Walk through most Fort Wayne offices right now and you'll find the same quiet story. Someone in sales writes proposals in half the time with ChatGPT. Someone in the back office cleans up spreadsheets with Copilot. Everyone is a little faster. And yet, when the owner looks at the numbers at the end of the quarter, the company isn't measurably faster. Revenue per employee looks about the same. The backlog hasn't shrunk. Nobody can point to the return.
That gap — faster people, same company — is now the central finding of the biggest teamwork research of the year, and it has a name. New research from Atlassian's Teamwork Lab, first surfaced this week by VentureBeat, calls the drag an “AI fragmentation tax.” The headline number is stark: 89% of executives say AI has increased the speed of work, but only about 6% are confident they can point to clear AI ROI across their organization.
This isn't an argument against AI. Your people really are faster. The problem is that the speed is trapped at the individual level and never wired into the way the company actually runs. Below, we break down what the research found, why Fort Wayne and Northeast Indiana businesses are especially exposed to it, and the concrete moves that lift AI from a personal productivity trick to organizational return.
Key Takeaways
- Atlassian research found 89% of executives say AI increased speed, but only about 6% are sure of clear org-wide ROI.
- Atlassian estimates the resulting “fragmentation tax” at roughly $161 billion a year across large enterprises.
- Only about 14% of teams have cracked organizational AI ROI — and they're far more likely to say AI improves planning and collaboration.
- 85% of knowledge workers use AI at work, but only about 29% have integrated it into their actual workflows.
- The fix isn't more individual AI seats — it's operating AI at the team and process level, owning a workflow end to end.
- A local diagnostic can tell you in an afternoon whether your AI spend is stuck at the individual level.
What is the “AI fragmentation tax”?
The core idea from Atlassian's State of Teams 2026 research is that AI has been adopted one person at a time, and individual speed doesn't automatically add up to organizational speed. When everyone accelerates independently, without shared context or connected workflows, you don't get compounding gains — you get duplication, rework, and misalignment.
Atlassian's behavioral scientist Liz Fosslien put the paradox plainly in the company's analysis of why AI speed isn't delivering ROI: “If everyone's AI-efficient...you'd expect 10x momentum across the company. But what's happening instead is that teams are often getting 10x the chaos.” The research puts a price on that chaos — an estimated $161 billion a year in lost value across large enterprises, the “fragmentation tax.”
The survey behind these figures was substantial: a double-blind study of 12,035 knowledge workers and 173 Fortune 1000 executives, fielded in January and February 2026, plus qualitative interviews with 25 Fortune 500 leaders. These are enterprise numbers, but the mechanism scales all the way down to a 12-person firm in Auburn. If you gave everyone a faster tool and changed nothing about how work moves between them, you bought speed and left the return on the table.
The word “tax” is doing real work in that phrase. A tax is a cost you pay automatically, whether or not you notice it, simply for operating the way you do. The fragmentation tax works the same way: adopt AI one seat at a time with nothing connecting the seats, and you pay it by default — no invoice ever arrives, just a company that compounds slower than the tools it bought promised it would.
Why doesn't faster staff mean a faster company?
The math feels like it should work. Ten people each save an hour a day, so the company gains ten hours a day, right? The research says no — and the reason is coordination.
Atlassian found that knowledge workers spend the majority of their time on collaborative work, yet only about 24% of AI adoption is aimed at teams rather than individuals. So the tool speeds up the 20% of work a person does alone and leaves the 80% that involves handoffs, reviews, and shared decisions untouched. Worse, Atlassian's coverage of the efficiency paradox notes that 87% of knowledge workers say they lack the time or capacity to coordinate because everyone is stuck in execution mode — heads down, moving fast, in parallel, and increasingly out of sync.
There's an adoption gap underneath it, too. Roughly 85% of knowledge workers use AI at work, but only about 29% have actually integrated it into their workflows. The rest are using AI beside their work — a faster way to draft an email — rather than inside their work, where it could own a repeatable process. That distinction is the whole game, and it's the same theme we covered in why execution beats strategy: the value isn't in having the capability, it's in wiring it into how work actually gets done.
Picture a 15-person insurance agency in Fort Wayne. A producer uses AI to draft a quote in ten minutes instead of thirty. That quote lands in the service coordinator's inbox with no notes on why certain coverages were chosen, so the coordinator spends twenty minutes reconstructing the reasoning before it can go to the client. The producer got faster; the agency didn't. The half-hour “saved” upstream reappeared — with interest — as rework downstream. Multiply that one broken handoff across every quote, every claim, and every renewal, and you can see exactly where the fragmentation tax hides: not in the drafting, but in the seams between people.
What do the teams that cracked it do differently?
Not everyone is paying the fragmentation tax. Atlassian identifies a group — roughly 14% of teams — that has translated individual AI speed into organizational return. According to diginomica's analysis of the research, these teams cut the fragmentation tax nearly in half, and the performance spread between them and everyone else is widening.
The gap is not subtle:
| Outcome | Top-performing teams vs. the rest |
|---|---|
| AI improves planning and prioritization | 5.6x more likely to report |
| AI enhances collaboration | 9.4x more likely to say so |
| Fully trust AI to surface relevant information | 2.3x more likely |
“Shared context” sounds abstract until you make it concrete. It means the AI working a lead can see the same CRM history, past quotes, and account notes the whole team sees — so its output arrives ready to use, not ready to re-check. It means one connected workflow instead of five people prompting five private chatbots about overlapping work. The top teams didn't buy smarter models; they gave ordinary models the same situational awareness a good employee builds over months on the job, and then let the tool act on it across the whole process.
What separates them isn't more licenses. It's shared context, connected workflows, and a culture that treats AI as a team capability instead of a personal shortcut. Accounting Times' reporting on the study ties the gap partly to training and integration — organizations that invest in wiring AI into how the team operates get the return; those that just hand out seats don't. It maps closely to what we've written about treating AI as an operating layer rather than a scattered set of tools, and about rethinking org design for agentic AI so the structure supports the technology.
A Fort Wayne diagnostic: is your AI stuck at the individual level?
Here's a plain-language test any Northeast Indiana owner can run this week. If you answer “yes” to most of these, your AI spend is trapped at the individual level and you're paying a local version of the fragmentation tax.
- The demo test: Can anyone in the company point to a workflow — not a person — that AI now runs end to end? Or can they only point to individuals who “use AI”?
- The handoff test: When your AI-drafted proposal, report, or quote leaves the person who made it, does the next person have to redo half of it because they lack the context?
- The duplication test: Are two people using AI to produce overlapping versions of the same thing because nothing connects their work?
- The P&L test: Can you tie any AI usage to a line on the income statement — faster cash collection, more qualified leads worked, shorter turnaround — rather than a vague “we're more productive”?
- The trust test: Would you let AI output go straight to a customer, or does everything still funnel through one person's inbox as a bottleneck?
If the honest answers are “individuals, yes-they-redo-it, yes-duplication, no-P&L-link, and everything-bottlenecks,” you don't have an AI problem. You have a fragmentation problem — and buying more ChatGPT seats will make it slightly worse, not better. The way to check whether the fix is working later is the same discipline we lay out in the metrics that actually matter: measure throughput at the process level, not keystrokes saved by one person.
How does an AI Employee fix the org-level ROI gap?
The reason a scattered pile of AI seats doesn't move the company is that each seat speeds one person's task. To move the company, you have to speed a process — and a process crosses people, tools, and handoffs. That's what an AI Employee is built to do.
Instead of making one salesperson's drafting faster, a Cloud Radix AI Employee owns a workflow end to end: lead intake → qualification → follow-up, or document review, or recurring reporting — running behind a Secure AI Gateway with the context, permissions, and audit trail to operate across the whole pipeline, not one desk. The throughput gain lands on the organization's P&L because the AI holds the shared context that used to leak out at every handoff. There's no re-explaining, no duplicated draft, no bottleneck inbox.
Three or four moves lift most Fort Wayne businesses from individual speed to organizational return:
- Pick one workflow, not one person. Choose a repeatable, cross-person process — the one where handoffs cause the most rework — and make that the AI Employee's job.
- Give it the shared context. The reason the top 14% win is context. Wire the AI into the systems and information the whole process touches, so it doesn't start from a blank slate at each step.
- Measure at the process level. Track the metric the workflow exists to move — days to follow up a lead, hours to turn around a report — not “time saved” by an individual.
- Keep governance in place. A gateway, documented behavior, and human checkpoints so the speed is safe, not just fast.
None of this requires ripping out the tools your team already likes. The producer keeps drafting quotes with ChatGPT; the salesperson keeps their Copilot. The difference is that the workflow around those drafts — the context that travels with them, the handoff to the next person, the follow-up that used to fall through — now runs on rails instead of tribal memory. That's the whole shift from a faster person to a faster company, and a local business can make it deliberately rather than drifting into it seat by seat.
Before committing, it's worth running the AI value audit to put a dollar figure on a specific workflow — so you're targeting organizational ROI from the start, not hoping individual speed adds up on its own.
The Fort Wayne bottom line
Northeast Indiana runs on lean teams. In professional services, healthcare, manufacturing, and home services around Auburn, Fort Wayne, DeKalb County, and Allen County, most companies don't have a dozen people to throw at a problem — which is exactly why fragmentation hurts more here. When a big enterprise pays the fragmentation tax, it's a rounding error. When a 15-person Fort Wayne firm pays it, it's the difference between AI being a real advantage and being an expensive habit.
The research is a gift precisely because it reframes the problem. You don't need your people to be faster — the data says they already are. You need the speed to reach the company. That's an operating-model fix, not a software-license fix, and it's one a local business can make deliberately instead of drifting into more seats and more chaos. If you want to see where your AI spend is stuck and what an org-level fix would look like, talk to Cloud Radix about an AI Employee for your Fort Wayne business — one built to own a workflow, not just speed up a person.
Frequently Asked Questions
Q1.What is the AI fragmentation tax?
The AI fragmentation tax is a term from Atlassian's State of Teams 2026 research for the value lost when individuals adopt AI independently and their speed gains don't connect into organizational results. Atlassian estimates it at roughly $161 billion a year across large enterprises. It shows up as duplicated work, rework at handoffs, and misalignment — the chaos of everyone moving fast in parallel without shared context.
Q2.Why does AI make employees faster but not the whole company?
Because AI is usually adopted one person at a time and speeds the work someone does alone, while most business value comes from collaborative work involving handoffs and shared decisions. Atlassian found only about 24% of AI adoption targets teams, and 87% of workers say they lack time to coordinate. So individual speed rises while the connective tissue of the organization stays the same — or gets noisier.
Q3.How much AI ROI do companies actually see?
In Atlassian's research, 89% of executives said AI increased the speed of work, but only about 6% were confident they could point to clear AI ROI across their organization. The roughly 14% of teams that have cracked it are far more likely to report AI improving planning (5.6x) and collaboration (9.4x). The gap is driven by whether AI is wired into workflows, not by how many people have access to it.
Q4.How can a Fort Wayne business tell if its AI is stuck at the individual level?
Run a quick diagnostic: Can you point to a workflow AI runs end to end, or only to individuals who "use AI"? Does AI-produced work get redone at handoffs? Are people duplicating each other's AI output? Can you tie any AI use to a line on your P&L? If AI only lives in individual tasks and can't be tied to a business outcome, it's stuck at the individual level.
Q5.How does an AI Employee deliver organizational ROI instead of individual speed?
An AI Employee owns a full workflow — such as lead intake, qualification, and follow-up — rather than speeding one person's task. Running behind a Secure AI Gateway with shared context, permissions, and an audit trail, it holds the information that normally leaks out at each handoff, so the throughput gain lands on the company's results rather than one person's calendar. You measure it at the process level, like days-to-follow-up, not keystrokes saved.
Q6.Should we just buy more AI seats for our team?
Usually not, at least not first. The research suggests more individual seats without connected workflows tends to increase fragmentation rather than reduce it. The higher-return move is to pick one cross-person process where handoffs cause the most rework, wire AI into that workflow with shared context and governance, and measure the business metric it's meant to move.
Sources & Further Reading
- Atlassian: atlassian.com/blog/state-of-teams-2026 — The State of Teams 2026.
- Atlassian: atlassian.com/blog/ai-at-work/why-ai-speed-isnt-delivering-roi-for-cios — Why individual AI speed isn't delivering the ROI CIOs expected.
- Atlassian: atlassian.com/blog/ai-at-work/ai-efficiency-paradox — The AI efficiency paradox: What to do when AI boosts productivity but not results.
- diginomica: diginomica.com/ai-strategy-what-best-teams-do-differently — AI strategy: what the best teams do differently.
- Accounting Times: accountingtimes.com.au/technology/lack-of-ai-training-leads-to-fragmentation-tax-atlassian — Lack of AI training leads to fragmentation tax: Atlassian.
- VentureBeat: venturebeat.com/orchestration/atlassian-why-ai-speeds-up-employees-but-not-organizations — Atlassian: why AI speeds up employees but not organizations.
Turn Faster People Into a Faster Company
We'll show you where your AI spend is stuck at the individual level and what an org-level fix looks like — an AI Employee that owns a whole Fort Wayne workflow, not just one person's calendar.
Schedule a Free ConsultationNo contracts. No pressure. Just an honest conversation about what would help your business.
