Nobody can forecast AI in 2030, but plans need a horizon. Three illustrative scenarios, a short list of signposts and five no-regret moves help leaders decide now without betting on one future.
Nobody can say what AI will look like in 2030, and anyone who claims certainty is selling something. But leaders of mission-driven organisations still have to make commitments now that play out over that horizon: hiring plans, licence contracts, training investments, data arrangements, funder promises. Deciding as if the present will simply continue is a forecast too, and usually the least likely one.
Scenario thinking is a more honest alternative. It does not predict. It describes a few plausible futures, identifies the signals that would tell you which one is forming, and asks which moves make sense in all of them. The three scenarios below are illustrations for planning, not forecasts.
The purpose of a scenario is not to guess the future. It is to find the decisions you would not regret in any of them.
Force one: how concentrated AI capability becomes. At one end, a handful of providers supply most of what organisations use, with prices and terms set by them. At the other, capable tools are widely available from many suppliers, including open and regional ones, and switching is easy.
Force two: how hard funders and regulators ask for evidence. At one end, expectations stay loose: a sentence in a proposal and a general assurance. At the other, funders routinely ask for documented policies, records of review and demonstrated capability, much as they now ask for financial controls.
Capability is concentrated in a few providers and evidence expectations remain light. AI works like electricity or cloud storage: essential, cheap for most tasks, and priced and governed by a small number of suppliers. Small organisations gain productivity but have little bargaining power. Terms change and organisations adapt or absorb the cost.
What it rewards. Clear exit plans, disciplined data classification, and staff who can move between tools without losing their workflows. What it punishes. Deep dependence on one vendor's proprietary features, and workflows that no colleague can explain.
Capability is widely available from many suppliers and open models, and funders ask for hard evidence. Tools are plentiful and the scarce resource becomes trust: proof that a given organisation uses AI responsibly and effectively. Documented policy, records of review, measured outcomes and demonstrated staff capability start to function as conditions of funding, much like audited accounts.
What it rewards. Organisations that already keep the evidence file, can show measured results and can explain their reasoning under questioning. What it punishes. Enthusiastic adoption without records, and claims that cannot be verified.
Capability is widely available but evidence expectations stay loose and uneven. Every team, partner and country uses different tools under different rules. Individual productivity is high and organisational coherence is low. Data ends up in many places, quality varies from output to output, and incidents are discovered by accident.
What it rewards. A simple, permission-based policy that people follow, a live tool inventory and a habit of verification. What it punishes. Assuming that access to good tools equals good practice, and relying on rules that staff quietly bypass.
The organisations that do best across all three worlds are not the ones that predicted correctly. They are the ones that can adapt quickly.
A scenario is only useful if you can tell whether it is forming. Keep a short signpost list, reviewed twice a year by the leadership team, with one named person responsible for updating it.
Make signposts specific and dated. "Funders are asking more about AI" is an impression. "Two of our five largest funders added an AI question to their reporting template this year" is a signpost. The second form lets the leadership team argue about evidence rather than mood.
Five actions pay off whichever scenario arrives.
Also keep a short list of moves to avoid: long single-vendor commitments justified by a discount, and headcount plans that assume a single future.
Plan for three plausible futures rather than one predicted future: watch the signposts, make the no-regret moves now, and keep the flexibility to change course when the evidence shifts.
At your next leadership meeting, test the current three-year plan against all three scenarios, choose five dated signposts with a named owner, and schedule the first review within six months.
No external statistics cited. The three scenarios are illustrative planning constructs proposed by M.A.I. Consulting, not forecasts.
One AI supplier means one set of prices, terms and roadmaps. The Portability Ladder tests four things: method, data, workflow and account control. Diversify by being able to move, not by buying twice.
Custom Agents & Tools · 6 minGuide · 27 September 2026The five documents that together let an organisation prove its AI governance to an auditor, board or funder: policy, tool inventory, training record, competency assessment and review record.
AI Use Policy · 4 minGuide · 11 September 2026A five-column spreadsheet of the AI tools staff actually use is the prerequisite for any policy or assessment, and it only works if built in an amnesty spirit.
AI Readiness Assessment · 4 minIf this is the question on your desk, a thirty-minute call tells you whether the service fits, or that you do not need us yet.