Dear COO, This Is Your Moment. Here Is What Will Sabotage It.
Share
BCG just handed the COO the biggest mandate of the decade. In its 2026 report, BCG argues that the COO's job is no longer running a function. It is orchestrating agentic AI across the entire enterprise (BCG).
Read that again. The COO — historically the deepest technical operator in the company — is now the person who decides how AI reshapes the business.
This is your moment.
But there is a problem sitting between you and that mandate. A problem BCG has documented three separate times, and almost nobody in your C-suite is talking about it.
The Three Numbers That Should Terrify Every COO
Before you accept the AI orchestrator mandate, look at what BCG's own research says about the odds.
Number one: 75% of transformations fail. BCG's global analysis found that only 1 in 4 transformations deliver value-creating, enduring change. That failure rate has not improved in decades (BCG).
Number two: 70% of AI value is not the AI. BCG's 10-20-70 rule states that "10% of a company's efforts should be focused on algorithms, 20% on technology and data, and the remaining 70%—the lion's share—on people and processes" (BCG).
Number three: the COO now owns all of it. BCG's 2026 piece is explicit — the COO must "move beyond optimizing individual parts of the value chain" so that "the whole system can respond as one" (BCG).
Stack those three numbers on top of each other and the math is brutal. You are being asked to lead the highest-stakes transformation in your organization's history, in a discipline where three out of four attempts fail, using a lever where the technology itself accounts for less than a third of the value.
The AI is not going to save you. The vendor demo is not going to save you. The consultant deck is not going to save you.
Why Most COOs Will Fail at This
Here is the trap. Most COOs will read the BCG article, agree with the thesis, and immediately do the wrong thing.
They will buy a forecasting agent. They will pilot an orchestration copilot in procurement. They will approve a supply-chain AI workstream. They will spin up what BCG calls "a thousand blooming flowers" — a scatter of point solutions that look attractive in isolation and shatter the moment you try to connect them (BCG).
The result is predictable. Six months in, the pilots deliver marginal gains. Twelve months in, the board asks why the AI investment has not moved the P&L. Eighteen months in, the COO is back to defending headcount reductions instead of designing new value pools.
This is the 75% failure path. And it fails for one reason.
You cannot orchestrate broken processes.
An AI agent placed on top of a broken workflow inherits every hand-off, every wait state, every duplicate approval, every reconciliation loop, every silo. It runs faster. It fails faster. It does not transform anything.
BCG's own recommendation makes this concrete. They tell COOs to resist mapping the current process in detail before applying AI, because "this approach can place artificial constraints on the design." They tell you to start with value pools, decisions, and data — not with the workflow you already have (BCG).
That is a nice principle. It is also almost impossible to execute inside a bank, an insurer, a manufacturer, or any enterprise where processes have accumulated over decades in dozens of disconnected systems. You cannot redesign what you cannot see. You cannot see what your policy library, your SOP repository, your BPM tool, your audit archive, and your intranet each describe differently.
This is the layer BCG names but does not solve. In the 2026 article, BCG explicitly admits: "I have not seen anyone fully solve that yet" (BCG).
Somebody has to solve it. It has to be the COO. And it has to happen before the orchestrator agents get switched on.
What "Process Transformation" Actually Has to Do
If 70% of AI value is people and processes, then a serious COO needs a serious answer to seven questions. Not one of them is about the AI model.
- Which end-to-end processes actually generate value in this business, and which just generate paperwork?
- Where does value stop flowing — where are the hand-offs, wait states, duplicate approvals, and reconciliation loops?
- What does the redesigned future-state process look like, before any agent is deployed?
- Who owns each decision in that future-state process, and how do those decision rights change when an AI is in the loop?
- What documentation — policy, procedure, RACI, OLA, FMEA, control matrix — has to exist for auditors, regulators, and risk to sign off?
- How will front-line staff actually adopt the new process, and where will they push back?
- How will the process keep improving after go-live, so that the transformation compounds instead of decaying?
Every one of these questions is a process problem. None of them is solved by buying an orchestration agent.
The COOs who win the AI decade will be the ones who build a process transformation layer underneath the orchestrator. The COOs who lose will be the ones who skipped that layer and wondered why the agents did not scale.
Meet ESSAM: The Process Transformation Layer for AI Orchestration
This is the gap ESSAM was built to close.
ESSAM is an E2E Agentic Process Transformation OS designed for the exact situation BCG describes — an enterprise that needs to redesign the processes underneath AI before it deploys the agents on top.
It is not a policy library. It is not an SOP repository. It is not a generic BPM tool. It is a closed-loop transformation cycle that takes a process from raw baseline to redesigned future state, complete with the documentation, controls, and adoption mechanics an enterprise actually needs.
Here is what the cycle does:
- Baseline — ESSAM captures the current process through conversational mapping, document ingestion, or native Visio import. No multi-year data-cleansing program required.
- Analyze — The platform runs value and non-value classification, waste detection, and process statistics to expose where flow actually breaks.
- Redesign — One prompt generates a redesigned future-state process, complete with change-degree assessment and FMEA on the risks the redesign introduces.
- Document — ESSAM auto-generates the SOP, OLA, RACI, board-ready deck, PDCA A3 export, and policy or procedure draft that governance, risk, and audit will demand.
- Govern — Approval routing, version control, audit trail, and signature workflows sit inside the same platform, not in a separate document management stack.
- Deploy — Front-line staff receive the new process on the channel they already use — Microsoft Teams or WhatsApp — instead of being forced into a new portal.
- Feedback and re-optimize — Every deployed process collects staff feedback and feeds it back into the AI, which surfaces the next waste to remove.
- Discover AI use cases — Once a process is redesigned and running, ESSAM identifies where agentic AI belongs on top of it, ranked by impact.
That last step is the one that matters for BCG's orchestrator mandate. You do not deploy AI agents into a fog. You deploy them into a redesigned, documented, governed, adopted process — with a ranked backlog of exactly which decisions inside that process should become agentic.
That is what "the whole system responds as one" actually requires.
The Assurance Layer Enterprise Buyers Ask For First
Every COO reading this has the same next question. What about the auditors, the regulators, the risk committee?
ESSAM is certified GDPR, ISO/IEC 27001:2022, and SOC 2 Type II. Documentation control is treated as the floor of the platform, not the ceiling — versioning, routing, approval, audit trail, signatures, and controlled libraries are all built in. The category claim is not about better paperwork. It is about making value flow faster by identifying and removing waste, on top of the assurance base that enterprise procurement demands.
This is what ESSAM ships right now; not in the future.
The Decision You Are Actually Being Asked to Make
BCG's 2026 report is a gift to every COO who reads it carefully. It says, in plain language, that the mandate is yours, the failure rate is 75%, and 70% of the value lives in people and processes — not in the algorithms.
That leaves you with a simple decision.
You can accept the AI orchestrator mandate and skip the process layer. That is the 75% path. You will spend the next two years explaining pilots that never scaled.
Or you can accept the mandate and build the process layer first. That is the 25% path. It is harder to sell to a board that wants to see AI in the demo. It is the only path where the AI actually works.
The COOs who choose the second path will define this decade of operations leadership. The rest will be replaced by the ones who did.
Your move.
If you are ready to see the process layer in action, book a 30-minute walkthrough of ESSAM at thestrategist.me/pages/essam. Bring one broken process. Leave with a redesigned future state, the FMEA on its risks, and a ranked list of the AI use cases that belong on top of it.
Written by Abdulla Al-Awadi — Chief Strategy Officer, Kuwait International Bank, and founder of TheStrategist.me and ESSAM.ai.