.
- 10
- Portfolios, each holding multiple products, in one operating view
- 4
- Plan states between a submission and a commitment worth measuring against
- 140%
- One specialist’s combined allocation across three roadmaps, invisible until the plans shared a view
- 1
- Unbroken line from the original plan to the delivered impact
An enterprise too large to hold in anyone’s head
Ten portfolios hold the enterprise’s products, each portfolio spanning multiple products and the teams that build and run them. The work lives where large technology organizations keep it: ServiceNow for demand, Jira and Azure DevOps for delivery, GitLab, spreadsheets, and Power BI at the edges. Every one of those systems held part of the truth. None of them could draw a trusted line from an idea to the demand it became, the project, the product, and the outcome it was supposed to produce.
The front door had a defect nobody had named. Demand entered through the service platform, whose intake process assigned a solution owner almost immediately. That single step decided the shape of everything downstream: the system started solving before anyone had established that the problem was real, or worth solving at all. Early ideas faced a heavy form and often never entered cleanly. The ones that did enter carried their submitter’s assumptions into solution documents, and every artifact afterward inherited them.
Planning had the mirror-image problem. Teams submitted annual and quarterly plans, and leadership treated a submission as a commitment. Reporting on those plans was assembled by hand, stale by the time anyone reviewed it. The enterprise did not lack process. It lacked proof at both ends of the work: proof, before money moved, that an idea deserved resources, and proof, while the work ran, that the approved plan was the work actually happening.
Section one
Validation: idea to demand
The answer was an AI operating system configured for this enterprise specifically: its portfolio hierarchy, its governance rules, its own artifact standards, its context. That context works at three levels at once: the idea itself, with its files, notes, and the gaps the system prompts its owner to fill; the portfolio it belongs to, with its strategy, resources, and standards; and the artifact being produced, because a charter, a decision record, and an executive summary are written for different readers. The frameworks and artifact agents are customizable per kind of work; the configuration belongs to the enterprise, not to a vendor’s default.
The workflow inserts a problem-validation stage before any solution work begins. What problem are we solving, and who experiences it? Where else in the enterprise does this same problem appear? Is it a one-time solve or a durable product need? And before anything is built: does the enterprise already own a capability that addresses it? Validation maps the problem against the enterprise’s business capabilities and its application registry, and checks it against hundreds of demands already in the system. In a building this size, “do we already have this?” is the cheapest question there is, and the old process never asked it systematically.
Strategy enters here, deliberately and lightly. Each idea is validated against the stated direction of the portfolio it would draw resources from, so alignment is a gate at intake, not a hope reviewers bring to a meeting months later. An idea that cannot connect itself to that direction is not argued about. It gets answered early, while saying no is cheap.
The road from an idea to a demand
The idea enters
Through the front door the enterprise already runs, at the weight an early idea can carry
The problem is validated
Proven, framed, and checked against what the enterprise already owns
A framework is matched
Scrutiny proportioned to the kind of work: new product, enhancement, build versus buy
The artifact set is drafted
Charter, problem, solution, scope, stakeholders, generated from context and owned by a person
Demand, ready
Funded, staffed, and baselined as a plan reality can be measured against
Not every idea deserves the same scrutiny, so the framework is matched to the kind of work in front of it:
The deep path
New products
Full problem validation, capability mapping, evidence drawn from the enterprise’s own user research, and the complete artifact set before a demand is approved. New products make the largest claims on the future, so they carry the heaviest burden of proof.
The lighter path
Enhancements and optimization
The product and its users are known, so validation confirms the problem, the expected gain, and the fit with the product’s direction, without the ceremony a new product carries.
The decision is the deliverable
Build versus buy
Validation lays out what already exists, what building would require, and what buying would commit the enterprise to. The reasoning is preserved in a decision record, so the choice can be revisited on evidence instead of memory.
The standing path
Sustainment and minor work
The categories of work that keep existing products alive. Small enhancements run as a governed allowance against a product, outside the project pipeline, planned honestly as the cost of standing still well, not forced to masquerade as projects.
What a validated idea produces is the enterprise’s own artifact set, defined with its program management office: the project charter as the anchor, the business problem it answers, the proposed solution, scope, and the stakeholder map, with decision records and build-versus-buy analyses added where the path requires them. The AI drafts each artifact from the validated context rather than a blank page; a named person reviews, corrects, and owns it, because in this enterprise the individual remains responsible for AI-generated output. Governance stops being paperwork chased after decisions and becomes a byproduct of making them.
The planning cycle proved the discipline at scale. When the annual plans arrived, they ran through the same validation as everything else, and the pass caught what submission alone had always hidden. One team’s plan listed projects that did not line up with each other; its resource totals did not roll up to the people the team actually had; work sat in it carrying no line back to the portfolio’s strategy. Other plans arrived perfectly flat, the same staffing in every month of the year, with ideas whose validation scores were zero. Under the old rhythm, those plans would have been treated as commitments, and their defects discovered quarter by quarter as missed delivery. Instead, a plan now moves through states: submitted, reviewed, approved, committed. Only a committed plan becomes the baseline reality gets measured against.
Section two
Plan vs reality
A committed plan is only useful if someone can see it whole. The enterprise’s hierarchy now lives in one view: portfolio, product line, product, application. Across ten portfolios, leadership can see what is funded, who is committed, and what each commitment should return, without ten groups assembling ten versions of the truth. The reconciliation underneath matters as much as the view: every person is resolved across plans into a single record, so commitments add up across the whole enterprise instead of inside each portfolio’s silo.
That reconciliation surfaces what silo planning structurally cannot. One specialist appeared at fifty percent on one product’s roadmap, thirty percent on a second, and sixty percent on a third. Read alone, each plan looked reasonable. Read together, the enterprise had promised one hundred forty percent of one person. Nobody had made an arithmetic error; nobody had ever seen the three plans in the same view. Multiply that person by every scarce expert across ten portfolios, and the capacity problem changes character: not a staffing shortage, but a visibility defect, and a correctable one. Over-allocation now appears as an exception to resolve in planning, not a surprise at delivery.
Once a plan is committed, execution flows in from the delivery tools the teams already use, and the plan is measured as the work happens, not reconstructed after the fact. What leadership watches is the exception list: staffing gaps, missing dates, commitments drifting from the plan, decisions waiting too long for an owner. Drift is treated as information, delivered early and specifically, while correcting it is still cheap. A plan that slips in silence for a quarter is a management failure the system no longer permits.
New ideas do not stop arriving because a plan exists. When one lands mid-year, it is evaluated against the committed plan before anyone argues about it: what it would need, whose time it would take, and what would have to move to make room. The levers are named in advance. Work can shift to a future year. Partially completed work can be moved, at a visible cost. The run-and-maintain load can be trimmed, at a visible risk. The price of saying yes sits in the room as a number, not a later consequence nobody chose.
A committed plan changes only through a change order: a recorded decision naming what moved, why it moved, and what it displaced. The money stays attached to the same record, so a commitment can be followed through its whole life: the initial estimate, the approved budget, the committed resource plan, the forecast, the actual spend, and the variance against all of it. Put together, leadership can walk one unbroken line from the original plan to today’s reality and account for every step between. What was promised. What changed. Who decided. What it produced. That line is the audit trail, from plan to reality, and from execution to impact.
What this makes possible
The enterprise can now answer two questions it could not answer before. Should this exist? Validation answers that one before the money moves, with the problem proven and the artifact set to show for it. And is the plan real? The plan-versus-reality view answers that one continuously, while the answer can change the outcome. Between those two answers, the distance between strategy and execution stops being a place where value quietly leaks out.
The pattern matters beyond one enterprise. This deployment began as a single Snapshot inside one division, proved itself against operational data, and moved into production. It is now expanding team by team. Large institutions do not need another system of record competing with the ones they already run. They need the layer above those systems where judgment happens before the commitment, and truth is kept after it.
White paper PDF
Get the PDF
The full paper, formatted for sharing.
Free Strategy Snapshot
Connect the dots between your strategy and the work your team actually does.
In one free 30-minute working session, we take in your strategy and the work data you already have, and build a fully functioning connected view of execution against strategy, plus the repeatable model your team can keep running. You see where execution matches the plan, where it has drifted, and what to do next.
What you get:
- Your strategy, priorities, and active work connected in one live view
- Where your product and engineering capacity is actually going, against the priorities the business set
- A repeatable model your team can keep running, not a one-off report
Free. 30 minutes. No account needed. No credit card required. No pitch deck; if the fit is not there, we will tell you.