The Jira Life Down Under: Key Takeaways from Our Session on Atlassian's Pricing Changes

On 24 September 2026 we ran a live session with Alex Ortiz and Rodney Nissen, hosts of The Jira Life, alongside our founder Brett Celliers.

The reason for the session was pretty simple. Atlassian has announced two things ten weeks apart: a Cloud list price increase from 13 October 2026, and usage-based billing for AI and automation from 3 December 2026. Everyone has read both announcements. Almost nobody has sat down and run the numbers on what they mean together.

Roughly an hour, no slides, plenty of questions from the audience. Here are the themes that stood out.

1. The first real numbers are in, and the surprises aren’t where people expected

When the metering announcement first landed, the question everyone asked was “can we see the numbers yet?” Now the reporting is there, and the early read is less dramatic than feared.

Rodney’s take was that Enterprise customers, the group he was most worried about, are largely fine. The limits are generous enough that most are still comfortably inside them.

But the shape of the change is worth noticing. Alex has seen orgs go from using one to three percent of their automation allowance to seventy or eighty percent. You’re still under the cap. You’re just not remotely near it anymore.

And the genuine surprise? Bitbucket build minutes. Not automation. That was the one that caught a customer Rodney works with off guard, and a lot of it came back to pipelines that were never set up with efficiency in mind.

2. The winners and losers are already obvious

Rodney framed it neatly. Teams that are heavy on licences but light on usage are coming through this fine. No real issues.

The teams that are hurting are the ones who went the other way: trimmed licences down and automated the gap. If you’ve spent the last few years replacing seats with rules, the meter is about to price that decision differently.

Alex added the governance angle. Organisations that put guardrails around automation, and ran internal sessions on when to build a rule and when not to, are going to be fine. It might sting, but they’ll cope. The ones who have been moving fast without those boundaries are the ones who get bitten harder.

3. The headline percentage is not your real increase

This was the number that stopped the room.

Take a 100 user deployment on Service Collection. The October increase for Premium and Enterprise is around 7.5%, so far so manageable.

Now layer the metering on top. Atlassian’s own published guidance defines light Rovo adoption at 500 credits per user per month and heavy at 2,000. A mid-level adopter sits around 1,250. Service Collection includes 700. So a mid-level adopter is paying for roughly 550 credits per user per month that they don’t have.

Combine the two and that 7.5% becomes something closer to a 28% increase in total cost. On standalone products, where the included allowances are much smaller, Brett has modelled scenarios landing anywhere from 20% to 60%.

The uncomfortable part isn’t the price rise. It’s that the number you’ve been planning around might be off by a factor of four.

4. “Buy more usage” is on by default, and that’s a decision, not a setting

Every org has a toggle that lets it automatically purchase extra usage when it runs out. It’s on by default.

Alex’s view was clear: turn it off. Work inside the constraint first.

His reasoning is worth repeating. Sloppiness expands to fill whatever budget you give it. Automation sloppiness, Rovo sloppiness, all of it. Put a hard boundary in place and your team will refactor, rethink and find a way to work inside it. Leave the tap running and they never will.

There’s a second reason, which is visibility. If overages quietly become a line item you never see, you lose the signal that something is broken. Hitting a limit should prompt a conversation about whether you need to buy more or clean something up. It can’t do that if it resolves itself in the background.

One gotcha on the buying side: Atlassian’s documentation puts the floor at a US$20 block, which buys 40,000 automation steps at 50 cents per thousand. Go 500 steps over and you’re still buying the whole block.

5. Start with what you can actually control

The October increase isn’t something you can negotiate away. It happens every year, and as Alex pointed out, October 2027 is already on the horizon.

What you can control is what you’re paying for right now.

• Inactive users – if someone hasn’t touched the tool in 90 days, that’s a conversation with their manager, not a renewal line item.
• Invited users – this one catches people out. Every outstanding invite bills you monthly, whether that person ever logs in or not. Share a work item with a vendor, they loop in two colleagues, and suddenly you’re paying for people you’ve never met.
• Marketplace apps – Atlassian isn’t the only one putting prices up. Add-ons are frequently the biggest driver of true per-seat cost, and several vendors have announced increases of their own.

Usage data used to mean exporting CSVs and modeling them in a spreadsheet. It’s now built into the admin UI with 30, 60 and 90 day activity windows, which makes the tier-down conversation much easier to have.

This is the work that pays. When we ran a subscription optimisation programme with Zip, licence hygiene and removing add-ons made redundant by native functionality took 36% off their licensing in year one and more than US$1M over five years. None of that required a pricing announcement to justify.

6. For a lot of organisations, collections are the obvious move

The maths here is genuinely straightforward.

Service Collection costs the same as JSM standalone. You get Assets, you get CSM, and you get roughly ten times the usage allowance. Rodney correctly called the JSM to Service Collection move close to a no-brainer, and it’s worth noting new customers can’t buy JSM standalone at all anymore.

Teamwork Collection is the same price as Jira plus Confluence bought separately. Loom comes along for free, plus the bigger shared pool for Rovo credits and automation steps.

Where it doesn’t hold up: don’t move to a collection purely to chase allowance. Software and Strategy Collections go through a sales conversation rather than a self-serve upgrade, so if you’re not going to use the products, you’re buying headroom you could get more cheaply elsewhere. Run the comparison on your actual consumption, not on the brochure.

7. Your sandbox is quietly drawing from the same pool

This one surprised a few people in the chat.

Automation rules live at the org level, not the site or product level. So every step you burn experimenting in sandbox comes out of the same allowance as production. Worse, adding a new licensed site adds to your pool. A sandbox doesn’t. It draws down without contributing.

Two practical things to do this week:

• Turn off scheduled rules in sandbox. If you’re debugging an automation and leave a scheduled rule running, it keeps consuming after you’ve stopped paying attention
• Check when your rules actually run. Alex’s point: if nobody works weekends, the rule probably doesn’t need to run on weekends either

8. Rovo is very good at one specific thing

We asked both hosts how Rovo stacks up against the alternatives, and the answer was more nuanced than either camp usually allows.

Alex’s case for it: Rovo knows your data better than any general model does. Twenty-plus years of work history, team dynamics, where things get blocked. If the question is “what’s the best AI to help me do the job I’m actually paid to do,” that context is hard to replicate.

Rodney’s counter, as an active Rovo user: it’s good inside the Atlassian world, particularly for pulling together Confluence content and work items into something usable. It’s much better than it was. But when he’s spending his own time and money, he’s using Claude with MCP access to Jira and Confluence instead.

And Alex’s concern is timing. Across the dozen or so teams he works with weekly, two use Rovo. Metering an AI product before adoption has taken hold puts a price on something many teams haven’t yet found a reason to use.

The honest question to ask yourself: do you need an AI that knows your work that well? If the answer is no, you have cheaper options.

9. This is an enterprise strategy, and it explains everything else

Alex had been through Atlassian’s Q4 earnings report the morning of the session. The recurring theme was enterprise customers: year on year growth, the number spending over a million a year, the percentage above three million. Alex noted the C-suite hires over recent years have come predominantly from enterprise application backgrounds.

Read the pricing changes through that lens and they make more sense. A 30% increase might push smaller customers toward ClickUp, Asana or Monday. That’s a trade Atlassian appears willing to make, because acquiring and servicing one 30,000 seat enterprise customer is far more efficient than doing it six or seven times over with small ones.

The honest bit: if you’re frustrated and thinking about leaving, that’s fair, and both hosts said as much. But there is no single product that replaces the full Atlassian suite. ClickUp is a genuinely strong competitor and we work with it, but it wasn’t built for software engineering teams. Linear is winning startup attention. GitHub and GitLab are creeping into the project management space. None of them is a one-for-one swap, and the switching cost is usually the number that ends the conversation.

Which is exactly why Atlassian feels comfortable raising prices.

Where to start

If you do one thing before December, do this. Sign in as an org admin, go to the gear icon and into your org administration at admin.atlassian.com, then look for Insights and Platform usage in the bottom left.

You’ll see 90 days of history across Rovo credits, automation steps, CSM agent resolutions and Assets objects, broken down by daily consumption. That’s enough to model the next twelve months with a sensible buffer, and enough to know whether you’re a “turn the toggle off and tidy up” org or a “have a conversation about collections” org.

You don’t have to guess. Atlassian has given you the notice, the data and the tools.

What’s next?

The full recording is available on demand on the event page. If you want the detail behind the two announcements, we’ve written them up separately: the October price increase and the December metering changes.

And if you’d rather not model this yourself, we’re offering a free Atlassian Usage Readiness Check. We’ll look at your actual consumption against the December limits, flag where you’re exposed, and tell you whether a collection move or a licence tidy-up is the better lever. It’s the same approach we took with Harrison AI, where consolidating onto Teamwork Collection Premium avoided AUD 229k over 24 months.

Get in touch with our team for a no-obligation conversation.