The 9 AI tools I use most often
There are now thousands of AI tools available but which are the ones I use most often and why?
Below is a table of nine of the AI tools I turn to on a regular basis, what I find they’re best for, along with ethical considerations you might want to weigh up when deciding if and when to make use of them.
Research
Perplexity has replaced Google as my research tool of choice. Always getting a quick, readable response, with citations I can check if necessary, makes wading through reams of search results, many of them sponsored and/or spammy, decidedly unappealing.
Idea generation / summarisation / analysis / drafting
Claude is currently my preferred AI chatbot (nicer tone and user experience plus more ethical responses). Common uses include idea generation, document summarisation/analysis and feedback on drafts. I still turn to ChatGPT when my prompts require web access, knowledge about me or I want to use my voice instead of a keyboard (recently discovered use case: brainstorming whilst driving).
Image generation
Midjourney has been my go-to image generator for a long-time due to the quality of output and the level of control (e.g. the ability to easily regenerate parts of an image), with Ideogram my fallback when the image needed to include text. However, I find I’m increasingly turning to Ideogram first, which now seems to be matching Midjourney on quality and doing slightly better on prompt adherence.
Image editing
Removebg is a single use tool which does one thing well: removing backgrounds from images. A task that would previously have involved painstakingly tracing the outline of the part of the image I wanted to retain.
Transcription & audio clean-up
Descript is a veritable Swiss-army knife of audio (and increasingly video) tools. I primarily use it for cleaning up audio and creating audiograms.
Sounds effects & voice cloning
ElevenLabs’ laser focus on audio has kept it ahead of those who are trying to do AI audio alongside a bunch of other things and it’s still the place I turn for voice clones and sound effects (although I’m now going to be checking the recently expanded BBC Sound Effects archive first). Their ElevenReader mobile app is great for listening to articles on the go (everything sounds better when narrated by Burt Reynolds - one of their licensed voices).
Translation
DeepL is another company whose laser focus is helping it outrun Google when it comes to accurate translations. My go-to when needing an accurate translation for client work.
Why no Gemini, Copilot or Llama?
Google Gemini continues to improve but I find it’s only my first port of call when the tasks requires a bigger context window. Despite being built on top of OpenAI’s models, I’ve always found Microsoft Copilot pretty underwhelming and my first experiences of its upgraded voice mode haven’t been great. I have Meta’s Llama 3.2 3B running locally on my iPhone, although I tend to only turn to it when I don’t have an internet connection.
What are the AI tools you use on a regular basis and why? Comment below👇
Great list, Dan. Very similar to my own daily use, but surprised you’re not using NotebookLM more.
I only realised it had rolled out in the UK thanks to the publicity around its slightly gimmicky podcast creation feature, but have found its core functionality incredibly useful. You can upload mountains of project docs or notes and interrogate them using natural language. E.g. “what’s the link to that Miro board we made at the start of the year to sort out that crisis?” I use it to consolidate all of my notes about something in one place, then give the output to Claude or ChatGPT for brainstorming (Gemini, which powers NotebookLM, is fine for retrieving info, but still lags far behind when turning it into something new).
As always, really clear. Does anyone else offer this kind of detailed comparative analysis?