Roles and Responsibilities

The short answer

If you're choosing between RACI and RASCI, the decision comes down to one letter. The two models are identical except that RASCI adds an S - Supportive - for the people who actively help deliver a piece of work without owning it. If your work genuinely has those people - a specialist lending expertise, a neighbouring team providing data or systems, a delivery team supporting a named lead - use RASCI, because naming them is the whole point of the extra letter. If it doesn't - one team, a simple approval chain, everyone already at the table - use RACI, because an empty column helps nobody. And if the thing that keeps going wrong isn't the work but the decisions, the answer may be neither: that's DACI territory, and we'll come to it.

That's the choice. The rest of this article shows what the S actually changes - including the same task list mapped both ways, so you can see the difference cell by cell rather than take it on trust.

What RACI and RASCI share

Take the S out and the two models are the same tool: a responsibility matrix, sometimes called a responsibility assignment matrix. Tasks and decisions run down the side, people or roles run across the top, and each cell gets a letter:

• R - Responsible. The person or people who do the work.
• A - Accountable. The one person who owns the outcome and signs it off.
• C - Consulted. The people whose input is sought before the work moves - a two-way exchange.
• I - Informed. The people who are told afterwards - one-way.

The disciplines are shared too, and they matter more than the choice of model. Exactly one Accountable per task - ownership can't be split, so a row with two A's is a row nobody owns. Keep the Consulted list short, or reviews turn into a queue. And leave cells blank freely - not everyone needs a letter on every row.

Everything that makes RACI work makes RASCI work. The choice between them comes down entirely to whether the fifth role earns a column.

What the S adds - and when it matters

Supportive names the people who actively help those doing the work - lending expertise, resources or hands - without being responsible for delivering it or accountable for the result. The specialist who advises for an afternoon a week. The data team that pulls the numbers. The comms team that turns a policy into something people will actually read.

RACI gives you nowhere honest to put those people, and teams solve that in one of two bad ways. Either the helpers get promoted to Responsible - and now several names share the R, ownership blurs, and you've recreated the exact ambiguity the matrix was meant to remove. Or they get parked as Consulted - which keeps the R clean but misdescribes them, because Consulted people give input while these people do work. Filed as C, their actual workload is invisible: nobody plans for their time, and the work is late the day their "help" turns out to be a dependency.

The S fixes both mis-filings at once. It says: these people are inside the work, their contribution is real, and their time is a named dependency someone plans for - but the Responsible person still owns delivery.

When does that genuinely matter? A useful test: could this work miss its date because a named person outside the core team doesn't show up? If yes, that person is Supportive, and you want the letter on the page. That's common wherever work crosses team boundaries - a process redesign that needs IT, a policy that needs comms, a bid that needs finance's modelling.

And when is it noise? When the honest answer to that test is "nobody". One small team, everyone at the table, help ambient rather than planned - add an S column there and you get a column to argue about, ten-minute debates over whether someone is S or C, and no better delivery. If the S column ends up empty, or everyone is in it, that's the model telling you RACI would do.

One example, mapped as RACI and as RASCI

Take a small, real-shaped piece of work: refreshing a flexible working policy. An HR business partner is drafting it. The HR director owns it. Legal needs to check it before sign-off. The comms team will turn it into manager guidance and publish it. And line managers need to know when it lands.

Mapped as RACI, the honest version looks like this:

Look at the comms team's column. RACI leaves nowhere honest to put them: they're not Consulted - nobody is asking their view on flexible working - so they end up Responsible, and now two teams share the R on the guidance and the announcement. When the guidance slips, whose task was it? That's the exact question the matrix exists to prevent. (The other common dodge is to file them as C, which keeps the R clean but makes their real work invisible.)

The same tasks as RASCI:

Two cells change, and both now say something more truthful: the comms team actively helps deliver, the HR partner still owns delivery, and the comms team's time is a named dependency that gets planned rather than assumed. That's the whole difference between the two models. If those two cells wouldn't matter in your work, neither would the change - and RACI will serve you fine.

RACI, RASCI or DACI - how to choose

In practice the choice usually announces itself. A project where a delivery team supports a named lead - a service redesign with IT and a data analyst alongside an operations lead, say - is RASCI-shaped: the support is real, and naming it is the point. A simple approval chain - a report drafted, checked and signed off inside one team - is RACI-shaped: the leaner model is easier to keep alive, and an S column would sit empty. And if what keeps failing is decision-shaped - the call stalls for weeks, or gets made and then quietly remade - the fix isn't a better task matrix at all. The DACI Framework maps decision rights rather than task roles: who drives the decision, who approves it, who contributes, who's told.

The models also combine without fuss. Plenty of teams run a RASCI for the delivery work and reach for DACI on the two or three decisions that genuinely need their rights spelled out.

Conventions, not standards

One honest caveat before you commit to either. There is no standards body for responsibility matrices. RACI, RASCI and DACI - along with RASIC, CAIRO and the other variants you'll meet - are conventions that grew out of decades of practice, not specifications. There's no official definition to comply with and no canonical version to appeal to when colleagues disagree about what a letter means.

Which is exactly why the choice of model matters less than the discipline around it. Pick one. Write down what each letter means in your context - one line per letter is enough. Use it consistently across the team, and resist improving it mid-project. A team with four letters and one shared definition will beat a team with five letters and five private interpretations every time.

If you land on RASCI, our full RASCI Framework guide covers each of the five roles in depth, how to build the matrix step by step, and the mistakes that trip teams up - and there's a free RASCI template (PDF) to run it with.

RACI vs RASCI FAQs

Is RASCI better than RACI?

Neither is better; they fit different work. RASCI is the better tool when delivery genuinely depends on people who help without owning - it makes that support visible and plannable. RACI is the better tool when it doesn't, because the leaner model is easier to keep alive. The worst outcome isn't picking the "wrong" one - it's an untended matrix of either kind, describing a team that no longer exists.

What's the difference between RASCI and RASIC?

Only the order of the letters. RASIC names the same five roles - Responsible, Accountable, Supportive, Informed, Consulted - and works exactly the same way. Use whichever spelling your organisation already recognises - the reordering isn't a different method.

Can we just add an S column to our existing RACI?

Yes - that's literally what RASCI is. If your RACI keeps producing rows where two people share the R and one of them is really a helper, adding the Supportive role is the fix. Define it in writing - actively helps deliver, doesn't own the result - tell the people it applies to, and carry on. No relaunch required.