AI is here.
Now what?
AI makes every individual faster.
Not your organization.
You pay for speed, and your structure burns it in approvals, reviews and meetings. Your competitors have the same tools; the edge is decided by the organization behind them.
Run the system checkThanks to AI, you deliver by the minute. But everything else in your organization, the reviews, alignments, compliance checks, operations, is just as slow as it ever was, and the new load makes it slower still. Everyone is annoyed, and yet hardly anything reaches the customer. And the mood tips into “Thank God it’s Friday”.
That is exactly what we turn around. “Thank God it’s Friday” becomes “Finally, Monday!” again.
“We give our best, but …”
This is not an opinion.
This is measured.
faster: that is how much quicker an individual completes a clearly scoped task with AI, depending on the task type: +14% in customer support, +25% in consulting, and writing tasks take 40% less time. The oft-quoted +55.8% comes from GitHub’s own lab experiment. The gain is real and has been independently replicated several times. QJE 2025 (n=5,179) · Science 2023 (n=453) · Harvard/BCG 2023 (n=758) · GitHub/MIT 2023 (n=95) 1, 2, 3, 4
more time is what experienced developers with AI needed as soon as the task was no longer isolated but sat inside their own, grown codebase (randomized, 246 real tasks): the time moves from generating to checking and fitting the AI output. They still felt 20% faster. This is exactly the gap organizations pay for: perceived speed, measured congestion. METR 2025, randomized controlled trial, arXiv:2507.09089 5
longer is how much reviews now take, while at the same time 98% more pull requests are created: 154% bigger change sets, 9% more bugs, and no measurable improvement at company level. DORA measures the same pattern: more AI, 7.2% less delivery stability. The bottleneck relocates, from the work into the alignment. Faros AI 2025, telemetry of 10,000+ developers · DORA/Google 2024 · Stack Overflow 2025 6, 7, 8
of companies see no measurable effect of generative AI on their bottom line. The rework is now quantified: 40% of knowledge workers receive “workslop”, AI output that looks like work, with almost 2 hours of rework per incident on average. McKinsey “The State of AI” 2025 · BetterUp Labs × Stanford, Harvard Business Review 2025 9, 10
The mechanism behind this has been described for decades: communication paths grow quadratically with the number of people involved, n(n−1)/211, and organizations systematically underestimate integration work12. AI multiplies output per head, not the system’s integration capacity. The measurements come from software teams; the mechanism applies wherever output has to be checked, aligned and fitted in: in a credit decision as in a proposal, in an expert report as in a campaign.
AI does not liberate. AI does not oppress.
It amplifies the system you plug it into.
No coincidence, but a known pattern: Kentaro Toyama’s Law of Amplification (Geek Heresy, 2015)13 shows that technology amplifies an organization’s existing forces instead of replacing them. Put bluntly: AI makes fast organizations even faster and slow ones even slower. And it is measurable: in the 2025 German Socio-Economic Panel (SOEP), the positive link between AI use and perceived autonomy disappears once existing working conditions are controlled for14: the system dominates the tool. In a command structure, AI amplifies the congestion; in a decentralized organization, the value creation. That is why tool rollouts and control towers fail alike: both amplify the wrong system.
Not another tool. An organization that carries speed.
Today, every decision travels up the pyramid. Tomorrow, small autonomous teams decide directly at the market. You do this rebuild yourselves: everyone invited, no one conscripted. Only one person has to truly want it: someone at the top with real authority who opens the space instead of ordering the change. We bring the experience from such rebuilds, you bring what matters most: the knowledge of your business.
Because the experts sit in silos, every idea travels through half the hierarchy: over a hundred alignments, and in the end nobody has decided.
Alignment stays inside the team, then the idea goes straight to the market: a few steps, and the value is delivered.
- Cells
instead of silos - Mini-enterprises with their own results responsibility: five to nine people, functionally complete, close to the customer. Not every team faces the market: internal teams serve the customer-facing ones instead of sitting above them.
- Seams
instead of interfaces - Teams connect through agreements. Day-to-day business runs without escalation ladders and without a boss as traffic light.
- Mastery
instead of cascades - Authority follows the situation, not the position. Reviews become spot checks instead of bottlenecks.
- Guardrails
instead of control loops - Governance is built into the flow of work, not bolted on afterwards. Speed and safety stop being enemies.
Beta is not an experiment.
Beta is proven.
The idea is older than any AI tool: Mary Parker Follett described leadership as power-with instead of power-over as early as the 1920s, with authority that follows the situation instead of the position15. Douglas McGregor showed in 1960 that the image of humankind behind a control system produces exactly the behavior it assumes16. And W. Edwards Deming estimated after decades of process work: 94% of performance belongs to the system, not the person17. Beta is the application of a hundred years of organizational research.
is how long Handelsbanken has steered its branches decentrally, entirely without a classic budget: decisions are made locally, performance is measured relative to competitors instead of against plan figures. More profitable and more cost-efficient than the industry average for decades, two financial crises without state aid. Wallander, “Budgeting: An Unnecessary Evil”, Scandinavian Journal of Management 1999 · Bogsnes 2023 18, 19
is what Buurtzorg needs, having grown from four nurses to over ten thousand in self-steering teams without middle management. The independent review: the country’s best patient experiences, case-mix-adjusted costs in the 38th percentile, despite higher hourly wages. Gray, Sarnak & Burgers, The Commonwealth Fund 2015 · KPMG cost analysis 2015 20
of the US market for tomato paste and diced tomatoes is handled by Morning Star: entirely without supervisors. Colleagues record their commitments in mutual agreements (CLOUs) instead of having them handed down from above. Hamel, “First, Let’s Fire All the Managers”, Harvard Business Review, December 2011 21
These organizations were fast long before AI existed. Because a good system amplifies any input: clear responsibility, short paths and open numbers are the context in which new tools, AI or not, get checked, fitted and shipped right away, by the team that gets to decide. Exactly this kind of system turns potential into speed instead of letting it seep away in congestion.
What does such an organization look like from the inside? We have modeled the pioneers in BetaOS: Buurtzorg as a living structural picture, and with Manic Media a media company from our own practice.
Three steps. One path.
No mammoth project: a scoped entry, a determined shift, a system that stays.
Understand your system.
Your AI makes individuals faster, yet the speed seeps away in the system. We show exactly where, and enable your people to see it for themselves going forward.
A system check without maturity theater
From Alpha to Beta.
Carve small teams with their own results responsibility, agree collaboration directly instead of escalating, measure against the market instead of plan figures: with the whole system, radically inviting, timeboxed. No pilot-group cosmetics.
The core program
The system that stays.
Once your system stands, it gets its operating system: BetaOS keeps cells, seams and value creation permanently visible, the living picture of your organization instead of org chart and budget binder. Not a control tower: BetaOS shows the structure, it doesn’t direct it, and the numbers belong to the teams, not to a head office. Because the shift from Alpha to Beta applies to your software, too.
Your Beta operating system
Good questions. Beta answers.
These questions are not invented: this is how we have been hearing them in our first conversations for years, from the board to the works council. And this is how we answer them.
“Shouldn’t we finish the AI rollout first?”
The other way round: every month of AI rollout in the old system feeds the congestion you see in the findings above. The organization is the lever that makes your AI spending pay off: the earlier you pull it, the more of it reaches the customer.
“We’re already rolling out AI, with licenses and training.”
Exactly those rollouts are what the findings above measure: the individual gets faster, the organization does not. Licenses and training raise output per head, but not the system’s capacity to check, fit and ship it. Without the rebuild, you buy speed and pay for it in congestion.
“For us, AI is an IT matter.”
But the bottleneck is not in the technology. What eats the AI gain is approvals, alignment and committees, in other words: organization. IT can introduce a tool; whether it turns into speed at the customer is decided by the structure everyone else works in.
“AI will soon replace so many roles that the rebuild is pointless.”
The opposite: the more output per head, the more depends on the system that integrates it. Communication paths grow quadratically, not headcount. A smaller, AI-amplified workforce in the old pyramid just jams up faster.
“This only works in IT.”
The numbers in the findings come from software development because every step there leaves telemetry: nowhere else is the congestion documented this well. The mechanism is industry-blind: wherever output has to be checked, aligned and fitted in, the bottleneck moves into the alignment. Software is just the industry with the best measuring device.
“What about day-to-day business?”
The rebuild is not a move. The design emerges in a few workshop days with everyone, the flip is a fixed date, followed by 90 days of settling in. Delivery continues throughout: in our mandates, day-to-day business kept running on every single day of the rebuild.
“Couldn’t we start with a pilot group?”
Tempting, but proven ineffective: the island gets reshaped by the surrounding system as long as budgets, targets and reporting lines stay the old ones. Hence the opposite path: the whole system, but timeboxed. What is small in the rebuild is not the group, it is the time frame.
“We’re regulated. Reviews are mandatory.”
And they stay that way. Beta does not abolish compliance, it builds it into the flow of work: four-eyes checks where the regulator requires them, instead of a committee loop around every decision. Handelsbanken has shown for fifty years that this works in the most regulated industry of all.
“What will the works council say?”
Beta takes away no one’s income or security: no individual performance ratings, no bonus arbitrariness, fixed salaries stay. And the rebuild is an invitation, not a decree: employee representation is at the table from day one.
“Can the works council block this?”
It rightly has a say in much of this, from work organization to pay principles. But the real question is a different one: we do not rebuild around the works council, we rebuild with it, from week one. Beta abolishes what works councils have been fighting for decades: individual performance ratings, bonus arbitrariness, hidden monitoring. Whoever wants to block that would have to argue for the pyramid.
“Our leaders will never go along with this.”
Leadership work does not disappear, it becomes more demanding: strategy, the playing field, mastery, instead of approval loops. Nobody loses pay or face: those who steer today are needed tomorrow where value is created. And in our experience, the biggest skeptics often become the strongest drivers.
“Our people don’t even want responsibility.”
After years of command and control it looks that way: working to rule is learned behavior, not a character trait. The image of humankind behind a control system produces exactly the behavior it assumes. People do not refuse responsibility, they refuse responsibility without the power to decide.
“And what becomes of HR?”
HR turns from a steering body into a service provider for the cells: contracts, payroll, labor law and good procedures remain, as an offer the teams pull. What HR hands over are appraisals, bonus rounds and top-down staffing: recruiting happens in the team, development happens on the job. Less administration of people, more craft for people.
“Who is accountable in the end?”
Accountability is not abolished, it becomes visible. Every cell has its own profit and loss statement, decisions are made where the knowledge sits, and all numbers are open. Today, accountability often sits where the knowledge is not: at the top.
“Without bosses there will be chaos.”
Beta is not less structure, it is a different structure: named cells, written agreements between teams, open numbers. That is more visible and more binding than any org chart, through which the real work never flowed anyway.
“Without targets, nobody knows where to go.”
Direction comes from strategy and the market, not from a planned figure: performance is measured actuals against actuals, the team compared to the market and the previous period. That is tougher than any budget, because you cannot negotiate with the market.
“Without bonuses, nobody will push themselves.”
Research has said the opposite for decades: individual incentives crowd out exactly the motivation they claim to buy, and reward hitting targets rather than creating value. Beta pays fair fixed salaries plus profit sharing for everyone: the pie grows together instead of being defended piece by piece.
“We’re too big for this. Or too small.”
Cell structure scales by division, not by layers: it works from the 60-person unit to the group of tens of thousands. Haier runs on about 4,000 micro-enterprises, Buurtzorg grew from 4 to 10,000 nurses without middle management. What matters is not your size, it is the dynamics of your market.
“Nice for startups. We run a factory.”
The pioneers are not tech hipsters: a foundry (FAVI), a tomato processor (Morning Star), a nursing service (Buurtzorg), a bank (Handelsbanken). Beta comes from production, not from Silicon Valley.
“This is just agile with a new label.”
No. Agile methods optimize teams inside the pyramid, and that is exactly where they get stuck: the review and stability numbers in the findings above measure agile organizations. Beta changes the operating system: who decides, how performance is measured, where the money flows.
“Zappos tried this. It went wrong.”
Zappos failed with Holacracy: a rule constitution with its meeting liturgy. Beta is the opposite: few principles, market pull, agreements between teams. What failed there was not decentralization, it was its bureaucratization.
“Peer pressure just replaces boss pressure.”
The most honest objection on this list. Yes, small teams with open numbers create social pull, that is part of the effect. The difference: the load is visible and negotiable, in written agreements and open capacities, instead of quietly landing in one person’s inbox.
“What about my people’s careers?”
In Beta, advancement does not mean climbing, it means mastery: more responsibility, bigger topics, growing pay, without having to become a boss for it. Those who want to develop people keep doing so, as a role rather than a position.
“And without titles? What do I show when I change jobs?”
Responsibility you can prove. Anyone who has carried a cell with its own profit and loss statement talks about customers, numbers and decisions in the next interview, instead of reporting lines. A title says where someone sat. A P&L shows what someone can do.
“Who hires, who lets people go?”
The people who will work with them afterwards: teams recruit their colleagues themselves, with a right of veto. And separations follow clear, fair procedures instead of a superior’s mood. Both become more careful, not more casual.
“Our customers want fixed points of contact.”
They get them: a cell that stays. Except their contact is now allowed to decide instead of forwarding. Customer relationships become more stable, because teams are built to last, not until the next reorganization.
“And who coordinates the big cross-team project?”
Cells form coalitions for that, with their own agreement and a shared result, for the duration of the endeavor. That replaces the project overhead: coordination emerges from the work, not above it.
“And if a cell performs poorly?”
Then everyone sees it, early: every cell has an open profit and loss statement. First comes help from other cells, then a clear deadline, as a last resort dissolution. That is corrected faster than in any line organization, where losses hibernate in budgets for years.
“We’re bound by group-level policies.”
The shell stays: reporting lines, legal and group compliance keep running through the formal structure, that is what it is for. Inside that shell, you organize value creation as a cell structure: the pyramid stands on the outside, the market governs on the inside.
“And if top management changes?”
The most common cause of death of “liberated” companies is the departure of their liberator. That is why Beta anchors the change in the system, not in the person: cells, agreements and open numbers survive any change at the top. Charisma leaves, structure stays.
“And if it doesn’t work for us?”
Beta is no silver bullet. It needs three things: a sponsor with real authority who genuinely wants it; a business whose value creation is complex rather than purely regulated and repetitive; and the courage to shift the whole system instead of a pilot group. If one is missing, we say so openly and advise against it. If they’re there, we know of no real flip that hasn’t held. And you don’t have to take our word for it: timeboxed, the numbers open, and after 90 days everyone in the company can measure it.
We have been rethinking organizations since 2015.
Long before it became urgent.
“With Scalamento, we are preparing Sparkasse HRV for working in complexity. We have become more customer-centered, and our culture has become even more attractive to new employees.”
“Scalamento has successfully supported us in transforming from a publishing and printing house to a modern media group, significantly aiding in the organizational development for our Digital First strategy. Thank you!”
“During my time at PwC, together with Stefan and the Scalers, I found my way into the world of agility and into organizational design at large. Stefan was and is a sparring partner for me, one who keeps opening up new perspectives and demonstrating the necessity of modern organizational design.”
“No consultancy has ever brought us as far forward as Scalamento.”
“In our collaboration, Scalamento played a significant role in the creation and scaling of MOIA. They helped us build a resilient and scalable organization. Hamburg, as an attractive location for the start-up scene, needs more innovative companies like Scalamento.”
“Working with Scalamento has initiated a true cultural change for 58 Agents. The introduction of new ways of working has brought our teams closer together and strengthened our innovation power.”
“Scalamento has paved the way for genuine autonomy in our teams. With their support, we were able to establish self-organized product teams through a structured process, significantly enhancing efficiency in collaboration.”
“Scalamento accompanied us in defining our True North and finding ways to work more closely together as a team, regardless of location. Personally, I have felt understood and acknowledged by my employer ever since.”
You rebuild.
We bring the cadence.
Beta cannot be delivered to your door: you do the rebuild yourselves, with the whole system. We bring method, tooling and the cadence of the rebuild, timeboxed instead of endless. After that, the cadence of your business is set by the market, not by us. Four people, no consulting apparatus, close to the customer.
Dr.-Ing. Stefan Link
Organizational developer, startup founder, tech nerd
More
I built my first startup in 2003 as a decentralized network organization with self-organized teams. Since then, as CTO and COO at companies including Bauer Media, Mondia Media and Finanzcheck, I have experienced how the same principles work in large enterprises. Today I accompany organizations from startup to corporation through the system shift to Beta.
Ute Athen
Organizational developer, coach, cultural scientist
More
As a cultural scientist, I am fascinated by the power that emerges when people are not constrained by rigid processes, budgets and departmental thinking. At and for Bertelsmann, Vodafone and O2, I explored what self-organized teams can unfold. Today I bring this experience into mandates for small and large companies.
Katie Lucke
Organizational developer, entrepreneur, team player
More
As an entrepreneur in my own butchery business and as a transformation coach at Osteria and REWE, I have experienced the power companies develop when they bet on trust and on people’s potential. The foundation for this is a positive image of humankind. In coaching, I recognize the big picture with a systemic view and make it discussable in a trusting way.
Anja Link
Lover of numbers, keeper of everyone’s back
More
As a certified accountant, I hold numbers, data and facts close to my heart: I keep everyone else’s back free of all things administrative, so that “Finally, Monday!” applies internally too. Having grown up in classic Alpha finance systems, I am fascinated by the freedom that value creation accounting unleashes for us and our clients. Beta principles are arriving ever faster, even in the conservative world of finance.
The knowledge has long been there.
We put it to work.
For more than ten years we have been mining the sources of good organization and putting them to work in real rebuilds, not just on the shelf. The best of it, curated: 42 books, whitepapers, podcasts and videos on the system shift.
Nothing in this combination. We keep reading.
Finally, Monday!
A contemporary organizational model puts your horsepower on the road: speed reaches the customer, and on Mondays everyone is glad to be back. The first step is a system check.
Every number has a source.
The evidence above is weighted by strength: experiments and telemetry before surveys, primary sources before retellings.
- Brynjolfsson, Li & Raymond (2025): Generative AI at Work. The Quarterly Journal of Economics 140(2). n=5,179.
- Noy & Zhang (2023): Experimental Evidence on the Productivity Effects of Generative AI. Science 381. n=453.
- Dell’Acqua et al. (2023): Navigating the Jagged Technological Frontier. HBS Working Paper 24-013, with BCG. n=758.
- Peng et al. (2023): The Impact of AI on Developer Productivity. arXiv:2302.06590. Vendor lab RCT, n=95.
- Becker, Rush, Barnes & Rein (2025): Measuring the Impact of Early-2025 AI on Experienced Developer Productivity. METR, arXiv:2507.09089.
- Faros AI (2025): The AI Productivity Paradox. Telemetry of 10,000+ developers across 1,255 teams.
- DORA / Google Cloud (2024): Accelerate State of DevOps Report.
- Stack Overflow (2025): Developer Survey. Over 49,000 responses.
- McKinsey (2025): The State of AI. Global survey.
- BetterUp Labs × Stanford Social Media Lab (2025): AI-Generated “Workslop” Is Destroying Productivity. Harvard Business Review. n=1,150.
- Brooks (1975): The Mythical Man-Month. Addison-Wesley. Communication paths n(n−1)/2.
- Heath & Staudenmayer (2000): Coordination Neglect. Research in Organizational Behavior 22.
- Toyama (2015): Geek Heresy. PublicAffairs. Law of Amplification.
- de Morree (2026): AI and the Future of Work: Liberation vs. Control. Corporate Rebels, with SOEP 2025 panel analysis.
- Follett (1942, posthumous): Dynamic Administration. Harper. Lectures from the 1920s.
- McGregor (1960): The Human Side of Enterprise. McGraw-Hill.
- Deming (1994): The New Economics for Industry, Government, Education. MIT Press.
- Wallander (1999): Budgeting: An Unnecessary Evil. Scandinavian Journal of Management 15(4).
- Bogsnes (2023): This Is Beyond Budgeting. Wiley.
- Gray, Sarnak & Burgers (2015): The Netherlands’ Buurtzorg Model. The Commonwealth Fund, with 2015 KPMG cost analysis.
- Hamel (2011): First, Let’s Fire All the Managers. Harvard Business Review 89(12).