Ask a leadership team how to close accountability gaps in organisations, and most of them start in the same place, which is the wrong place. It comes up almost every time I speak at a quarterly business review or a functional leader forum: the leader raising it has already gone back through the RACI, checked that every task has a name against it, and concluded the accountability is sound. It usually is, inside the boxes. The gaps are never inside the boxes. They sit in the joins between functions, and that is precisely why a document built to describe boxes cannot see them.
This is a system problem, not a person problem. The chief operating officers, chief transformation officers and divisional managing directors I work with are not short of owners. Every workstream in their transformation has a name attached. What they are short of is a mechanism for the moment two of those named owners look at the same trade-off and reach different conclusions. That moment, not the org chart, is where cross-functional outcomes are won or lost.
Accountability Gaps Are in the Joins, Not the Boxes
Picture a fairly ordinary functional organisation working on a cross-functional outcome: cutting the time it takes to resolve a complex customer complaint end to end, which touches operations, technology and customer service. Each function nominates a lead. Each lead owns their slice cleanly. Operations owns the resolution process, technology owns the way a case is tracked between teams, customer service owns what the frontline says and does. On paper, full coverage.
The gap shows up at the edge, where operations needs a change technology has not prioritised, or where customer service needs two extra weeks that operations does not have. Nobody owns that seam, because the org chart was never drawn to show seams, it was drawn to show functions. The result is Ownership Fog, the condition where every task in a cross-functional initiative has a named owner and the outcome itself has none.
This is the pattern I keep returning to in my work on cross-functional execution: the outcome sits across functions, but the accountability tools we use were built one function at a time.
What a RACI Can Tell You and What It Structurally Cannot
I want to be fair to the RACI. It is not a bad tool, it is a misapplied one, genuinely good at answering a task-level question: who does the work, who signs off, who needs to be asked first, who needs to be told afterwards. For a defined task with a defined owner, that structure holds up well.
What it cannot do, by design, is tell you who decides when two accountable people disagree about something the task itself did not anticipate. RACI confusion is rarely a data entry problem, where someone simply forgot to fill in a cell. It is a structural blind spot: the format has no cell for the question that matters most, which is who decides when this goes sideways. You can fill in every row of a forty-row RACI with total diligence and still have zero mechanism for resolving the one disagreement that actually threatens the outcome.
Three Accountable Owners Is a Record of an Unresolved Argument
Here is a composite I see often enough to call it a pattern rather than an anecdote. Call it the complete RACI. Forty rows, every cell populated, submitted on time, genuinely well-intentioned. Against the outcome that mattered most, though, three names sat in the accountable column, not one.
Nobody did that carelessly. Naming a single accountable owner meant telling two other senior, capable leaders that their function did not have the final say on something they had real skin in. Writing three names in felt like respect for people who had earned it.
What it actually was, was a record of an argument that never got resolved, written down in a format that made the lack of resolution look like completeness. Nobody would ever write “we could not agree who decides” into a governance document. Three accountable owners says exactly that, in a shape that passes every audit, because an audit checks that the cell is populated, not what decision it can produce. In that composite, the first real trade-off, delay the delivery date or proceed without the integration, sat unresolved for five weeks. Nobody was negligent. The document simply never said whose call it was.
The Question That Actually Matters: Who Decides When Two People Disagree?
Every accountability gap I am asked to help close comes down to one missing sentence: when these two people disagree, this one decides. Not consulted, not informed, decides.
Diffusion of responsibility is doing quiet work underneath that missing sentence. The social psychology research on the phenomenon, well established since the 1960s, shows that the more people who share responsibility for an outcome, the less likely any one of them is to act, because each assumes someone else will. Three accountable owners does not triple the chance the trade-off gets resolved, it divides it. Naming a single decider is not a demotion for the other two, it is the only structural move that actually produces a decision instead of a stalemate.
How to Write Decision Rights for a Cross-Functional Outcome
Decision rights are less fashionable than a RACI, and considerably more useful, because they answer the question the RACI cannot. For every cross-functional outcome that matters, I ask leadership teams to write down four things, not forty rows, four:
Who decides. One name, not a committee, not a function, a person.
Who must be consulted before that decision is made. This preserves the input the other accountable leaders genuinely deserve, without diluting who holds the call.
What the decision maker can decide without escalating. The boundary of their authority, stated plainly, so they are not guessing whether this is theirs to settle.
How long a disagreement can stay open before it escalates automatically. This is the line almost every leadership team leaves out, and it is the one that turns a document into a mechanism.
Put a Clock on Disagreement
This is where I apply The First 48 Hours, which in my work generally means a decision is only believed when something visibly stops, starts or changes in the first 48 hours after it is made. Applied to accountability gaps specifically, the corollary is this: any cross-functional disagreement that has not produced a decision within 48 hours escalates automatically, without anyone needing to find the courage to raise their hand.
That last clause matters more than it sounds. The most expensive accountability gaps are rarely the ones nobody noticed. They are the ones several people noticed, where escalating would have meant naming a peer as the source of the delay, so everyone waited. A 48 hour clock removes that interpersonal cost. Nobody has to decide to be the one who speaks up, the clock decides for them.
How to Audit Your Seams in an Afternoon
You do not need a quarter-long review to find where your accountability gaps sit. Gather the leaders of every function involved in your top three or four cross-functional outcomes and ask one question of each: name the last disagreement between your function and another that took more than a week to resolve. You will not need many examples before a pattern appears, usually the same two or three joins, showing up in every leader’s answer independently.
For each one, write the four decision rights lines above. Where two accountable leaders currently sit against the same outcome, decide, together and out loud, which one holds the call, and treat that conversation as clarity, not demotion. Where a chief financial officer or divisional leader can already point to a disagreement sitting unresolved right now, that is not dysfunction, it is the seam showing itself, and it is the fastest place to start.
For a closer look at what this same problem does when a person genuinely reports to two bosses with competing legitimate claims, rather than the ordinary functional setting I have described here, see my piece on matrix organisation accountability.
Further reading: Harvard Business Review’s research on decision rights, Who Has the D? How Clear Decision Roles Enhance Organizational Performance, is a useful companion to this argument.
If a cross-functional disagreement in your organisation has quietly outlived its second week, that is the clearest evidence I can point to in front of a leadership team about where execution is actually breaking down. If you want that named at your next leadership forum or quarterly business review, book a conversation with me here.

