Right now, unlimited AI usage is cheap. £200 a month can buy you access to some of the best models available, used as much as you like. It won’t stay this way.
Prices for AI access move in one direction as demand grows and providers work out what customers will actually pay. The £200-a-month “all you can use” plan of today could easily be £2,000 next year, or gone altogether, replaced by a metered plan that charges per token, per seat, or per “agent.” Providers change their terms. Models get deprecated. Free tiers disappear. None of this is hypothetical — it has already happened repeatedly across the AI industry in the last two years.
The real risk isn’t the price. It’s the lock-in.
If your business has built its workflows, your customer history, and your team’s institutional memory entirely inside one vendor’s chat window or one model’s context window, you don’t have a tool. You have a dependency. When the price changes, or the model is retired, or the provider changes its policy on how your data can be used, you’re not in a negotiating position — you’re stuck.
The businesses that will do well over the next few years are the ones that treat AI models as replaceable components, not as the foundation everything else sits on.
What actually needs to be permanent
Three things should never live inside a single AI provider’s product:
- Your history. Every conversation, every decision, every piece of context your business has built up with AI assistance is worth more than any individual model. That record should sit in your own system, not scattered across whichever chat tool happened to be cheapest that quarter.
- Your memory. The facts, preferences, and context that make AI responses useful to your specific business shouldn’t need re-explaining every time you switch tools. If moving providers means starting from zero, you’re not free to move — and providers know it.
- The ability to swap. Different models are good at different things, and that will keep changing month to month. The businesses with the most flexibility will be the ones that can point their workflows at a new model or a new harness without rebuilding everything underneath it.
How WI1 fits into this
WI1 is being built so the CRM — your leads, customers, documents, tasks, and communications — stays in one place regardless of which AI model or provider sits behind it on any given day. Your history and memory live in WI1, not inside a chatbot’s session. The model doing the thinking is a component we can swap out as the market shifts, not the thing your business is built on top of.
We’re rolling this out gradually — merging in model flexibility, better memory handling, and support for swapping between models and harnesses as those pieces mature, rather than waiting for one big-bang release. The direction is the same throughout: you keep control of your data and your history, no matter what happens to any one model’s price or availability.
Plans and prices for AI access are going to keep moving. Make sure your business doesn’t have to move with them.