The AI Tools Starter Guide

This is a field-tested shortlist of the AI tools worth using in 2026, grouped by the job each one does, written by Amanda Johnstone, an AI technologist and ML patent holder and Founder of Transhuman, who also advises Fortune 500 companies, governments and founders on what to adopt next.

How to read this guide

A woman sitting at a desk holding a smartphone, with a silver laptop and a large green potted plant nearby. She is smiling and looking at her phone.

by Amanda Johnstone.

I have grouped the tools by the work they help you do, not by the company that makes them. That is deliberate. The question is never which AI model is the cleverest in the abstract. The question is which tool belongs in which part of your day.

Start with the category that is slowing you down the most right now. Pick one tool. Give it a fortnight of real use, not a curious afternoon. Then come back for the next one.

Writing and thinking

General purpose AI models are the closest thing we have to a thinking partner on tap, and they sit at the centre of almost every serious AI workflow. They write, edit, summarise, brainstorm, code and explain. If you adopt nothing else, adopt one of these.

ChatGPT is the household name for a reason. It writes, edits, summarises, brainstorms, codes and explains across almost any topic. A strong default first tool if you are new to the category.

Claude is my pick for long form reasoning and tone. Claude handles nuance, voice and extended context particularly well, which makes it a favourite for writers, strategists and anyone drafting serious documents.

Google Gemini is integrated directly into Google Workspace, so it shows up inside Docs, Gmail, Sheets and Meet. The easiest entry point if your team already lives in Google.

Microsoft Copilot is embedded across Microsoft 365. If your organisation runs on Word, Excel, Outlook and Teams, this is where AI meets you in the tools you already use.

Perplexity is AI search with sources. It is built for research and fact finding rather than open ended chat, and it shows you where its answers came from.

Image creation

AI image tools have moved from novelty to production grade. Used well, they compress days of creative iteration into minutes. Used carelessly, they produce the same glossy sameness we all recognise. The skill is in the brief.

Canva AI is a set of AI design tools built into the Canva environment. Ideal for non designers producing social content, presentations and marketing assets at speed.

Midjourney is a high quality image model with a strong aesthetic sensibility. Beloved by art directors and brand teams for the richness of its output.

DALL·E is built into ChatGPT. Convenient if you are already working inside that environment and want image generation in the same place as your writing.

Adobe Firefly is commercially safe image generation trained on licensed and public domain content. The sensible choice when you need images you can confidently use in paid campaigns.

Video creation

Video was the last creative medium to feel the full force of generative AI. That has changed. The tools below are rewriting what a solo creator, a marketing team or a production house can do in a week.

Runway is an AI video generation and editing environment in one. A serious tool for creatives who want control over motion, style and sequence.

HeyGen makes AI avatars for video. Used widely for training, localisation and internal comms, where you need a presenter at scale without filming every take.

Sora is OpenAI's text to video model. Still early, but it is setting the benchmark for what short form generated video can look like.

CapCut is AI video editing built for creators. Fast, mobile friendly and deeply capable, with a feature set that rewards both beginners and professionals.

Audio and voice

Voice is one of the quieter revolutions inside AI and, in my view, one of the most consequential. It changes content, accessibility, customer experience and the texture of our digital interfaces.

ElevenLabs is text to speech and voice cloning at a quality that still surprises first time listeners. Used everywhere from audiobooks to product voiceovers.

Otter.ai is a transcription and meeting summary tool that has quietly become indispensable. Helpful for anyone who attends more meetings than they would like to remember.

Descript lets you edit audio and video by editing the transcript. A different mental model to traditional editing software, and a faster one once it clicks.

Whisper is OpenAI's speech recognition model. More often seen powering other products than used directly, but worth knowing about as the engine behind many audio features.

Productivity and automation

This is where AI stops feeling like a demo and starts feeling like leverage. Used properly, the tools in this section quietly remove hours of repetitive work every week.

Notion AI is AI built into your notes, docs and workflows. Excellent for summarising long pages, drafting inside existing structure and turning messy notes into usable output.

Zapier is workflow automation that connects thousands of apps. The backbone of countless small and medium business operations, now with AI steps layered through it.

Slack is workplace messaging with AI built in. Summaries, catch ups and search that make it easier to stay oriented across channels and threads.

Business and commerce

Every serious business platform is now an AI platform. These are the tools where AI is already changing how money moves, how customers are served and how regulated work gets done.

Shopify has AI woven through ecommerce, from product descriptions to merchandising and insights. A core platform for anyone building a modern online store.

Heidi Health is AI medical documentation built for clinicians. A strong example of vertical AI, designed around the realities of a specific professional workflow.

Xero is AI accounting for small and medium businesses. Helpful for automating reconciliation, reporting and the administrative edges of running a company.

QuickBooks is AI accounting that surfaces patterns in your numbers. Useful for business owners who want a clearer read on cash, margin and trends without hiring a team.

Salesforce Einstein is AI across the Salesforce CRM. Used by enterprise sales, service and marketing teams to forecast, personalise and automate customer interactions at scale.

Research and analysis

One of the fastest moving and most underrated parts of the AI stack. If your work involves reading, synthesising or making sense of a lot of information, this is where AI pays for itself first.

Consensus is an academic search engine that summarises what the research actually says on a given question. A thoughtful alternative to generic search for evidence based work.

Elicit is a research assistant for literature review, study synthesis and structured analysis. Particularly strong for academics, analysts and policy professionals.

NotebookLM is Google's AI layer over your own documents. Upload your sources and have a grounded conversation with them, with citations back to the original material.

Questions I'm often asked

What is the best AI tool to start with if you have never used one?

ChatGPT is the strongest default starting point. It writes, edits, summarises, brainstorms, codes and explains across almost any topic, and its interface is the most familiar on-ramp for new users. Once you are comfortable, Claude is the better pick for long form reasoning and tone, while Gemini and Copilot are worth using if your organisation already lives inside Google Workspace or Microsoft 365.

Which AI tool is best for long form writing and nuanced work?

Claude, every time. It handles nuance, voice and extended context in a way the other models do not yet match. ChatGPT is stronger as a general all-rounder, and Perplexity is better for research with sources, but for substantive writing that needs to sound human, Claude is where I land (as at 20 April 2026).

What AI tools should executives and leadership teams actually adopt?

The foundational stack for most leadership teams is a general purpose model (ChatGPT, Claude, Gemini or Copilot depending on your existing software), Perplexity for research, Otter or Descript for meetings, and whichever AI layer is native to your core platforms, such as Salesforce Einstein, Shopify, Xero or Notion. The mistake is collecting tools. The win is choosing one tool per category, using it on real work every day, and building the instinct for when a new one deserves your attention.

How should leaders think about adopting AI over the next twelve months?

Use AI on something real every day. Notice what it changes. Notice what it does not. That is how you develop the instinct that no guide, including this one, can give you. The organisations pulling ahead right now are not the ones with the longest tool lists. They are the ones who have built the habit of experimenting, choosing and integrating rather than accumulating.

How to think about this list in twelve months

Half of the names on this page will look different by the end of the year. Some will become quietly essential. A few will be absorbed into larger platforms. New categories will appear that I have not even named here.

That is not a reason to wait. It is a reason to start. The people and organisations who are pulling ahead right now are not the ones with the longest tool lists. They are the ones who have built the habit of experimenting, choosing and integrating AI into how they already work.

My own rule is simple. Use AI on something real every day. Notice what it changes. Notice what it does not. That is how you develop the instinct that no guide, including this one, can give you.