How AI Is Rewriting the Logic of Work

In 2010, the big conversation inside startups was about the cloud. How fast should we move? What about security? What happens to IT once everything lives in AWS?

Fifteen years later, those questions sound quaint. The new ones are harder.

Every operator, founder, and product lead is asking the same thing in different forms: How fast should we lean into AI? How do we measure productivity when half the work is invisible? When your software starts doing the work for you, what does “operations” even mean?

We’re still early, but the answers are starting to take shape.

After raising $200M+ from top VCs, Pacaso made a bold move: they opened their cap table to retail investors. The result? $72.5M from 17,500+ everyday investors.

Join Pacaso CEO Austin Allison, CLO David Willbrand, and DealMaker CEO Rebecca Kacaba for an exclusive conversation on why growth-stage companies are combining institutional and retail capital, and how 17,500+ investor-advocates provide value.

Cloud adoption was a grind. Convincing a bank or healthcare provider to move data off-premises could take years. Banks were among the first to adopt cloud computing at scale around 2010 because they had the capital and interest in emerging trends, but widespread enterprise adoption came slowly. Industry analysts predicted that the transition from early adopters to mass adoption would take until 2015.

AI is the opposite. Enterprises aren’t asking whether they should use it—they’re asking how many experiments they can run.

There’s a reason for that. Cloud was infrastructure. It required faith. AI is visible. Every executive has seen ChatGPT, every team has a pilot project, and the ROI—while fuzzy—feels immediate.

The numbers bear this out. Generative AI adoption jumped from 65% in early 2024 to 71% of organizations by 2025, with healthcare, manufacturing, and IT sectors experiencing the most dramatic growth. Overall AI adoption surged from 50% in 2022 to between 72% and 78% in 2024—a pace that took cloud computing years longer to achieve. Enterprise AI adoption reached 78% in 2024, representing a decisive shift from experimental to operational deployment.

But here’s the catch: even if adoption is faster, the integration problem is harder. The cloud reshaped where data lived. AI reshapes how work happens. That means the next decade of operations won’t be about migrating systems. It’ll be about reimagining workflows.

In cloud-era SaaS, you sold tools to people. In AI-era software, you sell outcomes to workflows.

The unit of value is shifting. Old model: pay per seat, per license, per human. New model: pay per task, per query, per agent.

That might sound abstract until you look at how usage-based pricing hits a P&L. When your costs depend on model queries, not logins, predictability vanishes.

The data reveals a fundamental pricing transformation underway. Seat-based pricing for AI companies dropped from 21% to 15% in just 12 months, while hybrid pricing models surged from 27% to 41%. 59% of software companies expect usage-based approaches to grow as a percentage of overall revenue in 2025, representing an 18% rise from 2023. Companies that stick with traditional per-seat pricing for AI products see 40% lower gross margins and 2.3 times higher churn than those adopting usage or outcome-based models.

Organizations spent an average of $400,000 on AI-native applications in 2025, a 75.2% year-over-year increase. The financial impact is real and growing, forcing operators to rethink everything from forecasting to customer success metrics.

Box’s hybrid model is a sign of what’s coming: seat-based access for humans, credit-based pricing for AI usage. It keeps finance sane while letting heavy users experiment. Expect that hybrid to spread.

Operators who still think in seats and licenses will feel the ground move beneath them.

Most companies still talk about AI as if it’s a faster spreadsheet—a way to remove steps, automate tasks, trim headcount.

That misses the point. The real leverage comes from rearranging work, not removing it.

Every process has two axes: how often it happens and how much judgment it needs.

The sweet spot for AI isn’t high-volume, low-thinking work—that’s RPA territory. It’s in the middle: repeatable processes that still require reasoning.

Manufacturing has embraced AI at a remarkable speed, with 77% of manufacturers now utilizing AI solutions, reporting an average 23% reduction in downtime from AI-powered process automation. Retailers that deployed AI-driven chatbots during the 2024 Black Friday sales reported a 15% increase in conversion rates, while AI-powered inventory systems reduced overstocking by an average of 18%.

Operators who find that zone—where automation meets context—will extract the most ROI. Those who chase full automation will hit governance walls, security risk, or user backlash.

In the cloud era, scale and integrations were moats. In the AI era, entanglement is.

The more your system touches a company’s real workflows—their data, permissions, and agent logic—the harder it is to replace.

For operators, that means two things:

Winning in this environment means thinking less about user retention and more about workflow gravity.

There’s no such thing as an AI budget yet. Most enterprise teams are pulling from “innovation funds” or R&D lines. That’s 1–10% of IT spend at best.

At the same time, the pitch flood has hit absurd levels. Everyone’s “AI for X.” Customers are exhausted.

In a 2025 survey1, 48% of IT buyers planned to increase their spending on AI and GenAI in the next 12 months, with 68% of vendors charging separately for AI enhancements or including them exclusively within premium tiers. The willingness to pay is there, but pricing models remain in flux.

73% of AI companies are still experimenting with their pricing models, with the average company testing 3.2 different approaches in their first 18 months. CFOs at data infrastructure companies admit they’re not monetizing AI to juice revenue—they’re monetizing to avoid eating $10,000 of costs on a $500 plan.

That’s why smart operators are embedding AI as a native feature, not a paid add-on. If your value is real, it shows up in adoption metrics, not upcharges.

For the next few years, pricing will look like controlled chaos. The experiments that stick will mix predictability for finance (base subscription) with flexibility for users (credit pools, usage tiers).

You can’t bolt security onto AI. It is the product.

Every company that’s scaled AI internally has run into the same walls: data visibility and permissioning, unintended disclosures, shadow prompts that surface sensitive info.

The governance crisis is real and urgent. 45% of enterprise employees already use generative AI tools, with 67% of AI usage occurring through unmanaged personal accounts, leaving security leaders blind to data flows. 40% of files uploaded into GenAI tools contain personally identifiable or payment card industry data, with employees using personal accounts for nearly four in ten of those uploads.

69% of organizations cite AI-powered data leaks as their top security concern in 2025, yet nearly half have no AI-specific security controls in place. 64% of organizations lack full visibility into their AI risks, and almost 40% admit they lack the tools to protect AI-accessible data.

Governance used to be invisible to end users. Now it defines user trust. Operators who treat guardrails as UX, not compliance, will win.

The goal isn’t to control everything. It’s to make sure the system acts as expected, with context and traceability. The most underrated skill inside AI-era operations will be policy design.

It’s tempting to build low. Every company wants its own model, its own infrastructure, its own GPU cluster.

That’s a trap. The infra war is already lost to the giants. The advantage now lies at the application layer—how well you orchestrate and apply models, not how you train them.

The competitive landscape has shifted dramatically, with Anthropic capturing 32% of enterprise market share compared to OpenAI’s 25% and Google’s 20% in 2025. 37% of enterprises now use five or more models in production environments, indicating that multi-model deployment has become the norm.

The new leverage is vertical expertise and trust. If you run operations, your job isn’t to train models. It’s to build a system that uses them responsibly, measurably, and repeatably.

The cloud changed how teams collaborated. AI will change how they think.

In every company, there’s a quiet divide forming: One group sees AI as relief—fewer repetitive tasks, faster iteration. The other sees it as threat—loss of control, loss of craft.

42% of C-suite executives report that AI adoption is tearing their company apart, with 68% reporting friction between IT and other departments and 72% observing that AI applications are developed in silos. Yet the research also shows a clear path forward: At companies without a formal AI strategy, only 37% of executives report being very successful at adopting AI, compared to 80% at companies with a strategy.

Operators bridge that divide. You set the norms for experimentation, transparency, and trust. If your AI rollout feels like a secret project, adoption will rot.

The best operators right now are doing the opposite. They document, share, and normalize the weirdness of building with AI. They let teams test, fail, and iterate publicly. Over three-quarters of employees using AI already self-identify as AI champions or see the potential to become one, suggesting that when employees engage with the right AI tools, they become enthusiastic advocates.

Culture, not tooling, becomes the compounding asset.

AI adoption won’t be a sprint. It’ll be a 10-year migration, just like the cloud—faster, but just as deep.

The challenge won’t be convincing people. It’ll be integrating systems. The cloud changed infrastructure. AI changes cognition.

According to the Census Bureau’s Business Trends and Outlook Survey2, AI adoption among US firms has more than doubled in the past two years, rising from 3.7% in fall 2023 to 9.7% in early August 2025. Despite this rapid growth, the vast majority of firms in the US do not yet report using AI in their production processes, with usage remaining unevenly distributed across the economy.

In 2025, 31% of AI use cases studied reached full production, which is double the amount compared to 2024, but expectations that AI would cut costs and boost productivity are underdelivering. The gap between pilot and production remains the defining challenge.

That’s why this shift belongs to operators. You’re the ones sitting between models and people, between strategy and execution. Your choices—on pricing, data, workflows, governance—will decide which companies make it through the transition.

The cloud rewarded teams who scaled fast. The AI era will reward teams who adapt fast. The difference is subtle but defining. You can’t brute-force your way into this shift. You have to understand it, layer by layer, like an operator who knows the system inside out.