By Shannon H. Boykin · 6 min read
A stalled portfolio rarely announces itself. There's no outage, no audit finding, no headline. Work is happening — meetings, briefings, sprints — and yet the mission capability at the end of it isn't improving. Here are five signs I've learned to watch for, and what actually gets a portfolio moving again.
New for 2026: under each fix, a note on where AI genuinely helps — and where it just produces confident noise faster. AI accelerates the work; it doesn't replace the judgment, or the need to verify that “done” is actually done.
1. Your roadmap hasn't really changed in a year
Modernization is perpetually “in progress,” but look closely and the same systems are on the same slides. Nothing has been sunset. Cloud migrations are counted, but the legacy estate underneath them never shrinks. That's not modernization — it's accumulation. Real progress shows up as things retired, not just things added.
What unsticks it: aggressive, honest rationalization. Pick the legacy systems you will actually turn off, put dates on them, and defend those dates. Consolidation is where the capital and the capacity come from.
⚡ Can AI help?Yes — on the analysis. A working-AI agent can compress a months-long rationalization study into weeks: inventory the estate, map dependencies, and flag the redundant or barely-used systems so you can see what's safe to retire. It surfaces the candidates and the evidence; the decision to pull the plug stays human. Verify the list against reality before you act — a confident inventory is not a correct one.
2. Every decision dies in the “frozen middle”
Leadership sets the intent. The mission owners know what they need. And somewhere between them — in the program offices, the resourcing panels, the cross-command seams — initiatives quietly lose momentum. Everyone is doing their job, and nothing moves. The frozen middle isn't a people problem; it's a misalignment problem.
What unsticks it: translation and partnership. Someone has to speak both “general” and “program office” — turning executive intent into acquisition reality and back again — and build the coalition so the people who have to execute actually own the plan.
⚡ Can AI help?This is where AI earns its keep on information flow: translating executive intent into program-office-ready language and back, drafting the briefs in both directions, and keeping everyone working from one synthesized picture — shrinking the latency where momentum usually dies. It removes the friction; it doesn't build the coalition. That's still human work.
3. You're defending budget on efficiency, not mission
When your program justifications lead with “reduces technical debt” and “improves efficiency,” you're speaking a language the resourcing process has learned to discount. In an era focused on mission and great-power competition, IT investments that can't articulate operational impact get marked — and you spend the year defending what you should have been growing.
What unsticks it: reframe every line item around mission impact and risk. I've watched a portfolio recover a mid-eight-figure cut simply by re-articulating what the systems do for the mission, not what they cost.
⚡ Can AI help?AI is genuinely strong at reframing. Feed it the technical justification and it drafts the mission-impact narrative, models the operational risk of a cut, and produces richer POM and budget-defense material far faster than a team can from a blank page. You bring the mission judgment; it accelerates the articulation.
4. Your data is a swamp, not an asset
You have dashboards. You have an analytics platform. And yet, in the meetings that matter, leaders still decide on gut — because they don't fully trust the numbers. When data isn't visible, governed, and trusted, it isn't an asset; it's overhead. Decision-making stays slow no matter how much you've spent on tooling.
What unsticks it: radical transparency and a single source of truth. Get the functional owners into the data, make performance inescapable, and adoption follows. A platform only creates advantage when the community actually runs on it.
⚡ Can AI help?Once the data is governed, AI makes it usable — plain-language questions over trusted sources, auto-generated summaries, anomalies surfaced for leaders without an analyst in the loop. The catch is the whole point: AI on an ungoverned swamp just produces confident nonsense faster. Fix the single source of truth first — then AI multiplies it.
5. Audit and compliance are a perpetual fire drill
Every audit cycle feels like the first one. Findings recur. Remediation is a scramble, handled by heroics rather than by design. If FIAR readiness lives in a spreadsheet that gets rebuilt every year, your portfolio is spending energy standing still — energy that should be going into capability.
What unsticks it: systemic remediation, not seasonal. Restructure the rhythm, embed automated controls and “minimum viable governance,” and close findings at the root so they stop coming back.
⚡ Can AI help?AI can turn seasonal heroics into continuous readiness: monitoring controls, keeping FIAR evidence current instead of rebuilt yearly, flagging drift, and drafting remediation the moment a gap opens. Building AI can embed the automated controls in the systems themselves. Demand the proof, though — a control that reports itself compliant still has to show it actually ran.
The common thread
Notice what these five have in common: none of them is really a technology problem. Portfolios stall on alignment, ownership, and the translation of IT value into mission terms — and they get moving again through transparency, partnership, and the discipline to retire the old to fund the new. That's the work. It's rarely glamorous, and it's almost never done alone.
If two or three of these felt uncomfortably familiar, the good news is that stalls are fixable — usually faster than people expect, once the right constraints are named and the right people are aligned around them.
