When your AI programs give your team 14,000 hours back next year, who decides how the team spends them?
In my upcoming book, Coherence at Velocity, I argue that change capacity is a strategic asset. Every launch draws on managerial attention, training bandwidth, exception handling, redesign effort, and judgment at the edge, and most enterprises launch more change than their people can turn into stable operating behavior. The limit has a name, absorptive capacity: an organization's ability to recognize the value of new knowledge, assimilate it, and apply it to produce outcomes. Leaders running AI programs now manage a second constraint alongside it. When a team automates a task, it gets hours back, and those hours become capacity only after a leader assigns them to specific work.
On Friday I argued that every AI program needs a one-page report for the CFO showing its outcome, total cost, Return on Agent, and owner. This edition covers the decision that comes before the report: where the freed time goes, and who has the authority to send it there. Analysts and authors at Gartner, HBR, and Forrester published on that decision between October 23 and 26.
Decide where freed capacity goes before you count it as savings
CV Viverito and Arj Bagga of Gartner report that 77 percent of board members expect AI to deliver productivity and cost savings, and that 3 percent of organizations exceed their cost-savings goals through productivity. Employees in a separate Gartner survey redirected 21 percent of the time AI freed up into more innovative work. Adam Job and four Boston Consulting Group colleagues write in HBR that roughly three-quarters of more than 600 US public companies have moved past AI pilots without earning a shareholder-return premium, and that the companies pulling ahead follow one of three paths to value. Martin Gill of Forrester writes that employees have quietly filled the gaps in most CX strategies with their own judgment, and that agents need those gaps written down as decision rules. Amii Barnard-Bahn writes in HBR that executives stall AI programs when they claim sole control of a decision or step back from decisions outside their expertise.
Read together, these authors describe one planning step most marketing organizations skip. For every AI program, write down which work receives the freed hours, which value path that work serves, the decision rules the work follows, and the executive who approved the assignment.
Assign every freed hour to named work before the program goes live
CV Viverito, senior director analyst, and Arj Bagga, VP analyst, presented new Gartner HR research at Gartner HR Symposium/Xpo in Orlando on October 26. Seventy-seven percent of board members expect AI to deliver productivity and cost savings, and 97 percent of organizations cannot reach their desired cost savings through productivity alone. Viverito reported that 38 percent of companies are over budget on AI and 85 percent plan to spend more next year. That combination will feel familiar to anyone who has remodeled a kitchen.
Marketing leaders should read further down the release. In a July 2026 Gartner survey of 297 CHROs, 78 percent said their leadership teams have no alignment on where AI creates value. In a second-quarter survey of 1,865 employees, Gartner found that people redirected 21 percent of the time AI freed up into more innovative work. Viverito put it this way: "What organizations are struggling with is how to make that capacity usable." Gartner analysts recommend treating capacity gained as an asset, starting with the skills people need to take on higher-value work: AI fluency, human connection skills such as stakeholder management, and adaptability and creativity.
In my book, Marketing Operations 3.0, I include capacity released and redeployed in the value side of Return on Agent, with one condition: time savings count only when a team redirects the released hours into higher-value work such as strategy, exception handling, creative development, or customer empathy. Hours that disappear into meetings, rework, and drift stay out of the calculation. Apply that rule to Gartner's numbers and roughly four of every five freed hours fall out of an honest ROA. Agile Brand Principle 4 asks leaders to respect employees and their time, and an hour nobody assigned usually ends up in a meeting about AI.
It reminds me of what Aytekin Tank, Founder and CEO of JotForm, said when I interviewed him on The Agile Brand podcast: "Wherever I look—marketing, growth, development—people are using AI and are much more productive; instead of wasting time on manual, boring tasks they give those to AI and spend more time on things that actually make a difference."
Tank describes the outcome every CMO wants, and Gartner analysts measured the share of freed time that reaches that kind of work at about one-fifth.
Take a hypothetical specialty retailer with a 120-person marketing organization. Its AI programs for content production and performance reporting cost $900,000 a year in licenses, consumption, and governance, and the vendors promise 14,000 hours back, about seven full-time roles. At an $85 loaded hourly rate, those hours look like $1.19 million on a slide.
In the first version, the CMO lets the hours land wherever they land. Employees redirect 21 percent of them, or 2,940 hours, into higher-value work, worth about $250,000. Return on Agent comes to ($250,000 minus $900,000) divided by $900,000, about negative 72 percent. The marketing operations lead presents strong usage numbers and a negative return.
In the second version, the marketing operations lead builds a capacity ledger before go-live, and the CMO, CFO, and CHRO approve it together. Seven thousand hours go to the CRM team for a lapsed-customer reactivation program, which brings back 4,000 customers at $210 of annual margin each, or $840,000. Three thousand hours go to writing decision rules and escalation thresholds for the service chat agent the company plans to scale in the spring, which replaces a $240,000 consulting engagement. Two thousand hours cover a departing analyst's workload, and the team leaves that role unfilled, saving $170,000. The final 2,000 hours fund continuous enablement for the new workflows, and the team counts no dollar value for them, because without enablement the other three numbers would not hold. Total value comes to $1.25 million, and ROA comes to about 39 percent.
Between the two versions, the CMO changed only where the hours went.
My take: Before your next AI program goes live, list the hours it should free up and assign every one to named work with an owner. Report only the assigned hours in your Return on Agent, and treat the unassigned remainder as a planning gap.
Choose which value path each AI program serves, then build the capability that path needs
Adam Job of the BCG Institute, Ulrich Pidun, Matthew Kropp, Tania Wang, and Matthew Molloy of Boston Consulting Group published 3 Paths to Delivering Shareholder Returns with AI in HBR on October 26. They studied more than 600 US public companies and found that roughly three-quarters have invested in AI and moved beyond pilots without earning a shareholder-return premium over industry peers. The companies pulling ahead follow one of three paths: doing existing work more efficiently, improving current products, or creating new offerings. The authors tie each path to a capability the company has to build, among redesigned workflows, broad employee AI fluency, and specialist talent, and they argue that adding tools and expanding deployment will keep a company on the plateau.
BCG Henderson Institute researchers published the analysis behind the article with more detail. Six percent of the companies in the sample qualify as adoption leaders, and their industry-adjusted total shareholder returns over three years ran 9 percentage points above the median. Revenue growth and margin expansion account for that outperformance, with no contribution from higher price-to-earnings multiples. Growth is the dominant path among the leaders, and between 2022 and 2025 they increased headcount at a compound annual rate 3 points higher than the median firm.
Most AI roadmaps I review group programs by vendor or by function, and leaders reading them cannot tell which programs serve growth. Label each program with its path and with the capability the authors tie to it, then check whether you have funded that capability. In the retailer example, the 7,000 hours sent to reactivation put the largest share of capacity on the growth path, where BCG found the leaders.
Section 14.5 of Marketing Operations 3.0 gives leaders five decisions to make from Return on Agent: scale, fix, reallocate, constrain, or kill. Reallocate applies when the capability has merit and the deployment target is wrong. Many efficiency programs that look weak on ROA belong in that category, because the freed hours went to the wrong work.
My take: Add a path column to your AI program list and fill it in for every line. Where a program sits on the efficiency path with no workflow redesign attached, schedule the redesign or move the program to the reallocate list.
Rewrite your CX strategy as decision rules before an agent has to follow it
Martin Gill, VP and research director at Forrester, published Why AI Is About To Stress-Test Your CX Strategy on October 23. He argues that most CX strategies consist of aspirational statements about delighting customers and earning trust, which inspire employees and give them little guidance for consistent decisions. Experienced employees cover the gap every day: a frontline adviser weighs satisfaction against policy, and a service manager decides when fairness outranks short-term revenue. Agents cannot infer those trade-offs from aspirational language. Gill lists questions a CX team would have to answer before an agent could serve customers, including what delight means in practice, which customers get priority service, and when customer value outweighs margin protection. He frames preparation for agents as focusing "as much on removing ambiguity as on adding intelligence."
Gill prescribes two sets of steps. First, make the strategy executable: define customer value by segment, specify acceptable trade-offs between customer outcomes and business performance, set measurable success metrics, document escalation thresholds and exceptions, and translate principles into decision rules. Second, use agent deployment to tighten decision-making itself, by mapping the customer decisions that matter most to owners, aligning incentives, instrumenting decision points, and refining rules from observed outcomes.
In my AI Capability Framework, goals sit at the apex, and the work of carrying a goal faithfully through hundreds of automated steps lives in the Trust and Identity/Permissions layers. Gill's decision rules are how a CX team writes its goal into those layers. When a CX team hands an agent only aspirational language, the agent's owner learns where the boundaries belonged from customer complaints.
It reminds me of what Jared Navarre, Founder and CEO of Keyni, said when I interviewed him on The Agile Brand podcast: "Judgment often looks like the expensive mistake that never happened…if leadership still calls that hard to measure, the problem is the measurement."
The judgment Gill describes appears nowhere in a budget today, because employees supply it for free inside their regular jobs. Once a team writes it into rules, it gets an owner, a review date, and a line in the capacity ledger, which is why the retailer example assigns 3,000 hours to the work.
My take: Pick the three customer decisions your service and marketing teams argue about most often. Write the trade-off rule for each, name an owner, and put the document in front of every agent that touches those decisions before you expand its autonomy.
Settle decision rights among your executives before anyone presents an AI roadmap
Amii Barnard-Bahn, an executive coach and former Fortune 5 executive, published The C-Suite Behaviors That Stall AI Transformations in HBR on October 23. She opens with a CIO who presents a technically sound AI roadmap to the executive team and a CHRO who learns about it in the same meeting as everyone else, then raises workforce impact and reskilling questions and disagrees with several of his answers. Barnard-Bahn names two behaviors that stall AI programs: claiming sole control over decisions, and disengaging from decisions that seem outside one's expertise. Because AI crosses functional boundaries, she argues, executives have to share authority, resources, and accountability. She recommends that CEOs define the business problem before assigning ownership, build cross-functional relationships before leaders need them, treat early AI plans as experiments, reward jointly owned outcomes, and set rules for resolving disagreements.
Put her CHRO next to Gartner's 78 percent. Most leadership teams have not agreed on where AI creates value, and the executive who owns reskilling often hears about the roadmap last. A CMO who builds a capacity ledger has to hold that conversation, because nobody can assign 14,000 hours without the CHRO, the CFO, and the CMO in the room.
In Marketing Operations 3.0 I recommend a CO2 Council, a cross-functional body that owns the decisions no single function can make, with a charter that spells out its decision rights. Barnard-Bahn's rules for resolving disagreements belong in that charter. Add one decision to the council's list: approving the capacity ledger for every AI program before launch.
My take: Give your CHRO a seat on the body that approves AI programs, and write down who breaks a tie when marketing, finance, and HR disagree about where freed hours go.
Take these actions before your 2027 plan locks
Build a capacity ledger for each AI program. List the hours it should free, the work that receives them, and the owner of that work.
Count only assigned hours. Keep unassigned time out of your Return on Agent and report it as a planning gap.
Label each program with its value path. Efficiency, product improvement, or new offerings, along with the capability each one needs.
Write decision rules for contested customer decisions. Start with the three trade-offs your teams argue about most.
Put the CHRO on the approving body. Workforce decisions and capacity decisions belong in the same meeting.
Fund enablement from the ledger. Assign hours to continuous enablement and leave them out of the value calculation.
Show your CFO where the hours went
I took piano lessons for 12 years, and my teachers never accepted "I practiced for an hour" as a report. They asked which passage I worked on, at what tempo, and what changed by the end of the week. An hour I spent without an assignment usually went to playing pieces I already knew.
Coherence at Velocity asks leaders to measure absorption across three layers. The first is efficiency, meaning less avoidable effort, rework, and resolution time. The second is experience, meaning less friction and more consistent outcomes for customers. The third is equity, meaning protection of Business Lifetime Value by cutting the recurring costs of poor adoption and workarounds. A team that leaves freed hours unassigned improves none of the three layers, and in Gartner's survey employees redirected only about one-fifth of freed time into more innovative work.
Answer the opening question this quarter. For each AI program in your 2027 plan, decide how your team will spend every hour it frees up, write that assignment into a ledger, and have your CMO, CFO, and CHRO sign it before the program goes live.
Featured this cycle:
Gartner HR Research Finds 77% of Board Members Expect AI to Deliver Productivity and Cost Savings, Yet Just 3% of Organizations Exceed Cost Savings Goals Via Productivity. CV Viverito and Arj Bagga, Gartner, October 26, 2026
3 Paths to Delivering Shareholder Returns with AI. Adam Job, Ulrich Pidun, Matthew Kropp, Tania Wang, and Matthew Molloy, Harvard Business Review, October 26, 2026
Why AI Is About To Stress-Test Your CX Strategy. Martin Gill, Forrester, October 23, 2026
The C-Suite Behaviors That Stall AI Transformations. Amii Barnard-Bahn, Harvard Business Review, October 23, 2026