Pradeep Kanagaraj is Vice President of AI Operations at Ensemble Health Partners.gettyThroughout my career leading operations and strategy, I've watched an old assumption collapse: that strategy can be set once a year and executed on schedule. The annual plan is already dead; the quarterly plan is dying slowly. Most organizations still fund work the way they did years ago, locking in budgets months in advance and treating reallocation as a planning failure rather than a feature.Consider how fast the ground moves. AI compute costs have collapsed by orders of magnitude, repricing business cases overnight. Major model releases land weeks apart, so bets get leapfrogged before they ship. 2023's hottest job titles are obsolete. Compliance exposure is now a board-level line item.When inputs reprice this quickly, the problem isn't that your plan was wrong. Any fixed plan is wrong by the time it's funded. The question for every operating leader: How do you reprice strategy continuously without chaos?​Stop Planning. Start Trading​The reframe starts with the question leadership asks. Old: "What's our road map?" New: "Where does the next dollar of capacity earn the most return?" Every executive manages a portfolio of bets. The only choice: Defend it as constructed, or trade it as conditions change.Trading cannot be improvised. Repricing without chaos requires a system, not a habit: three layers in a loop, re-scored every cycle. Fund, execute and measure.​Fund: Price The Backlog With Value, Confidence And Cost Of Delay​Most prioritization debates are won by the loudest voice. To remove that distortion, I price every initiative with three multipliers:• Value. The size of the prize: revenue won, cost removed or risk retired.• Confidence. How sure are we? This deliberately discounts hype.• Cost of delay. How fast the prize decays. What a week of waiting erodes, in dollars per week.Multiply the three, and the backlog has a price you re-score instead of arguing about. In one illustrative portfolio, the “moon shot" everyone was excited about scored 5.4 (9 x 0.3 x 2): high value but low confidence, with almost no decay from waiting. A boring operational fix scored 32.0 (5 x 0.8 x 8) because it was bleeding $8,000 a week in manual labor and rework. The exciting project funds last; the bleeding one funds first. Cost of delay ends the argument.Cadence matters too. I re-score quarterly, long enough to ship and short enough to correct cheaply, while watching signals weekly for drift. Most of the portfolio holds steady through each cycle, so reallocation becomes a scheduled ritual rather than a fire drill. You move the marginal dollar, not the whole board.​Execute: Convert Outputs Into Owned Inputs​Funding the right work is half the system; knowing execution is on track is the other. Traditional OKRs focus on the what but neglect the how. An output metric like $10 million in ARR lands only at quarter-end, and because no one owns it alone, "70% done" is a guess.The fix is what I call OKIRs: objectives and key results with input metrics. Decompose the output into the inputs that produce it: 10 deals at $1 million each, requiring 50 late-stage opportunities, requiring 500 prospects contacted. Each input gets one named owner and a real count, so 25 of 50 opportunities is exactly 50% done. A number, not a vibe. Defining inputs up front pushed our key result success rate above 70%.Control matters too. Gates ask permission; guardrails set boundaries. Work-in-progress caps, spend ceilings, risk thresholds and experiment budgets let teams move freely inside agreed limits. Speed rarely dies in the work. It dies in the approval queue.​Measure: What Gets Measured Gets Delivered​Metrics aren't reporting. They are a leader's strongest lever over organizational behavior. People optimize toward whatever the scorecard rewards; the metric, not the memo, sets the real priority.That conviction led to the Qualume score: quality times throughput, benchmarked to a common baseline for fairness and tied directly to bonuses. It rejects the false choice between speed and quality by scoring both simultaneously. Collective throughput rose from under 30% of target to over 100%, while quality held steady. When a critical tool went down, scores dropped, and overnight, uptime became everyone's problem. No executive escalation required.Measurement also reshapes capacity. Headcount is a snapshot of last year's priorities, rigid and tied to org boxes; skills are a liquid asset, rotated to the highest-value work. If the work can move but the people can't, you have silos with a spreadsheet on top.​AI Is The Processor That Runs The Operating System​AI fits not as another initiative, but as the processor that runs this operating system. The OS sets the rules; AI executes them continuously, faster than any committee. It re-scores the portfolio in minutes, chases dependencies and surfaces blockers like a program manager, and tracks quality and throughput live so confidence is real-time, not a quarter-end guess. Stop trading by hand. Automate the desk, with humans in the loop, not the weeds.​Action Plan​Most leaders already have the evidence in their portfolios. The zombie initiative "90% done" for two quarters is a funding problem that value, confidence and cost of delay would expose. The blind bet everyone calls 70% complete but no one can prove is an execution problem that input metrics would make real. The whiplash team, swinging between fast-and-sloppy and slow-and-perfect, is a measurement problem that a balanced score would resolve.Three steps for this week:1. Fund. Score your top 10 initiatives on value, confidence and cost of delay, and see what reorders.2. Execute. Pick one goal, define its input metrics and assign owners. Turn one OKR into an OKIR.3. Measure. Rate one team on quality and throughput together, not output alone.In the project economy, strategy isn't a document you write once a year; it's a portfolio you reprice every cycle. The winners won't have the best plan. They'll have the discipline to fund, execute and measure in a continuous loop and the courage to trade when the signals say it's time.​Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?