The Seat-Based Software Trap: How to Get More Output Without More Headcount

Every growing company runs the same experiment without meaning to. Work piles up, so you hire. The new person needs a login to the CRM, the ERP, the ticketing tool, the document store and the analytics dashboard, so you buy five more seats. The bill goes up in a straight line with your headcount. And a quarter later, the backlog of routine work is exactly as deep as before.

The reason is a mismatch nobody prices in. Per-seat software charges you for the number of people who can open it. What you are actually short of is capacity to finish the work sitting in the queue. Those are two different things, and buying more of the first does nothing for the second. Zylo’s 2025 SaaS Management Index puts average software spend at 4,830 dollars per employee, up almost 22 percent in a year, with the average company using just 54 percent of the licenses it pays for⁴⁵.

This piece is for the operations leader, CFO or managing director who keeps approving seat renewals and keeps wondering why the team never gets ahead. It names the trap, shows the maths, explains why the pricing model is quietly collapsing, and lays out the alternative: buying completed work instead of logins, so output grows without the headcount growing with it.

The Seat Math Problem

Per-seat pricing feels rational because it is simple. You pay a fixed fee for each named user, so the cost is predictable and easy to budget. The problem starts the moment you notice what the fee is attached to: the number of people, not the amount of work those people finish.

• Cost tracks headcount - Every hire multiplies across your whole software estate. One new coordinator can trigger new seats in the CRM, the ERP, the helpdesk, the BI tool and the collaboration suite, so a single salary quietly adds thousands in annual license fees on top⁴.
• Usage and cost rarely match - A license bought for a role that opens the tool twice a month costs the same as one used eight hours a day. The vendor charges for the door, not for how often you walk through it¹.
• Nearly half of every seat is waste - Zylo found the average organisation actively uses only 54 percent of its SaaS licenses. The other 46 percent are paid for and idle⁵.
• The waste compounds - Organisations lose an average of 21 million dollars a year on licenses nobody uses, a figure that rose more than 14 percent year over year as tool counts crept up⁴.
• Growth becomes a tax - The vendor’s revenue rises with your headcount, which is exactly the variable a scaling company least wants to make more expensive. You are effectively taxed for hiring¹.
• The backlog is untouched - None of that spend does the routine work. The invoices still need matching, the orders still need entering, the tickets still need triaging. More seats mean more people who can do it manually, not less of it to do.

The clearest way to see the trap is to separate two words that per-seat pricing treats as one: headcount and capacity.

A worked example: three hires against one backlog

Put rough numbers on it. Say a growing operations team faces a rising backlog and adds three coordinators over a year, each touching five per-seat tools. Watch what the two lines do.

• The cost line moves in a straight, permanent climb - three salaries and fifteen recurring licenses that renew every year whether the work shrinks or not.
• The capacity line barely twitches - the new people inherit the same manual process and lose the same half-week to email and information-hunting that everyone else does⁶.
• The backlog wins the race - because volume rises with growth, the queue is often deeper at year end than when you started hiring.
• The alternative changes the shape of the cost line - take the routine work off people and cost tracks work completed, so it flattens or falls as the backlog clears instead of climbing with every hire.

Once you see headcount and capacity as separate lines, every per-seat renewal reads differently. You are approving spend on the first line while your actual shortage is on the second.

What a Seat Actually Buys You (and What It Does Not)

A software seat is a right of access. It lets a named person log in and operate the tool. That is genuinely useful, but it is worth being precise about where the value stops, because the gap between access and output is where budgets leak.

A seat gives you access

• Permission to enter - The person can open the system, see the data and use its features. Without a seat, they cannot participate at all.
• A place to do the work - The tool provides the forms, screens and workflows. It is the workshop, not the worker.
• A record of what was done - Seats carry audit trails, permissions and accountability tied to a named identity, which matters for compliance.
• Network value, sometimes - For a communication or collaboration tool, each extra user genuinely adds value, because the point is that more people are connected. Here per-seat pricing is fair.

A seat does not give you the work done

• No hours are added - The seat does not process a single invoice on its own. A human still has to sit in it and do the task by hand.
• No routine load is removed - The queue of repetitive work is exactly as long after you buy the seat as before. Access does not subtract.
• No knowledge is retained - When the person in the seat leaves, their way of doing the work leaves with them. The license stays; the capability walks out the door.
• No leverage is created - The person in the seat is not faster because a fee was paid. Speed comes from tooling that does the work, not from the right to open a screen.

This distinction is not academic. It explains a pattern every operations leader recognises: the software estate keeps growing, the license bill keeps climbing, and the team still spends its days on manual routine.

Why More Seats Never Clear the Backlog

The routine-work backlog is not a fixed pile you can staff your way out of. It refills. Every order, invoice, ticket, form and status request that arrives adds to it, and it grows with the business. Adding seats adds hands to shovel it, but the shovelling is the problem, not the shortage of shovels.

The backlog grows faster than seats can keep up

• Volume rises with growth - More customers, more suppliers and more transactions mean more routine work, so the queue deepens exactly when you can least afford to add people.
• Knowledge workers already lose the day to it - McKinsey found the average interaction worker spends about 28 percent of the week on email and nearly 20 percent hunting for internal information or the colleague who has it⁶. Almost half the week goes before real work starts.
• Every new seat inherits the same manual process - A new hire in a new seat does the work the same slow way. You have multiplied the labour, not changed the method.
• Coordination overhead grows too - More people means more handoffs, more status updates and more meetings to align them. Some of the new capacity is eaten by the cost of having more people.
• The labour to fill the seats is not there - The German Mittelstand cannot simply hire its way out. The DIHK reports a persistent skilled-worker shortage, and the OECD projects Germany’s working-age population will shrink by 3.9 million by 2030¹⁸¹⁹.

What actually moves the backlog

Only two levers shorten a routine-work queue: reduce how much arrives, or finish each item with less human effort. Per-seat tools do neither. The lever that works is taking the routine item off the person entirely.

This is the heart of the trap. The tool sold to help with the work is priced by the number of people still doing the work, so the vendor has no reason to remove it and every reason to keep you adding seats.

The Pricing Shift Nobody Priced In

The seat-based model is not just inefficient for buyers. It is breaking for vendors too, and the market has started to move. If you renew per-seat contracts on autopilot, you are buying into a model the industry is actively leaving.

The market is moving off seats

• Pure per-seat is shrinking fast - Per-seat as the primary model dropped from 21 percent to 15 percent of companies in twelve months, while hybrid models jumped from roughly 27 percent to 41 percent².
• Gartner puts a number on the risk - Up to 234 billion dollars of enterprise application spending is exposed to agentic AI through 2030, about 20 percent of enterprise SaaS spend³¹⁰.
• The value link is broken - When one AI-augmented user does the work of ten, charging per user makes no sense. Buyers refuse to pay per seat for software that reduces the number of seats they need⁸.
• Outcome pricing is rising - Analysts expect a long shift from subscription toward usage and outcome models, with vendors charging per resolved ticket, per processed document or per qualified lead¹⁶¹⁷.
• The penalty for standing still is real - Vendors clinging to pure per-seat pricing for AI products are reported to see markedly lower gross margins and higher churn than those on usage or outcome models⁸.

Why buyers are the ones forcing the change

This is not vendors being generous. It is buyers refusing a model that punishes them for getting more efficient, and analysts telling software companies to adapt or lose the revenue.

But outcome-based pricing has a catch that most vendors gloss over, and it points directly at what has to change underneath the price tag.

Capacity Without Headcount: The Alternative to Buying Seats

If more seats do not add capacity, the question changes. Instead of asking how many logins the team needs, you ask how much of the routine work you want taken off the team entirely. That is a different purchase: you buy completed work, not access, and the economic unit stops being the person.

Two ingredients turn a tool into capacity

Outcome pricing only works if something actually owns the outcome from start to finish. A rebranded chatbot cannot be paid per resolution because it never completes one on its own. Real capacity needs two things a per-seat tool does not have.

• A Company Brain - A shared memory of how your company actually works: the people-knowledge, processes, rules and data that normally live in individual heads and scattered systems. It survives staff turnover, so the way the work is done stays even when the person who did it leaves.
• AI employees wired into your real systems - Not a copilot waiting in one window, but AI employees connected to the email, Teams, SharePoint, CRM and ERP you already run. They take a routine task end to end - triage the inbox, enter the order, match the invoice, draft the quote - and they learn your company, not the internet, by getting feedback from your team every day.

Why this breaks the seat link

When an AI employee owns a task, the seat leaves the critical path for that work. It runs whether or not a person is logged in, so cost stops tracking headcount and starts tracking work completed.

This is what leverage means in practice: the same team producing more because the routine load has been lifted off it, not because more people were added underneath it.

Nine places the routine load actually sits

The trap is easiest to see in concrete tasks. Each of these is routine work that today occupies a person in a seat and could instead be owned end to end by an AI employee.

1. Inbox triage - A shared mailbox where every message is read, classified and routed by hand. An AI employee sorts, drafts replies and escalates only the exceptions.
2. Order entry - Orders arriving by email and PDF are keyed into the ERP manually. An AI employee reads them, enters them and flags mismatches.
3. Invoice matching - Invoices checked against purchase orders and delivery notes line by line. An AI employee reconciles and posts the clean ones, queuing only discrepancies.
4. Quote preparation - A sales engineer rebuilds a similar quote from scratch each time. An AI employee drafts it from the Company Brain and the CRM, ready for review.
5. Status chasing - Someone pings colleagues for updates on shipments, approvals or tickets. An AI employee gathers the status across systems and prepares the summary.
6. Data entry between systems - The copy-paste economy: moving the same field from one tool to another. An AI employee keeps the systems in sync directly.
7. Document drafting - Routine confirmations, letters and reports written by hand from a template. An AI employee produces the first draft with the right data filled in.
8. Ticket resolution - Repetitive support questions answered one by one. An AI employee resolves the known patterns and hands over the genuinely new.
9. Morning briefings - A manager assembles yesterday’s numbers and today’s priorities manually. An AI employee prepares the briefing overnight so the day starts ahead.

None of these needs a new seat. Each needs the work taken off the person, which is precisely what per-seat software is not built to do.

How to Audit Your Own Seat Spend

Before you decide anything, get the real picture of where your software money goes and how much of it buys access versus output. This is a practical exercise any operations or finance lead can run in a week.

A five-step seat audit

1. List every per-seat contract - Pull the vendor, the number of seats, the annual cost and the renewal date for each tool. Most companies are surprised by the length of this list once it is in one place⁴.
2. Measure real utilisation - For each tool, find how many seats logged in during the last 30 days and how often. Expect the industry average: roughly 54 percent active, the rest idle⁵.
3. Separate access tools from work tools - Mark each tool as access value (communication, authoring, dashboards) or work-clearing value (the tool exists to get routine work done). The second group is where the trap lives.
4. Name the routine work behind each work tool - For every work-clearing tool, write down the actual task people do inside it by hand: entering orders, matching invoices, resolving tickets. That task, not the seat, is the thing to cost.
5. Cost the manual hours - Estimate the hours per week the team spends doing that routine work manually, times the loaded hourly cost. This is your real spend on the backlog, and it usually dwarfs the license fee.

Reading the results

• High license cost, low utilisation - You are paying for access nobody uses. The immediate win is right-sizing seats; the bigger win is questioning whether the tool should occupy a person at all.
• Modest license cost, huge manual hours - The classic trap. The license looks cheap while the real cost hides in the team’s time. This is the strongest candidate for taking the work off people.
• Access tools with high use - Leave these alone. Per-seat is fair here and the value is real.
• Renewals inside 90 days - Do not autopilot them. Each renewal is a decision point to ask whether you are buying capacity or just buying seats again.

How Superkind Fits

Superkind builds AI employees for your company: custom AI that lives inside your systems and gets better every day because your team works with it. The model is deliberately not per-seat. You are not buying more logins - you are handing routine work to AI employees backed by a Company Brain, so output grows without headcount growing with it.

• Company Brain at the core - Your processes, rules and people-knowledge become shared company memory that survives turnover. AI employees work from your reality, not a generic model’s guess.
• AI employees, not copilots - Each one owns a routine task end to end rather than waiting for a person to prompt it in a window. The seat leaves the critical path for that work.
• One layer over what you already run - Connects to email, Teams, SharePoint, CRM, ERP and any API-based software like Salesforce, SAP and HubSpot. No rip-and-replace, nothing new for the team to learn.
• Learns your company, not the internet - Your team works with it from day one and gives feedback, so it sharpens on your part numbers, your approvals and your exceptions.
• Live in about two weeks - First AI employees deploy in roughly two weeks rather than the six-month rollouts common with large software projects.
• Works the off-hours - Because an AI employee is not a seat, it runs at night and on weekends, so the queue is shorter when the team arrives.
• Priced by work, not by login - The unit is routine work taken off the team, which is what you were actually trying to buy when you kept adding seats.
• Your data stays in your infrastructure - Secure, API-based connections with data protection built in, aligned with DSGVO expectations for German companies.
• Leverage, not headcount - The point is more performance from the team you have, cutting manual routine work sharply so people spend their time on judgement.

Superkind is not the answer to every software line item. It is the answer to the one the seat model never solved: the routine work that keeps occupying people no matter how many seats you buy.

Decision Framework: Buy a Seat, Hire, or Add Capacity?

When work is piling up, you have three moves: buy more seats, hire more people, or take the work off people. Here is how to tell which one the situation actually calls for.

The direction is the same whether it comes from a Gartner analyst, a hyperscaler CEO or a Big Four chair: the cost of work is moving off headcount and onto systems that actually do the work. Per-seat pricing is on the wrong side of that shift.

Frequently Asked Questions

Related Articles

• The End of Per-Seat Software: Why AI Employees Are Priced by Outcome, Not by Login
• Hiring Freeze, Not Layoffs: How to Grow Output While Headcount Stays Flat
• The Routine-Work Tax: Why Your Best People Spend Half Their Day Below Their Pay Grade
• New Hire vs AI Agent: How the Mittelstand Decides Whether to Fill the Seat or Deploy an Agent
• The Always-On Colleague: What an AI Employee Gets Done Overnight and on Weekends

Henri Jung

Co-founder of Superkind, where he helps SMEs and enterprises deploy custom AI agents that actually fit how their teams work. Henri is passionate about closing the gap between what AI can do and the value it creates in real companies. Before Superkind, he spent years working with mid-sized businesses on digital transformation and saw first-hand how many companies keep buying seats when what they need is capacity. He believes the Mittelstand has everything it needs to lead in AI - it just needs the right approach.