To read an AI interview answer without losing eye contact, keep the assistant window narrow and directly under your webcam, make the text larger in Options, and read the key points, which come first, in one glance before you look back at the lens and speak. Use Ctrl+Shift+/ to scroll and Ctrl+Shift+H to show or hide the window so your hands never leave the keyboard. CoPilot Interview shows answers as on-screen text only; there is no audio playback, so your reading technique is what keeps the conversation natural. This page covers placement, text, timing and the habits that make it work.
Why eye contact on video is about geometry
On a video call, eye contact means looking at the camera lens, not at the interviewer's face on your screen. Most people already fail this a little, because the face is in the middle of the screen and the lens is above it. Anything you read is a further pull away from the lens, and the farther the text is from the lens, the more your eyes visibly travel.
So the rule is simple: the text should be as close to the lens as you can get it. For a laptop with the camera centered above the screen, that is a window pinned to the top center. For an external webcam clipped to a monitor, it is the strip right below where the camera sits. Small eye movements within a couple of inches of the lens read as thinking; a glance to a second monitor or the bottom corner reads as reading. The page on where to position an AI interview assistant window goes into screen layouts in detail.
Set the window up before the call
- Make it narrow and tall, not wide. A narrow column under the camera keeps every line of text within a small angle of the lens. Wide windows push the end of each line sideways, which is the eye travel you are trying to avoid.
- Enlarge the answer text in Options. Bigger text means fewer words per glance but far less squinting, and squinting is more visible than glancing. Set the size so that two or three key points fit in the window without scrolling.
- Place the meeting app's video below the assistant. Your eye path is then lens, text, interviewer's face, all in one vertical line. Nothing is off to the side.
- Turn on click-safe (Ctrl+Shift+F). The window then never takes focus from the meeting app, so you can leave the cursor in the call and use hotkeys for everything.
- Test with the camera preview on. Record ten seconds of yourself reading a sample answer in your meeting app's preview, and watch where your eyes go. Adjust until the movement is small.
Read the structure, not the sentences
The answer a real-time interview copilot produces arrives in about 2 seconds after the question ends, key points first and then detail. That ordering is designed for exactly this problem. You do not need to read the detail before you start talking; you need the two or three key points, which fit in a glance. Then you look at the lens and speak the first point in your own words. While you are speaking, you already know what the second point is. If you need the detail for the third, a quick glance brings it back.
If answers are consistently longer than you can take in, use the shorter answer or bullet points buttons, or set that preference and leave it. Bullet points are easier to scan from a few feet away than prose. For a coding round, where the answer may be a block of code you need to type, the 📝 Note option adds a 2-4 line approach and complexity comment at the top, which is the part you read first and say aloud while you begin typing.
| Habit | How it looks to the interviewer | What to do instead |
|---|---|---|
| Reading a full paragraph before answering | A long silence, then a fluent answer that sounds recited | Read the first key point only, start speaking, glance for the next |
| Window at the bottom or on a side monitor | Eyes visibly dropping or turning away | Narrow window directly under the camera |
| Small text | Squinting and leaning in | Enlarge the text in Options; fewer words per glance |
| Clicking to scroll or switch windows | Hand and attention leaving the conversation | Ctrl+Shift+/ to scroll, Ctrl+Shift+H to show or hide |
| Speaking the answer word for word | Phrasing that does not sound like you | Say the point in your own words; use the structure, not the script |
Glance, then speak: the timing
Here is the rhythm that works, written out as a sequence (example we wrote):
- The interviewer finishes the question. You nod and say "Sure" or repeat the question briefly. That takes about 2 seconds, which is the time the answer needs to appear.
- Glance at the window. Read key point one, and register that there are two more. Half a second.
- Look at the lens. Deliver point one in your own words, with your own example.
- Glance. Point two. Lens. Deliver.
- If the answer runs below the window, Ctrl+Shift+/ moves it while you are still talking. No click, no cursor movement.
Repeating the question is doing two jobs: it confirms you understood, and it covers the moment the answer is still arriving. On a slow connection, where the answer may take longer than about 2 seconds, a slightly longer restatement does the same work.
No audio playback, by design
Some people ask whether the answer can be spoken to them so they never have to look at the screen. CoPilot Interview does not do that. Answers are on-screen text, and the only thing the app reads aloud is the mock interviewer's questions during practice. The reason is practical: listening to one voice while another person is talking to you is a worse experience than a glance, and it produces the recited-sounding answers interviewers dislike. The on-screen structure, read in glances and spoken in your own words, is the method the app is built around.
Where this does not work well
The honest limitation: a desktop AI interview assistant cannot help you read when the answer is long and the question is one you know nothing about. If you do not understand point one, no glance is short enough, and you will end up reading. The fix is upstream. Paste your résumé and the job description so the answers are grounded in your real experience, which makes the key points things you recognize instantly. And do enough practice with the AI mock interview that most questions already have a shape in your head, so the window is confirming rather than teaching.
The other limit is multi-monitor setups: a second screen is a long way from the lens, and nothing about placement fixes that. If you have two monitors, put the call and the assistant on the one with the camera.
Choosing a tool with reading in mind
When you compare AI interview assistants, check three things that affect reading: whether the answer puts key points first, whether you can enlarge the text, and whether there is a scroll hotkey. Those three decide whether a glance is enough. For the mechanics of what the window shows and when, see how real-time interview assistants work, and for more on speaking the answer naturally, how to use AI interview answers in your own words.
Follow the rules of the employer or assessment you're in — many people use it only to practice, which works on the free plan.
Build the glance-then-speak habit
Run a free AI mock interview with your camera preview open and the window in its interview position. The AI interviewer reads each question aloud and scores your spoken answer.
Start a free mock interviewFAQ
Where should the AI interview assistant window be on my screen?
As close to the camera lens as possible, usually a narrow column at the top center directly under the webcam. Put the meeting app's video below it so your eyes travel in one short vertical line between lens, text and the interviewer's face.
Can the app read the answer aloud to me?
No. Answers are shown as on-screen text only. The only thing the app reads aloud is the mock interviewer's questions during practice. Reading key points in a glance and speaking them in your own words is the intended method.
How do I scroll a long answer without clicking?
Press Ctrl+Shift+/ on Windows or ⌘+Shift+/ on Mac. It scrolls the answer while your cursor stays in the meeting app. Ctrl+Shift+H shows or hides the window if you want the screen clear for a moment.
How do I make the answer text bigger?
The answer text size can be enlarged in Options. Set it so two or three key points fit in the window at once; bigger text with fewer words per glance is less visible than squinting at small text.
What if the answer arrives after I have already started talking?
That is fine and often ideal. Repeat the question or say a short opener while the answer appears, which takes about 2 seconds. Start with what you know, then glance for the structure to carry the rest of the answer.
Does a second monitor help with eye contact?
Usually it hurts. A second screen is far from the lens, so glancing at it is a visible head turn. Keep the call and the assistant window on the monitor with the camera.