Copilot does not have to live in its own app. AI Central's walkthrough, Use Copilot in Any App on iOS and Android, moves the assistant down into the keyboard. Install Microsoft SwiftKey from the App Store or Google Play, and Copilot rides into every text field on the phone. Prompt where you already are, then paste the answer in place. Images work the same way, generated in one turn and corrected in the next.
Reviewed August 2026.
The move is from destination to input layer
AI Central's guide rests on one structural idea, and it is worth stating plainly. An assistant you have to visit is an assistant you use less.
Most phone AI is a destination. You leave what you are doing, open a chat app, type the request, wait, select the answer, copy it, switch back, paste, then repair the formatting. That round trip is a tax, and it is charged on every single request. On a laptop it is annoying. On a phone, where switching apps means losing your place in a thread and re-finding the cursor, it is often enough to stop you asking at all.
The keyboard removes the trip. A keyboard is a system-level input surface, which means it appears in every app that accepts typed text: messages, mail, notes, comment boxes, search fields, forms. Put the assistant there and you have not integrated it with a list of apps. You have made the list irrelevant.
That is why AI Central's route has exactly one dependency rather than a settings tour. The whole technique rests on Microsoft SwiftKey, listed in the walkthrough as available on the App Store and Google Play. The store listing reproduced alongside it shows a rating of 4.6 stars, more than a billion downloads, and a note that the app contains ads.
Three moves, and that is the whole thing
AI Central breaks the setup into three steps, and the sequence matters more than any individual tap. Install the keyboard, prompt from inside whatever app you are already in, then move the result into that app.
- Install Microsoft SwiftKey, the keyboard that carries Copilot, from the App Store or Google Play.
- Open any app you already use and start typing. AI Central's second step is to use Copilot anywhere, and its instruction is that you then type your prompt as usual, with no separate syntax to learn.
- Take the result and drop it where you were working. Copy and paste is the handoff, and the guide repeats that line at nearly every stage.
AI Central states the payoff of that third step in one sentence.
Get your result and use it in any app you like
Read that as a scope statement rather than a slogan. Because the assistant sits in the input layer, the destination is not a supported partner app, it is wherever your cursor happens to be.
What prompting from a keyboard actually looks like
The worked text example in AI Central's guide is deliberately ordinary. The request is for the best steps to grow on LinkedIn, and the answer comes back structured rather than as a wall of prose. The first item is to optimize your profile, supported by two concrete points: use a high-quality profile picture and banner, and write a unique headline that shows what you do.
Structure is the quiet reason this works on a phone. A numbered answer survives being pasted into a post box, a note or a message far better than a paragraph does, because you can delete the parts you disagree with without unpicking the sentences around them.
AI Central attaches the same instruction to that example that it attaches to the rest.
Just copy/paste the answer into your app!
The detail that carries the technique is where the answer lands. It arrives in a live, editable field inside the app you were already using, so the answer is a first draft you finish in place rather than a finished block you import.
Images, generated and then argued with
The third step in AI Central's walkthrough is creating and editing images, and it is the part most people underuse. The example prompt is a single line of intent, a request for an image of a cat in a car, and Copilot answers that it will try to create it.
Then comes the step that changes the output. AI Central's instruction is to create your image, then tap to edit, and its own example runs a second turn that pushes the picture away from a cartoon toward a realistic version. The image is treated as a conversation, not as a slot machine you pull once and accept.
Build for two turns by default. Ask for the subject and the scene first, look at what comes back, then spend the second turn on the one thing that is wrong: the style, the realism, the framing. That habit costs a few seconds and lifts the hit rate more than any amount of prompt engineering typed with thumbs. When it is right, the same copy and paste move sends it wherever you needed it.
The Android note worth reading before you go hunting
Availability is the one place where the experience diverges by platform, and AI Central flags it rather than glossing over it. The guide notes that the feature is already available on Android, and that you can reach it by tapping the Bing icon and then choosing Chat.
That is a useful thing to know before you start digging through keyboard settings looking for a Copilot switch that may be sitting behind an icon you already have.
How to make this stick
The technique fails for the same reason most phone workflows fail. People install it, use it twice, and drift back to the chat app out of habit. Four things fix that.
- Point it at the places where your typing volume already lives, replies, email answered while moving, comment boxes, rather than saving it for special occasions.
- Keep prompts short. You are typing with thumbs, so state the job and the format you want back, and skip the preamble.
- Treat every answer as a draft in an editable field. That editability is the real advantage over a chat window, so use it instead of accepting the first output.
- Give images two turns, one to generate and one to correct, before you decide the tool cannot do what you asked.
None of this asks you to change tools or move your work. That is precisely the argument AI Central is making with this setup: the fastest AI workflow is usually the one that does not require you to go anywhere.
Do I need the Copilot app to use Copilot inside other apps?
Not for the route AI Central lays out. The keyboard is the delivery mechanism, so installing Microsoft SwiftKey puts Copilot inside any app where you can type, and you prompt from there instead of opening a separate assistant.
Does this work on both iPhone and Android?
AI Central's walkthrough covers both, and the keyboard is listed as available on the App Store and Google Play. The guide adds that the feature is already available on Android, reachable by tapping the Bing icon and then Chat.
Can it create images, or only text?
Both. Creating and editing images is the third step in the guide, with a worked example that generates a picture from a one-line request and then refines it in a follow-up turn before the result is used elsewhere.
How does the answer get from Copilot into the app I am using?
By copy and paste, which AI Central repeats at every stage. The answer is produced in the keyboard layer while you sit in your app, so moving it across is a paste rather than an app switch, and you can edit it where it lands.