A few years ago, I watched a “simple” project die in slow motion: not because anyone was incompetent, but because every team spoke a different dialect—spreadsheets here, PDFs there, a project tool nobody else logged into. We held meeting after meeting just to translate. That’s why the idea that AI’s Big Payoff is coordination—not automation—lands so hard for me. The real cost isn’t the work; it’s the handoffs.
The unsexy problem: Translation Costs in Core Business
In my Harvard Business Review article, “AI’s Big Payoff Is Coordination, Not Automation” (February 25, 2026), I argue that the biggest drag on Core Business performance is often Translation Costs: the hidden work of turning one team’s output into another team’s usable input.
What “translation” looks like in real Business Operations
It’s not abstract. It’s rework, meetings, misalignment, and the dreaded “can you re-send that in my format?” loop—usually across emails, spreadsheets, PDFs, and specialized industry software. I’ve watched a single “PDF-only” stakeholder become the bottleneck: everyone else had live data, but decisions stalled while someone manually retyped specs from a scanned attachment.
“Most organizations don’t lose time to ‘bad decisions’—they lose it to decisions that arrive too late, after the translation work has quietly eaten the calendar.” —Tom Davenport
Why standardization stalls (and why it hurts)
Most standardization programs promise Cost Reduction, but they bog down because consensus is slow. Worse, enforcing one workflow can crush local autonomy—teams lose tools that make them fast, accurate, and accountable. The result is a compromise system that nobody loves, plus a long tail of exceptions that recreates the translation mess.
My gut-check for Cross Functional friction
If a project needs a weekly “alignment meeting,” it probably has a translation problem, not a people problem. That’s where AI creates real Efficiency Gains in 2026: not by replacing every task, but by embedding a translation layer into core operations so cross-functional work can move without waiting for perfect formats. Automation and predictive analytics still matter—they’re just secondary benefits when coordination finally gets cheap.
Coordination Without Consensus: What AI Integration actually changes
When I talk about AI Integration, I’m not talking about replacing people or forcing everyone onto the same system. I’m talking about reducing the “translation” work that quietly eats budgets: turning one team’s output into another team’s usable input.
The mechanism: structure from mess, then continuous translation
Most coordination breaks down because the inputs are messy and scattered—emails, spreadsheets, PDFs, and notes buried in specialized tools. Modern Machine Learning can extract usable structure from that mess (entities, dates, quantities, dependencies), then keep translating in the background as new information arrives. The result is a searchable project record that stays current, even when teams never change their tools.
Why this isn’t old-school interoperability
Traditional interoperability tries to “standardize first”: agree on fields, formats, and workflows. That usually fails because it’s not a technology problem—it’s incentives and autonomy. As Andrew McAfee puts it:
“Interoperability is rarely a technology problem; it’s an incentives problem. AI helps by lowering the coordination tax without demanding a political referendum.” —Andrew McAfee
This is why I’m skeptical of “single source of truth” slogans. What AI Solutions make realistic is closer to a single source of usable answers, assembled from incompatible platforms and reconciled as conditions change.
Decision Making improves because the record becomes queryable
Once everyone can query the same living record, Decision Making gets faster and more Data Driven. Real-time analysis and predictive trends help teams spot conflicts early (scope gaps, schedule risk, cost drift) without another meeting. But it only works when data readiness is real—clean access, permissions, and consistent capture—so AI outcomes move beyond hypothetical New Capabilities into daily execution.
Construction’s Efficiency Gains: Autodesk, Procore, and Trunk Tools
Efficiency Gains in “many tools, one job” work
On a real project, the architect lives in Autodesk, the PM team tracks commitments in Procore, and the field crew speaks in RFIs, submittals, and “what changed since yesterday.” Same job, different vocabularies, different files, different clocks. That gap is where construction tech AI creates the biggest Efficiency Gains—not by replacing people, but by cutting the hidden “translation” work that slows Business Operations.
AI integration that reduces rework and improves Quality Assurance
When AI integration works, specs, drawings, schedules, and meeting notes stop being scattered artifacts and become a structured, searchable, current project record. Tools like Trunk Tools can pull from PDFs, plan sets, and platform data, then reconcile what’s “latest” across systems. That directly supports Quality Assurance: fewer teams building off old details, fewer late surprises, and fewer coordination meetings just to confirm the basics.
Here’s the lived-in moment: a foreman on the jobsite searches “latest doorway spec” and gets the current detail, the related RFI, and the approved substitution—without calling three people or waiting for a meeting. AI won’t stop change orders, but it can stop the avoidable ones caused by stale info.
“The jobsite isn’t short on data—it’s short on shared context. The winner is whoever turns scattered artifacts into a usable narrative.” —Jeanne Gang
Market signal: Procore + Datagrid
Procore’s acquisition of Datagrid is a clear bet that AI-powered translation will be core infrastructure. In 2026, the organizations seeing the best results embed AI into core operations for efficiency and profitability—not as a side tool, but as the coordination layer.
Claims Processing without retraining the world: CCC vs Tractable
CCC’s old edge: standard workflows (and massive switching costs)
For years, CCC Intelligent Solutions built its Competitive Advantage in Claims Processing by defining the “right” digital workflow. That delivered real Workflow Automation, but it also locked insurers in: changing vendors meant retraining adjusters, reworking integrations, and rewriting operating rules. The hidden tax wasn’t software—it was the translation work between people, systems, and formats.
What changed in 2023: AI coordination as a translation layer
In 2023, Tractable flipped the model. Instead of forcing a new standard, it used AI to read raw smartphone photos of vehicle damage and produce repair estimates that fit existing insurer processes. Tractable handled nearly $7 billion in claims that year, largely because it reduced friction: the AI translated messy inputs into usable outputs without demanding a new daily routine.
“In insurance, the fastest innovation is the one that respects the adjuster’s morning—because that’s where the real constraints live.” —Daniel Schreiber
Why this matters to business leaders: adoption rides muscle memory
The strategic lesson is simple: AI coordination wins when it rides existing muscle memory instead of fighting it. That’s where Cost Reduction shows up fastest—less rekeying, fewer escalations, fewer meetings to reconcile mismatched data. It also matches broader adoption patterns: 75% of workers report faster outputs with AI, and 87% report quicker IT resolutions; among SMBs using AI, 82% cut costs and 91% saw ROI.
- Wild-card hypothetical: an insurer pilots in one region in 30 days because the “translation layer” plugs in lightly—no workflow rewrite required.
Supply Chain coordination: project44, Maersk, and FedEx strategies
In Supply Chain work, I keep seeing three strategies as AI integration lowers “translation” costs: become the translation layer, double down on Accountability Systems, or monetize privileged views. I used to think “visibility” was a feature; now I think it’s a business model.
- Translation layer: unify messy data without forcing one standard.
- Accountability: pair coordination with guarantees, contracts, and Risk Management.
- Privileged views: control the network view and sell selective access.
project44: Operational Agility through “coordination without consensus”
project44 fits the translation-layer play. It offers an integrated, real-time view of global shipments while letting carriers keep their own tools, formats, and processes. That’s the hidden win: faster exception handling and higher Operational Agility without a painful standards war. As translation gets cheaper, the platform that reconciles status updates, ETAs, and exceptions becomes hard to replace.
Maersk and FedEx: coordination creates power—and liability
Maersk leans into end-to-end logistics with stronger governance and Risk Management. The more you coordinate across modes and partners, the more customers expect outcomes, not dashboards. That’s where responsible AI governance with human oversight matters in 2026 workflows: it builds trust when models disagree, data is missing, or disruptions hit.
FedEx invests in AI-driven routing to create a privileged internal view of its network, then selectively monetizes that insight. This is where agentic AI can work—if it operates with guardrails focused on execution, escalation, and audit trails.
“When everyone can see the shipment, the game shifts from visibility to responsibility—who pays when the prediction is wrong?” —Yossi Sheffi
Scaling AI in 2026: AI Governance, Data Readiness, and the messy human stuff
In my HBR piece, I argue AI’s biggest payoff is coordination—AI as the hidden translation layer. But Scaling AI breaks the moment trust breaks. Translation is easy to demo; it’s harder to run in production when liability, customer impact, and vendor contracts are on the line. As translation costs fall, issues of accountability and responsibility rise—fast.
“Governance isn’t the brake on AI—it’s the steering wheel. Without it, ‘coordination’ just means faster mistakes.” —Satya Nadella
AI Governance that survives real Business Operations
In 2026 workflows, responsible AI Governance with human oversight is what earns adoption and keeps it. I keep Compliance Management practical: enough structure to prove what happened, why it happened, and who can stop it.
- Human oversight for high-stakes decisions (safety, money, legal).
- Audit trails: prompts, sources, model versions, and approvals.
- Clear escalation paths: who gets paged, and when the AI must defer.
- Contracts that match reality: SLAs, data rights, and error ownership.
Data Readiness: the difference between pilots and measurable gains
Data Readiness isn’t glamorous, but it’s how coordination becomes Measurable Gains. We’re already seeing productivity lift—75% of workers report faster outputs and 87% report quicker IT resolutions. SMBs report outcomes like 93% revenue growth, 82% cost cuts, and 91% ROI when AI is operationalized with governance.
What I ask in the first steering meeting
- Executive sponsorship tied to one operational metric.
- Multidisciplinary team: ops, IT, legal, risk, and frontline users.
- Minimum data inventory: systems, owners, quality, and access rules.
- Manager enablement: HBR’s 10 Must Reads for New Managers and The Leadership Unblocked Toolkit.
Conclusion: The Big Payoff (and my Call to Action)
The Big Payoff: Coordination Not just Automation
If you remember one idea from my HBR work, let it be this: the Big Payoff from AI is Coordination Not automation. Yes, there are real Automation Benefits. But the bigger win is lowering translation costs—the meetings, rework, and “can you resend that in my format?” moments that quietly drain execution. In the intro anecdote, the problem wasn’t too few people. It was too many lost hours between people.
This is why the coordination win compounds. Every time AI translates a handoff—drawing to schedule, email to spec, photo to estimate—it makes the next handoff faster, cleaner, and easier to trust. That flywheel is one of the most important AI Trends 2026 for operators: speed comes less from doing tasks faster and more from removing friction between tasks.
“Strategy is choosing what not to standardize. AI lets you keep the differences that create value—while reducing the friction that creates waste.” —Michael Porter
My Call to Action for Business Leaders
My Call to Action for Business Leaders is simple: pick one translation choke point—one painful interface between teams, tools, or data—and fund it like a mission-critical system, not an “innovation experiment.” Put governance around it. Define accountability. Measure cycle time and rework, not just model accuracy.
Picture AI as the quiet interpreter in the corner of the room. If you treat it like a toy, you’ll never trust it with the negotiations. Treat it like infrastructure, and you’ll earn the coordination advantage that lasts.
