AI slop isn't limited to slide decks and marketing one-pagers. We might be creating an organizational version of it.
Think about that colleague who spends hours assembling a specialized, static data set, feeds it into AI, and produces a polished presentation in seconds. The slides look impressive. The output appears more sophisticated. But the underlying thinking hasn't improved, and nobody has questioned why the process needed to be so complicated in the first place.
Now consider what we're doing with AI across our businesses.
We're automating processes, accelerating workflows, and improving productivity within organizational structures that were designed long before these capabilities existed.
But how much of the work we’re automating needs to exist? How much of it is simply a consequence of how we’ve organized our companies?
Are we fundamentally improving how our businesses operate, or just becoming remarkably efficient at perpetuating the same organizational dysfunction?
What if the real opportunity isn't to automate the organization we've inherited, but to fundamentally rethink how it should work?
I've spent much of my career building and leading finance functions at companies ranging from early-stage startups to pre-IPO businesses to Fortune 500 enterprises. Over the past several months, I've found myself increasingly questioning some of the assumptions behind how we build and scale companies.
THE ECONOMICS ARE CHANGING
AI is changing the economics of software delivery. Traditional SaaS companies have benefited from relatively low marginal delivery costs and attractive gross margins. AI introduces potentially meaningful variable costs through inference, model APIs, and compute. The economics vary by product and architecture, but the relationship between usage, delivery costs, and profitability is changing.
AI is changing what customers pay for. Software monetization is evolving beyond licenses and seats toward usage, completed work, and measurable outcomes. Subscriptions aren't disappearing, but these models challenge longstanding assumptions about how companies generate revenue and measure customer economics.
AI is changing the relationship between staffing and organizational capacity. Historically, growing businesses have generally needed to add people as revenue, customers, and operational complexity increased. More customers required more support. More transactions created additional administrative work. More employees required additional people infrastructure.
AI creates the possibility of breaking that relationship. Companies may be able to serve more customers, manage greater complexity, and expand into new opportunities without proportionately increasing staffing.
That doesn't necessarily mean employing fewer people. It means reconsidering how much organizational infrastructure is actually required to support growth.
We've begun rethinking the economics of the products we sell and the revenue models we use. But are we applying the same scrutiny to the economics of the organizations we've built to deliver them?
OUR ORGANIZATIONS REFLECT YESTERDAY’S CONSTRAINTS
For decades, we've organized companies around functional specialization.
Finance manages accounting and financial performance. People teams oversee recruiting and employee administration. Operations manages processes and execution. Technology oversees systems and infrastructure.
Each function develops its own workflows, applications, data, reporting, and management structure.
There's good reason for this. Specialized expertise and clear accountability matter. Functional organizations have provided a practical way to manage complexity.
But they've also created inefficiencies we've largely come to accept as part of running a company.
Information gets trapped in separate systems. Departments develop different definitions of the same metrics. Work passes through multiple handoffs and approval processes. Employees spend considerable time reconciling data, coordinating activities, and moving information between functions.
And we build additional organizational infrastructure to manage that complexity.
Consider something as straightforward as hiring an employee.
The need originates in an operating plan. Finance evaluates affordability. A hiring manager defines the position. People operations coordinates recruiting and onboarding. Payroll and benefits systems are updated. IT provisions access. Financial forecasts and organizational reporting need to reflect the change.
Each step makes sense in isolation. But the process crosses multiple functions, systems, and decision points.
Much of the effort isn't actually about hiring someone. It's about coordinating the organization around the act of hiring someone.
What if we could eliminate much of the need for that coordination altogether?
DESIGNING AROUND THE WORK, NOT THE DEPARTMENTS
This is where I believe AI presents a much larger opportunity than incremental productivity improvements.
Imagine designing a company's operating environment around shared information, integrated workflows, and intelligent systems capable of coordinating work across traditional functional boundaries.
Instead of separate departments maintaining their own versions of the business, there are common definitions and connected sources of information.
Instead of employees manually transferring information between applications, workflows move across systems, with AI interpreting context, executing appropriate actions, and escalating decisions that require human judgment.
Instead of each function independently developing its own processes and automation initiatives, the company designs how work should happen from beginning to end.
Return to the hiring example.
An approved operating plan establishes the parameters for a position. A connected workflow coordinates recruiting, approvals, onboarding, payroll, systems access, and updates to financial and organizational reporting.
The appropriate people still make important decisions. Recruiting still requires expertise. Financial controls, privacy, and compliance obligations remain.
But much of the surrounding administrative work and coordination could be reduced or eliminated.
Consider what becomes possible when the operating plan, organizational structure, financial forecasts, and underlying business activity are connected rather than maintained independently. Decisions made in one part of the business can inform the others without requiring separate reporting cycles, reconciliations, and manual updates. Instead of periodically assembling a picture of the company from disconnected sources, leadership could operate from a more consistent, continuously updated understanding of the business.
Now extend that thinking across procurement, customer onboarding, revenue operations, financial management, and workforce planning.
The opportunity becomes considerably larger than automating individual tasks.
Importantly, none of this necessarily requires replacing established systems of record. Accounting platforms, CRMs, payroll systems, and other specialized applications continue to serve important purposes.
The opportunity may be less about replacing those systems and more about how information, intelligence, and actions connect across them.
Of course, integration and process redesign aren't new concepts. Companies have pursued them for decades through ERP implementations, shared services, and digital transformation.
What AI potentially changes is how much of the interpretation, coordination, and execution between systems can be handled intelligently rather than requiring people to manage every step.
We're not simply talking about making existing workflows faster. We're talking about designing workflows that might never have been practical before.
THE EFFICIENCY OPPORTUNITY IS STRUCTURAL
There's an important distinction between making an existing organization more productive and designing one that requires less work to accomplish the same outcome.
A finance team that uses AI to close its books faster has become more efficient.
But a company whose transactions, approvals, reconciliations, and financial information flow through integrated processes may require substantially less effort to close its books in the first place.
The first approach optimizes an existing function.
The second questions whether some of the work that function performs is necessary at all.
We may be carrying significant organizational costs that exist primarily to coordinate the organization itself.
Consider the resources devoted to moving information between systems, reconciling inconsistent data, managing handoffs, maintaining duplicative processes, and coordinating decisions across departments.
These activities are often necessary within today's operating structures. But they don't necessarily represent work that creates value for customers.
AI creates an opportunity to reduce those costs by eliminating unnecessary complexity rather than simply automating the activities that complexity produces.
For startups and scale-ups in particular, this could fundamentally change how organizational capacity scales with revenue.
Rather than continuously adding staff, systems, and management layers to accommodate growth, companies may be able to build operating environments that absorb considerably more complexity without proportionately expanding their administrative infrastructure.
WHEN THE WORK CHANGES, THE ORGANIZATION SHOULD FOLLOW
If we fundamentally redesign how work happens across a business, it follows that we should also question the structures we've established to manage that work.
As information becomes more accessible, workflows become more integrated, and intelligent systems assume more of the coordination burden, some of those boundaries may become less necessary.
That doesn't eliminate the need for expertise, leadership, or accountability.
It changes how those capabilities might be organized.
Rather than structuring the company around departments and then building processes to connect them, we can begin with the information, workflows, and decisions required to operate the business effectively.
That could mean different team structures, broader responsibilities, fewer layers of coordination, and leadership models that cross traditional functional boundaries.
But those outcomes should follow from the design of the work itself, not from an exercise in consolidating departments or changing reporting lines.
The objective isn't to eliminate organizational silos on an org chart. It's to eliminate the friction those silos create in how the business actually operates.
BUILDING DIFFERENTLY
For established companies, this presents an opportunity to distinguish between organizational complexity that's genuinely necessary and complexity inherited from historical decisions about how work gets done.
For founders building companies today, the opportunity may be greater still. They can begin with a fundamentally different set of assumptions.
Rather than establishing traditional departments, selecting applications for each function, and subsequently trying to connect everything together, they can start by considering how the business should operate as a whole.
What information needs to be shared? How should work flow across the organization? Where can AI reliably interpret, coordinate, and execute activities? Where is human expertise and judgment essential? How should accountability and oversight be established?
Then design the systems, processes, and organizational structure around those answers.
I'm not suggesting that every company needs to become an AI experiment or that established operating disciplines should be discarded. Financial controls, security, compliance, and human accountability remain essential.
But I do think we're at a point where the capabilities available to us warrant reconsidering some longstanding assumptions about company design.
The companies that benefit most from AI may not be those that automate the greatest number of tasks. They may be the ones willing to question why those tasks exist in the first place.
And that requires something more difficult than adopting new technology. It requires challenging the assumptions behind how we've built our businesses.
If you were building your company today, knowing what AI makes possible, how much of its existing structure, processes, and complexity would you deliberately choose to recreate?