PESTEL after AI: the weather is changing while we build

PESTEL is a way to scan the forces around a business: Political, Economic, Social, Technological, Environmental and Legal. It usually asks what is happening outside the company that could help or hurt it.

AI makes that scan more interesting. A company can now create a product faster, while its customers, competitors, regulators and infrastructure providers are changing just as quickly. The same trend can be a tailwind at one point in the business and a headwind at another. Cheaper software creation, for example, helps you build. It also helps everyone else build.

Here is what each lens reveals.

Political: AI is becoming a matter of national capacity

Governments increasingly see AI capability as something to secure, much like energy, manufacturing or telecommunications. National strategies and public investment are expanding, while countries seek more control over the computing infrastructure behind AI. That creates a tailwind for businesses that can help institutions use AI in local languages, public services and domestic industries. It also creates a headwind to watch: dependence on chips, computing power or cross-border data that a business does not control. Stanford’s 2026 AI Index^[1] documents the growth in national strategies and the uneven distribution of infrastructure.

The strategic question is no longer only, “Can we use AI?” It is also, “Whose systems do we depend on, and where must this capability work?”

Economic: the cost of making falls; the cost of standing out rises

AI can reduce the time needed to research, write, code, design and support a product. Small teams can attempt work that once required larger ones. That is a real tailwind, though faster production does not guarantee lower total costs.

The headwind is that customers gain more options at the same time. A feature built quickly is also a feature competitors may copy quickly. And adoption does not automatically produce a sound business. In McKinsey’s 2026 survey, nearly nine in ten respondents reported regular AI use in at least one business function, but 44% said their organizations were scaling it across the enterprise. Only 37% attributed a positive effect on operating profit to AI. These are reported experiences from survey respondents, not a measure of the return every company can expect. The opportunity is real; so is the gap between using AI and making it pay. McKinsey 2026 survey^[2].

A useful economic test is: Where does AI improve the whole outcome for a customer, after the costs of review, mistakes, integration and maintenance?

Social: people may use AI and distrust it at the same time

People do not have to choose between finding AI useful and feeling uneasy about it. Ipsos’s 2026 survey across 32 countries found excitement and nervousness at similar levels, sometimes in the same respondents. Its country average should be read as a view of the surveyed markets, not a count of world opinion. Ipsos 2026 findings and methodology^[3].

That tension creates a tailwind for services that give people useful help, clearer choices and more access to expertise. The headwind appears when a product asks for trust it has not earned: an unexplained decision, a synthetic voice presented as a person, or an automated answer with no clear route to correction.

For many businesses, the social question will be less “Will people try AI?” and more “When will they let it act for them?”

Technological: capability moves fast; dependable systems take longer

Better models make more ideas feasible. Teams can prototype interfaces, analyze documents and connect work across tools faster than before. That is the technological tailwind.

But a convincing demonstration is only one part of a working system. The headwind is in the surrounding work: accurate data, permissions, testing, security, recovery when something fails, and a person or process accountable for the result. Stanford’s 2026 AI Index reports major gains in agents’ ability to complete computer tasks, while noting that they still fail roughly one in three attempts on a structured benchmark. That result is not a failure rate for every real-world use; it is a reminder to test the task that actually matters. Stanford’s technical performance chapter^[4]. NIST’s voluntary AI Risk Management Framework^[5] offers one way to organize that wider work.

The advantage may therefore shift from possessing an AI feature to knowing exactly where it belongs, how well it performs and when it should stop.

Environmental: AI has a physical footprint

AI can help forecast demand, monitor equipment and use energy systems more efficiently. Those are meaningful tailwinds. But AI also needs data centers, electricity, cooling and grid connections. The International Energy Agency’s April 2026 central projection has electricity use by all data centers rising from 485 terawatt-hours in 2025 to about 950 in 2030. AI is a major driver, but the projection is not a measure of AI’s electricity use alone. It depends on efficiency, uptake, infrastructure constraints and the kinds of tasks people choose to run. IEA, Key Questions on Energy and AI^[6].

The headwind is local as well as global. A data center may face a constrained grid and a community concerned about power prices. A business choosing AI systems has to ask what each use is worth, including the resources required to run it.

Legal: responsibility is becoming part of the product

Rules around AI are moving from general principles toward duties tied to particular uses. In the EU, for example, the AI Act applies different requirements according to risk and includes transparency rules for certain AI interactions and generated content. The Act became broadly applicable in August 2026, while some obligations for high-risk systems take effect later, in 2027 or 2028. Which duties apply depends on the system and the organization’s role. European Commission’s AI Act overview^[7].

That is a headwind for teams that have not kept track of their data, model providers or automated decisions. It can be a tailwind for businesses that can show what their system does, explain its limits and give customers appropriate oversight. The legal question is practical: If this system affects someone, can we account for how it was built and used?

The point of the scan

These six forces do not stay in separate boxes. A political drive for domestic AI changes infrastructure spending. Infrastructure spending changes energy demand. Energy demand changes local sentiment. Public trust influences regulation. Regulation changes the economics of a product.

That is what PESTEL means in the time of AI: a way to trace those connections before making a bet. For any opportunity, name the tailwind, name the headwind it may create, and ask which parts of the business can still hold up when both arrive.