AI Tools for Every Task, September 2026
The seventeen AI tools I would hand you to get more done this month: make content faster, automate the boring parts, research in minutes, write in your own voice, and make images that look professional.
Save it, print it, read it offline. It's yours.
These are the tools I’d hand you if you told me you wanted to get more done this month without learning anything technical. Seventeen apps, sorted by the job you’re trying to do: make content, automate the boring parts, research faster, write better, think out loud, and make images that look like you paid someone. Every one of them saves you time or money, and most are free to start.
This is the September 2026 list. A few of these didn’t exist six months ago, and some will be replaced by something better before the year is out. When that happens, I’ll update it, so you don’t have to keep up with all of it yourself.
Each one earned its spot by doing one job noticeably better than whatever you’re probably using now. Open any of them tonight and you’ll get something back.






















Content gets four slots because it’s four different jobs: the graphics, the edit, the repurposing, and the brand voice. Nobody has one app that does all four well.
Canva
What it is. The design app for people who aren’t designers. Posts, carousels, thumbnails, one-pagers, slides, all from templates you can actually understand.
Why it’s on my list. I’ve used Canva for years and it’s where every graphic I post starts. The AI pieces (Magic Write, background removal, resizing one design into every platform’s size) mean I stopped opening anything else for static visuals. The cards you saw in the video live in a Canva folder so I can grab them from any device.
Best for. Anything static that has to look finished: a quote card, a carousel, a thumbnail, a PDF you’re handing to a client. Not for video editing, even though it can. CapCut does that job better.
Do this first. Open Canva, pick any Instagram post template, and use Magic Write with this: “Rewrite this headline in my voice, plain and direct, no exclamation points: [paste your headline].” Then hit Resize and make the TikTok, Facebook, and story sizes from the same design. That’s four platforms from one file.
Try it here: canva.com
CapCut
What it is. A video editor built for phone-shaped video. Auto captions, one-tap cuts, text that pops on beat.
Why it’s on my list. Every video I post gets edited in CapCut. The captions alone are the reason: most people watch on mute, and CapCut’s auto captions are accurate enough that I fix a word or two instead of retyping. The cards in this video were placed as overlays in CapCut in a few minutes.
Best for. Talking-head video, tutorials, anything short and vertical. Save your one template with your caption style and reuse it every time so you’re never rebuilding from scratch.
Do this first. Import one clip you already filmed. Text, then Auto captions. Fix any misheard words by tapping them. Save the caption style as your default so every future video gets it automatically. Do this once and you never rebuild captions again.
Try it here: capcut.com
OpusClip
What it is. You give it one long video. It hands back a stack of short clips, each cut around a moment it thinks will hold attention, with captions already on.
Why it’s on my list. Repurposing is the job everyone says they’ll do and nobody does, because cutting a 40-minute recording into ten clips by hand takes an afternoon. OpusClip takes the afternoon away. It’s the difference between a long video being one post and being two weeks of posts.
Best for. Anyone with long recordings sitting around: a workshop, a podcast, a Zoom training, a livestream, even a long voice memo. The longer the recording, the more it hands back.
Do this first. Upload the longest recording you have, even a Zoom call. Let it cut. Pick the two clips it scored highest, add your own first line as a caption, and post one this week. If you don’t have a long recording yet, prop your phone up and talk for ten minutes about the questions people ask you most. That’s your first one.
Try it here: opus.pro
Pomelli
What it is. A Google Labs experiment. You paste in your website address, it reads your brand (colors, tone, what you sell) and generates social content that already sounds and looks like you.
Why it’s on my list. It’s the first tool I’ve seen that starts from your brand instead of from a blank prompt. For a small business owner who has a website and no time, that’s the whole problem solved in the right order. It’s new, it’s free, and almost nobody’s showing it yet.
Best for. A business with a real website and a thin social presence. Give it your URL, let it draft a month, then edit the drafts instead of writing from nothing.
Do this first. Paste your website address and let it read your brand. Ask for “ten posts for the next two weeks about [the one thing you most want people to buy or book].” Delete the ones that don’t sound like you. Edit the rest instead of writing from nothing.
Try it here: labs.google/pomelli


Automation sounds technical. It isn’t anymore. Opal is where you start if you’ve never automated anything, and Grok Bot is what it looks like when the AI does the whole job while you’re away.
Opal
What it is. Another Google Labs experiment. You describe a small app in plain English (“an app that takes a customer complaint and drafts three reply options”) and Opal builds it, shows you the steps as a visual chain, and lets you edit or share it.
Why it’s on my list. Every automation tool before this asked you to learn the tool first. Opal asks you to describe the job. It’s the first no-code builder I’d hand to a total beginner without a tutorial. And it’s free.
Best for. Repetitive little jobs you do the same way every time: turning a meeting note into an email, turning a product description into three ad variations, checking a form against a checklist. Start with one that annoys you weekly.
Do this first. Build one app with this exact description: “An app where I paste a customer message and get back three reply options, one short, one friendly, one firm. Same tone I’d use, no corporate words.” Use it the next time a customer emails.
Try it here: opal.google
Grok Bot
What it is. xAI’s newest product, in beta since August. Instead of a chatbot that answers you, it’s a roster of always-on agents that each get their own cloud computer, sign into your existing tools, and work through multi-step jobs without you watching. Research accounts overnight, keep records clean, process invoices, flag risks.
Why it’s on my list. This is the version of “AI assistant” people have been promised for two years: it does the job, not a draft of the job. It’s early, and it’s the thing to know about now, because this is where every AI company is heading and you’ll want to be ahead of it.
Best for. A business with a repeatable back-office process that eats hours: invoice follow-up, CRM cleanup, lead research, the weekly report. Anything you’d describe as “the same thing every Monday.”
Do this first. Write down the one back-office job you’d hand to an assistant tomorrow if you had one: chasing invoices, cleaning up your contact list, researching leads. If it’s something you do the same way every week, that’s your first Bot. Keep the note. When you’re ready to try it, you’ll already know exactly what to ask it to do.
Try it here: grok.com


One looks outward at the whole internet. One looks inward at your own files. You need both, and people keep using one for the other’s job.
Perplexity
What it is. A search engine that answers the question instead of handing you ten links, and shows its sources next to every claim so you can click through and check.
Why it’s on my list. It’s the fastest way I know to go from “I don’t know anything about this” to “I have the shape of it and three sources to read.” The citations are the point. A chatbot that answers confidently with no source is how you end up repeating something wrong in front of a client.
Best for. Anything about the outside world: a competitor, a regulation, a tool you’re considering, a trend somebody mentioned. Ask it the plain question, then ask it “what would someone who disagrees say?”
Do this first. Ask it something you’d normally Google in three tabs: “What do I need to know before [buying / signing / choosing X]? Give me the three things people regret not checking, with sources.” Then ask: “What would someone who disagrees with this say?” Two questions, both sides, sources for each.
Try it here: perplexity.ai
NotebookLM
What it is. Google’s research notebook. You upload your own documents (a contract, a client folder, meeting notes, a 60-page PDF) and it answers questions using only what you gave it, with a pointer back to the exact passage.
Why it’s on my list. Most AI tools guess when they don’t know. NotebookLM is built to not guess. It stays inside your sources, which makes it the one I’d trust with a contract or a set of client notes. Students found it first. Business owners are the ones who should be using it.
Best for. “What does this contract say about cancellation?” “What did we agree to across these four meetings?” “Summarize this 60-page report and tell me what’s missing.” Anything where the answer has to come from your documents, not the internet.
Do this first. Upload one document you’ve been avoiding, a contract, a lease, an insurance policy, a 40-page PDF. Ask: “What in here can cost me money, and what would I need to do to avoid it? Point me to the exact section for each.” Then ask: “What’s missing from this that I’d expect to see?”
Try it here: notebooklm.google.com


I use both of these every day. Claude gets the long pieces, ChatGPT gets the quick ones, and between them almost nothing I write starts from a blank page anymore.
Claude
What it is. Anthropic’s assistant, and the best writer of the bunch.
Why it’s on my list. Claude sounds the least like a robot. It holds a long document together, it doesn’t pad, and when you hand it real source material it’s careful about not inventing details. Almost everything long on this site was drafted in Claude, including this guide.
Best for. Anything long or structured: a guide, a proposal, a report, an email sequence, a page of your website. Give it three samples of your own writing first and tell it to match your voice, and it will hold that voice across the whole conversation.
Do this first. Paste three things you wrote yourself (a caption, an email, a text) and say: “Read these and describe my voice in five bullet points: sentence length, how formal I am, words I use a lot, words I’d never use, how I open and close. Then keep that voice for everything you write for me in this conversation.” Everything after that comes out sounding like you.
Try it here: claude.ai
ChatGPT
What it is. The assistant everyone already has, and the fastest at short writing that has to sound like a person.
Why it’s on my list. For a caption, a reply, a subject line, a text to a client, ChatGPT gets to a good version in one pass. It’s also the one that’s easiest to talk to (see the next section). Claude for the long piece, ChatGPT for the short one, is how I actually work.
Best for. Captions, replies, quick rewrites, “make this sound less stiff,” anything under a couple of paragraphs. Use ChatGPT’s Projects feature (the folder in the left sidebar) so it remembers your voice between chats.
Do this first. Open Projects in the left sidebar, start one named after your business, and paste your voice notes from the Claude prompt above into its instructions. Then every quick caption or reply you ask for in that Project already knows how you talk. Ask: “Make this sound like me, shorter, no exclamation points: [paste].”
Try it here: chatgpt.com




Four slots because brainstorming has two halves: getting the ideas out of your head, and then sorting them. The first two are for talking. The second two are for sorting.
ChatGPT Voice
What it is. ChatGPT’s voice mode. Tap the sound icon, talk, and it talks back, in a real conversation instead of a typed one.
Why it’s on my list. You can talk faster than you can type, and you naturally give more context when you’re speaking. In two minutes of talking you’ll tell it things about your situation that you’d never bother typing, and the answer gets better because of it. It’s also the easiest way to start using AI if you’ve been putting it off, because it feels like a phone call, not a form.
Best for. Working through an idea while you drive, walk, or cook. Explaining a messy situation and asking what to do next. Practicing a conversation before you have it. Anything where you’d rather talk it out than write it down.
Do this first. Next time you’re in the car or on a walk, open voice mode and say: “I’ve got an idea I want to talk through. Ask me questions until you understand it, then tell me what you’d do.” Talk for two minutes. When you’re done, ask it to summarize what you said in five bullet points. That’s a brainstorm you’d never have typed.
Try it here: chatgpt.com
Wispr Flow
What it is. Voice typing for your computer. Hold down a key, talk, let go, and clean text appears wherever your cursor is: an email, a prompt box, a document, a text message. It cleans up the “ums” and the rambling as it goes, so what lands on the screen reads like you typed it carefully.
Why it’s on my list. I use Wispr Flow constantly, and most of my work happens at my desk. Almost every prompt I give Claude or ChatGPT starts with me holding the key down and talking, because I can say a full paragraph of context in the time it takes to type a sentence. Long, detailed prompts get good answers, and talking is how you write a long prompt without it feeling like work.
Best for. Prompts, emails, notes, first drafts of anything. It works in every app on your computer, so once it’s installed you’re never typing a long message again.
Do this first. Install it, open Claude or ChatGPT, hold the key, and say the whole situation out loud: what you’re working on, who it’s for, what you want, what to avoid. Let go and watch it land as clean text. Send it. Compare that answer to the one you’d have gotten from a one-line typed prompt.
Try it here: wisprflow.ai
Claude
What it is. The same Claude from the writing section, doing a different job.
Why it’s here twice. When the ideas are out and you need them sorted, Claude is the best at structure. “Here are 30 rough ideas, group them, rank them, tell me which three are worth doing and why.” It gives you an honest ranking instead of a cheerful list.
Best for. Turning a brain dump into a plan. Paste in everything from the voice session and ask for the three that matter.
Do this first. Paste the transcript from your voice session and say: “Group these into themes. Rank the themes by which one I could act on this week with the least effort and the most payoff. For the top one, give me the first three steps.” One prompt turns a ramble into a plan.
Try it here: claude.ai
Miro AI
What it is. A visual whiteboard with AI built in. Sticky notes, clusters, diagrams, and an assistant that can group and summarize what’s on the board.
Why it’s on my list. Some ideas need to be moved around, not listed. If you think in maps and arrows, a chat window is the wrong shape. Miro lets you drag thoughts next to each other and then have the AI name the clusters.
Best for. Planning something with a lot of moving parts: a launch, a content month, a workflow with five steps and three people. Also the right tool when you’re brainstorming with someone else, not alone.
Do this first. Start a blank board, add a sticky note for every task in a project you’re dreading (a launch, a move, a busy month), one per note, no order. Then select all and ask Miro’s AI to group and name the clusters. The groups are your plan, and you’ll see the whole project in one glance instead of a list in your head.
Try it here: miro.com




Four slots because “make me an image” is four different requests: fix a photo I have, make one from words, make a set that matches, and put readable text on it.
Nano Banana Pro
What it is. Google’s image model, the one behind Gemini’s image editing. Its strength is working on a photo you already have.
Why it’s on my list. Every other image tool wants to generate from scratch. Nano Banana Pro is the best I’ve used at taking a real photo and changing one thing (the background, the lighting, the outfit, the object on the table) while leaving your actual face and product alone. For a business, that’s the more useful skill.
Best for. Product photos on a new background, a headshot with better light, “remove the person walking behind me,” a mockup of your product somewhere it’s never been.
Do this first. Upload a real photo of your product or yourself and say: “Keep everything about the subject exactly the same. Change only the background to a clean, bright studio setting with soft light.” Then try: “Same photo, now on a wooden table by a window.” Two backgrounds, one photo, no reshoot.
Try it here: gemini.google.com
ChatGPT Image 2
What it is. ChatGPT’s image generator. Describe it, get it, then keep talking to refine it in the same chat.
Why it’s on my list. It’s the best at understanding a plain-English description and getting close on the first try, and the conversation makes fixing it easy. “Make the sky warmer, move the logo left.” No settings to learn.
Best for. A one-off image from a description: an illustration for a post, a scene, a concept, a quick visual for a slide. When you have no source photo, start here.
Do this first. Describe an image you’d use this week, with these five things in the sentence: subject, setting, lighting, mood, and what it’s for. “A flat-lay of a coffee mug and notebook on a marble counter, morning light, calm, for an Instagram post about slowing down.” Then refine in the same chat: “warmer, less clutter.”
Try it here: chatgpt.com
Leonardo.ai
What it is. A full creative studio that generates images and videos and gives you access to all the latest models in one place, with style presets and controls for keeping a consistent look across everything you make.
Why it’s on my list. I use Leonardo a lot, and the reason is consistency. Making one good image is easy now. Making twenty that look like they belong to the same brand is hard, and Leonardo is built for exactly that: pick a style, lock it, generate the whole set.
Best for. A batch that has to match: a set of thumbnails, product mockups in one style, illustrations for a whole guide. If you’re making more than three images for the same project, this is the one.
Do this first. Pick one style preset you like and generate three images of completely different subjects with it. If all three look like they belong together, you’ve found your brand’s look. Save that preset and use it for every image in the same project so nothing looks like it came from somewhere else.
Try it here: leonardo.ai
Ideogram
What it is. An image generator whose specialty is text inside the image. Signs, labels, posters, packaging, a quote on a background, all with the words spelled right.
Why it’s on my list. For years AI images had one famous failure: the text came out as gibberish. Ideogram fixed that first and it’s still the most reliable at it. If the image needs words on it, this is the tool.
Best for. Anything with lettering: a poster, a sign for a shop, a label, a quote card, a book cover mockup, a thumbnail with a title baked in.
Do this first. Ask for one image with words on it that you’d actually use: “A clean square graphic that says ‘[your offer, 4 to 6 words]’ in bold modern lettering, [your two brand colors], lots of empty space.” Check the spelling. That’s the one thing it does that the others still get wrong.
Try it here: ideogram.ai
A tool on its own saves you minutes. Tools chained together save you a job. These are the three chains I’d set up first, each one built from apps above.
A month of posts from one recording
Record yourself talking for twenty minutes about the questions your customers ask most. Phone propped on a shelf is fine.
- OpusClip. Upload the recording. Take the eight clips it scores highest.
- CapCut. Drop each clip on your saved template, run auto captions, fix the two misheard words, export. Eight videos.
- Claude. Paste the transcript and say: “Pull out twelve standalone tips from this, one sentence each, in my voice.” Those are your text posts.
- Canva. Put the twelve tips on your quote-card template. Resize to every platform.
- Pomelli. Give it your website and ask for ten more posts about the same topic to fill the gaps.
Thirty posts from one afternoon of talking. The recording is the only part that needs you.
Sign a contract without paying for an hour you don’t have
Someone sends you a lease, a vendor agreement, a client contract.
- NotebookLM. Upload it. Ask: “What in here can cost me money, and what would I have to do to avoid it? Point me to each section.” Then: “What’s missing that I’d expect in a contract like this?”
- Perplexity. Take the two scariest clauses and ask: “Is this normal in a [type] contract? What do people usually negotiate on this? Sources.”
- Claude. Paste both answers and say: “Write me a short, friendly email asking for the three changes that matter most, in my voice, no legal jargon.”
You still might want a lawyer for the big ones. But you’ll walk in knowing exactly which three questions to ask, and you’ll skip the hour on the ones that don’t need it.
A product shoot without a photographer
You have one decent photo of what you sell and you need it to look like ten.
- Nano Banana Pro. Upload the photo. Ask for the same product on five different backgrounds: white studio, wooden table, outdoor light, lifestyle scene, your brand color. Five photos.
- Ideogram. Make a matching graphic that says your offer in words, same colors.
- Leonardo.ai. Lock one style preset and generate three supporting images (a texture, a scene, a detail shot) so the whole set looks like one shoot.
- Canva. Lay all nine out on your templates.
That’s a product page and a week of posts from a photo you took on your phone.
You don’t have to learn all seventeen. Nobody does. Pick the two that match where you are right now, get comfortable with those, and the rest will be there when you want them. Here’s where I’d point you.
If you’re brand new to AI and a little nervous. Claude and ChatGPT Voice. Do the voice prompt from Part 4 first, because talking to it is the fastest way to stop being nervous about it. Then do the “describe my voice” prompt in Claude. Two conversations and you’ll get it.
If you run a business by yourself. NotebookLM and Opal. The contract workflow above will save you real money the first time you use it. Opal turns the customer reply you rewrite every week into something you never rewrite again.
If you own a small business with customers walking in. Pomelli and Nano Banana Pro. Your website already says what you do. Pomelli turns it into posts, Nano Banana Pro turns the photos you already have into the ones you needed. You can be posting this week without hiring anyone.
If you already make content and you’re tired. OpusClip and Wispr Flow. One long recording becomes a month, and every prompt you’d have typed, you say instead.
Add the others as you go. The day a contract lands in your inbox, you’ll remember NotebookLM. The tenth time you do the same task the same way, you’ll remember Opal. When twenty images need to match, you’ll remember Leonardo. They’ll be right here.
You're fighting the tool. Retyping the same instruction, fixing the same mistake, exporting and re-importing between apps. Every app on this list does one job well. When you're fighting it, you're usually asking it to do a different app's job. Check the table at the top and switch.
Not before you've hit a wall on the free version. Every one of these has a free way in, and for most people it's enough for a month. Pay for the one you open every day and run out of, not the one that sounds impressive. Prices change too often to print here, so check the app's own site.
Want more like this?
One simple AI win in your inbox every Sunday. No jargon, no overwhelm, unsubscribe whenever you like.
Want more like this?
Free AI guides added regularly. No jargon, no overwhelm.
kellystrattonai.com
