In this article
If AI is saving your people time every week, where exactly is that time going? Most owners cannot answer that question, and it is the reason AI productivity gains so rarely show up in profit. McKinsey's August 2026 State of AI survey found that 80% of respondents who use AI in their roles say it has improved their individual productivity. Only 37% attribute any positive EBIT impact to AI, essentially unchanged from 2025, and only about 6% qualify as high performers that attribute 5% or more of EBIT to AI.
The direct answer: individual AI productivity gains do not reach the P&L because saved minutes are absorbed by the existing workflow rather than redirected. A faster draft still waits in the same approval queue, and freed hours fill with whatever was already there. Profit moves only when leaders redesign the work and decide where reclaimed capacity goes.
We have written before about the measurement side of this problem, including why AI that feels productive still has to be proven with a baseline. This piece is about the step before measurement. Even a perfectly measured AI program will show a disappointing number if nobody changed how the work flows.
Why Don't AI Productivity Gains Show Up in Profit?
Because a task getting faster and a process getting better are two different events, and most companies only achieve the first. Techaisle's SMB and midmarket GenAI research, based on 3,450 businesses with 1 to 999 employees, makes the split unusually clear: 76% of firms report accelerated task completion, but only 15% say they have successfully translated those faster individual tasks into better automated business processes.
That 61-point gap is where the money disappears. An accounts payable clerk who drafts vendor emails in half the time has not reduced your cost of processing an invoice if the invoice still needs three manual approvals, a rekeyed entry, and a weekly batch run. A sales rep who writes proposals faster has not shortened the sales cycle if pricing sign-off still takes four days. The individual feels the gain. The business sees the same throughput.
Where does saved time actually go?
In our experience, it goes to three places, none of which appear on a financial statement. Some of it becomes slack that absorbs interruptions. Some of it fills with lower-value work that was previously skipped. And some of it gets consumed by AI itself, through reviewing, correcting, and reformatting output. None of that is waste in a moral sense, but none of it is profit either, and it is the default outcome when nobody decides otherwise.
The Belief Leaders Need to Drop: Adoption Is Not the Finish Line
The common assumption is that if enough people use AI, results will follow. The 2026 data says otherwise. McKinsey found that 89% of respondents report regular AI use in at least one business function and 44% say AI is scaling across their enterprise, up from 38% a year earlier, yet the share reporting EBIT impact barely moved. Usage rose. Financial results did not.
What did separate the winners was work redesign. In the same McKinsey survey, nearly three-quarters of AI high performers reported fundamentally redesigning workflows because of AI use, compared with about one-quarter of other respondents. That is the single clearest difference in the report, and it is a leadership decision, not a technology one. Our main-blog piece on what you gain when AI is done right makes the same case at small-business scale.
Spending more does not substitute for it. McKinsey also found that 20% of respondents had constrained AI use because of operating costs, including token costs. Adding seats or upgrading models on top of an unchanged workflow just raises the cost of doing the same work slightly faster.
Who Is Actually Redesigning the Work?
Fewer organizations than you would expect, and the people doing the work are often left out. Adecco Group's May 2026 survey of 2,000 C-suite executives across 13 countries found that only 39% are involving employees directly in redesigning their jobs around AI, and just 22% are highly confident their organization is building the capabilities needed to keep pace.
That matters because the person doing a task knows where the real friction is. They know which approval exists only because of a problem from 2019, which report nobody reads, and which handoff causes the most rework. A redesign done without them tends to automate the visible steps and leave the expensive ones alone.
Kyndryl's research points the same way. Its June 2026 People Readiness Report, based on 1,100 senior business and technology leaders, found that 57% say AI is embedded in core processes or deployed broadly, yet only 23% say their workforce is fully ready and only 11% have achieved both of their top two AI goals. Kyndryl CIO Kim Basile said organizations that invest in people, including rethinking roles and workflows, are seeing positive outcomes at a much higher rate.
What a Redesign Looks Like in a 50-Person Company
You do not need a transformation office to do this. You need to pick one workflow, map it end to end, and make deliberate decisions about what changes. The comparison below shows the difference between layering AI onto a process and redesigning around it.
| Layering AI on top | Redesigning around AI |
|---|---|
| Each person uses AI to speed up their own step | The whole workflow is mapped and redundant steps are removed |
| Approvals and handoffs stay the same | Approval rules change to match the new risk level of each step |
| Saved time is absorbed informally | Leadership decides in advance where reclaimed capacity goes |
| Success is judged by usage and anecdotes | Success is judged by cycle time, error rate, and cost per unit of work |
A practical sequence for the first workflow looks like this:
- Choose a high-volume, measurable process such as invoice handling, client onboarding, or ticket triage, where small time savings repeat many times a month.
- Map it with the people who do it, including every approval, rekey, and wait. The waits are usually where most of the cycle time hides.
- Decide what AI does, what humans check, and what disappears. Some steps should be removed entirely rather than accelerated.
- Assign the freed capacity before you launch. Name whether it goes to backlog, client work, growth projects, or avoided hiring. This is the step that turns time into money.
- Set the baseline and the owner, then review after 60 to 90 days. Our guide to measuring AI ROI without inventing the numbers covers which metrics hold up.
Our main-blog walkthrough of workflow automation use cases in accounting and auditing firms shows what this looks like in practice for professional services, and the end-to-end AI adoption framework places it in the larger sequence.
Does a smaller company have an advantage here?
Often, yes. Redesign requires the authority to change approvals, roles, and handoffs, and in a 40-person firm that authority usually sits with a handful of people who can meet in one room. Large enterprises need months of alignment to change a single approval chain. A smaller company can redesign one workflow in weeks, which is why being early on the adoption curve is not the disadvantage it appears to be.
The constraint for smaller firms is usually attention rather than authority. The same few people who can approve a redesign are also running sales, payroll, and client escalations, so redesign work gets postponed in favor of whatever is urgent this week. The fix is to treat the first workflow as a scoped project with a named owner and a date, not as something the team will get to once things calm down. Things rarely calm down, and the saved time keeps leaking while you wait.
The Risk Side of Redesign
Redesigning work around AI also changes where errors and data exposure can occur. When a step that used to involve a human check is removed or automated, the business needs to know what replaced that check. When AI starts handling client data inside a workflow, the data handling rules for that workflow have to be updated. Our post on why AI adoption is a governance problem, not a technology one explains why these decisions belong in the same conversation as the productivity goal. Done together, redesign and governance reinforce each other. Done separately, one usually undoes the other.
How Securafy Helps Turn Saved Time Into Business Results
When clients ask us why their AI investment is not showing up in results, we start with one workflow rather than the whole AI program. We map it with the people who run it, identify where AI is speeding up individual steps without changing throughput, and work with leadership to decide what gets removed, what gets automated, and where the reclaimed hours will go. Because we also manage the underlying Microsoft 365, identity, and security environment for many clients, we can make the technical changes the redesign requires, such as permissions, data handling, and integrations, without a separate project.
From there we help set the baseline, name the owner, and schedule the review, so the first redesigned workflow produces a number leadership can defend and a template the next workflow can follow. Teams that want a structured starting point can begin with an AI readiness assessment to identify which workflows are most worth redesigning first.
Where To Go From Here
If your people say AI is helping but your financials do not, the gap is almost certainly in how the work is designed, not in the tools. Pick one workflow, decide where the saved time should go, and redesign around that decision.
If your team is moving faster with AI than your guardrails are, start with structured training rather than another tool. AI University gives your people role-based AI training with security built into the material, not bolted on afterward.
If you would rather talk through your specific environment first, book a strategy call with Securafy and we will walk your current AI usage, exposure, and the fastest path to safe adoption.
Not sure where you stand? Take the AI Readiness Assessment before you commit budget to tools.
Take the assessment
Randy Hall is the CEO and Founder of Securafy, with decades of experience helping organizations make smarter, safer decisions about technology.
A frequent speaker and instructor at national IT events, Randy has advised thousands of organizations, from startups and SMBs to large enterprises and U.S. government entities, on secure, practical technology adoption. He writes about the decisions business leaders are often expected to make without enough context, including cybersecurity, compliance, AI, cyber insurance, IT strategy, and business resilience.
Outside the office, you’ll often find Randy on Lake Erie enjoying time on his 38-foot Chris-Craft.
Writes about: Cybersecurity strategy, compliance, AI security, business resilience, cyber insurance, SMB risk, IT leadership
Join the conversation
Have a question or a different take on this? Add it below.