Learn
Picker
Which AI model should you use?
Five quick questions about the job, your data, your budget and how you work. It gives you a plain-English recommendation: the kind of model to reach for, a named current example, an alternative and one honest caveat. No sign-up, nothing sent, worked out in your browser.
Your recommendation
- Also worth a look
- One caveat
Models change fast. This is a signpost as of July 2026, not gospel. To re-check, open thecomparison table and confirm the detail on each maker's own page before you commit.
How it works
No black box, just sensible rules.
- Five questions, weighted. Your answers on the task, data sensitivity, budget, priority and team each nudge the result toward a kind of model: a balanced all-rounder, a top-tier reasoning model, a fast and cheap one, a research and workspace model, or a self-hosted open one.
- The safe default wins ties. When it's close, it leans to a balanced all-rounder, because that is the right call for most business work most of the time.
- Data sensitivity shapes the caveat. Flag confidential data and it will remind you to check the provider's data-training terms, use a business plan, or self-host, and never paste secrets you cannot risk.
- Named examples are dated. We name a current example so the advice is concrete, and stamp it July 2026. Models move fast, so the fullcomparison table is the place to re-check before you commit.
Got a model in mind? Come and use it.
Bring your actual work to our AI Automation Masterclass in Manchester and we'll help you pick and set it up on the day. Tickets are normally £20. This one's free, a limited-time offer to launch the series.
Reserve your seat