Generative AI is no longer limited to writing text. It now generates stunning visuals, edits videos on the fly, subtitles, translates, animates… In short: it is profoundly transforming our relationship with images. For this first meeting AI Common ThreadThe We Are COM Club immerses you in the world of AI applied to the photo and videowith a clear objective: to understand, test, and adopt these new tools to put them at the service of our communications.
A look back at the key insights, the platforms tested, the methods to remember… And the questions to ask yourself to maintain creative control.
Watch the replay 📹 Download the slidesAI News (Q2 2025)
Communications professionals are embracing AI (but not without caution)
According to the We Are COM survey (April 2025), 70% of communications professionals say they are familiar with AI, and 40% use it weekly. The most common uses? Content creation (92%), strategy (33%), social media (23%), automation, and monitoring (21% each).
But AI in communication remains a complex issue, both for professionals and for their companies!
AI agents: the quiet… but radical revolution
This is the turning point of spring 2025: AI is no longer just an assistant, it's an autonomous agent, capable of acting on your behalf. An AI agent can make decisions, chain actions together, and work in the background. Google, Microsoft, OpenAI, Cisco, Monica AI: they are all launching their own agents for sales, customer support, or software development. We're moving from a tool to a virtual teammate.
Google is transforming the web… and search
With Gemini 2.5, the Chrome browser becomes intelligent: automatic summaries of web pages, contextual clarifications, near-instant voice translation in Google Meet and even AI summaries of YouTube videos.
Google is also launching Veo 3 (video generation with soundtrack) and Imagen 4 (better typography + 2K rendering), while testing the FLOW platform, an AI toolkit for creating on the fly.
And find even more news in the presentation support !
4 AI photo/video use cases 100% COM
Create an infographic
Do you need to synthesize complex information (results, processes, reports, key figures, action plans, etc.) into a visual, readable, and engaging format? AI can transform a complex diagram into an effective illustration. But with a few precautions!
- To be effective, your prompt must include
- A clear context (explain the project, the audience, the objective)
- A specific request
- A specific style (illustration, colors, etc.)
- Constraints (format, key steps, keywords, etc.)
The We Are COM trick : ask the AI "what it needs to create the best possible infographic before asking it to generate it.
Example of prompt
To correct these errors, you can use the “selection” feature of ChatGTP Pro / Plus or, more efficiently, download the image and edit it with Photoshop or Canva.
Another technique is to limit the text in your request:
Create a personalized and consistent image library
Do you want unique visuals to illustrate editorial content or internal campaigns without relying on a generic stock photo library? Generative AI image generation is increasingly enabling the creation of a style based on a prompt or existing visuals. It's becoming an attractive alternative to traditional stock photo libraries.
On CHATGPT, you can attach existing illustrations and request the creation of a new visual, in the same style as the previous ones.
The We Are COM trick : to have more control over style, colors, effects, it is preferable to have the style analyzed by AI, make it into a prompt, and then use the prompt to make new visuals.
results:
Speaking on RECRAFTThere is an "IMAGE SET" feature that allows you to generate six visuals in the same style at the same time. You can test this feature by creating a free account.
The We Are COM trick You can create your own style within the tool, based on 1 to 5 existing photos. In this "source," you can assign a weight to each image.
Result
Result :
Create a b-roll video
Do you need a few seconds of video to illustrate a concept, supplement a video, or open a presentation? We're not talking about Hollywood movies here, but rather illustrative videos, from 5 to 20 seconds, to enhance an existing video. Generative AI like SORA (from OpenAI, included with ChatGPT subscriptions), Kling AI (China), or LUMA AI (USA) allows you to generate original clips for your needs.
With AI-generated video content, your prompt is most often text-based. There are some things to keep in mind:
- Describe a scene, not just an idea: think like a director: where does the scene take place, who is there, what is happening, what is being seen? Avoid abstract formulations and favour a precise and visual description.
- Be precise about the movement
- Specify the type of camera movement desired (pan, tracking shot, fixed shot, etc.). This allows for better control of the visual storytelling.
- Describe the character visually: approximate age, clothing, hairstyle, accessories… The more precise you are, the more the result will meet your expectations.
- Pay attention to the setting and atmosphere: describe the location, the lighting, the style of furniture, the background.
- Using the language of cinema
- Specify the camera's point of view (back view, high angle, low angle, shoulder-mounted camera)
- Avoid abstract or symbolic requests
- List what you don't want: elements to exclude (special effects, text on screen, people in the background, logos, etc.) to avoid surprises.
The We Are COM trick : again, you can ask an AI (ChatGPT, Mistral, Gemini, COPILOT…) to generate the prompt for you.
Example of prompt (generated by ChatGPT based on “a video of a woman working at her desk”): A realistic video of a professional woman in her early 30s working at her desk in a modern office. She is focused, typing on a laptop, occasionally looking at a second screen. The workspace is clean and well-lit, with soft natural light coming from a window on the side. Subtle office background movement. The woman is casually dressed (blazer over a T-shirt), wearing glasses, with short curly hair tied back. Cinematic framing with shallow depth of field, smooth camera movement (pan or slight dolly in), no text, no direct eye contact with camera.
Create a "face cam" video, without filming (avatar).
Need to quickly create a 2-minute video for an advertisement but don't have a studio, professional equipment, or even a speaker for recording? Today, it's possible to bring a message to life in video using avatars.
You can then select one of the 4 profiles offered.
Then select (or create, from your voice or a prompt!) the voice of your avatar
It's ready! All you have to do is write the text that your avatar will say in the video 🙂
- create your own avatar from a photo
- create your own avatar from a video
- use an avatar available from an AI catalogue (based on a “real” person or generated by AI)
📹 Replay of the AI and S seriespresentation support
