The AI tool landscape changes fast and most "best AI tools" lists are either outdated or sponsored. This is a practical assessment based on real use — what each tool is actually good at, what it costs, and which ones you should be running in your workflow right now.
How to Read This List
This isn't a ranking — no tool wins across all tasks. This is a use-case map. The right tool depends on what you're doing. For each category, I've noted what it's best for and what it doesn't do well, based on actual daily use.
For General Reasoning and Writing
Claude (Anthropic)
Best for: Long documents, nuanced writing, following complex instructions faithfully, anything where tone and precision matter. Handles very long context windows better than most. Produces writing that sounds less machine-generated than many competitors.
Less suited to: Real-time web search without tools, image generation.
Verdict: First choice for serious writing, analysis, and complex reasoning tasks.
ChatGPT (OpenAI)
Best for: Breadth of tasks, plugin integrations, Code Interpreter for data analysis, image generation via DALL·E, and the widest range of third-party integrations.
Verdict: The Swiss Army knife. Not always the best at any single task, but the most versatile. If you're choosing one paid tool, this is still the default.
Gemini Advanced (Google)
Best for: Google Workspace integration, summarising YouTube videos and Gmail, real-time Google Search grounding, multimodal tasks.
Verdict: Underrated for people in the Google ecosystem. The Workspace integration alone justifies it for teams who live in Docs and Gmail.
For Research
Perplexity
Best for: Answering questions with cited, real-time web sources. Replaces significant portions of search for factual questions and market research.
Verdict: The best AI-powered research tool available. Anyone doing regular research should have this in their stack.
NotebookLM (Google)
Best for: Synthesising questions across multiple uploaded documents with cited answers. Free.
Verdict: Uniquely useful for document-heavy work. Underused.
For Coding
Cursor
Best for: AI-native code editor with deep context of your entire codebase. Better than traditional autocomplete tools for working in large existing codebases.
Verdict: The current best option for developers who want AI deeply integrated into their workflow.
The Stack That Makes Sense
For most knowledge workers:
- Claude Pro or ChatGPT Plus — primary AI assistant (~$20/month)
- Perplexity Pro — research (~$20/month)
- NotebookLM — document synthesis (free)
Roughly £30–40/month. For most knowledge workers, that's 30 minutes of recovered time per day minimum.