AI Email Generation: FAQ

How long does generation take?

Most content generates within a few minutes. Short prompts and lighter assets usually finish quickly. Larger themes, long documents, or heavy attachments can take a bit more time because the AI needs to process more context before it writes.

If something feels slower than usual, it’s typically because the system is handling a bigger load of information behind the scenes. The tradeoff is worthwhile: more context usually means better accuracy and stronger output.

How does Brand Voice impact AI output?

Brand Voice shapes the tone, wording, and overall personality of everything the AI generates. It tells the model how you speak, what you value, and which terms or phrases matter most. When your Brand Voice and Brand Voice Rules are set up properly, the AI mirrors your style instead of producing something generic.

Think of it as the difference between “write something” and “write something the way we write.” It keeps your content consistent, recognizable, and on brand every time.

If you haven’t set Brand Voice up, but want to, check out this article.

Is my data safe?

Yes. Knak AI only uses the content you provide for your generation and does not use your inputs to train external models. For more info, check out this article.

What is Knak’s AI Policy?

To understand how Knak uses artificial intelligence while keeping your data secure and private, check out this article.

What does the setup look like?

Check out this article to help you get started.

What permissions do I need?

You’ll need the Email Creator role with the “create email with AI” permission enabled. Your Customer Success Manager can help you set up a new Email Creator role if needed. Once that permission is turned on, it automatically applies to all users with that role.

Can I use AI Email Generation without Brand Voice set up?

Yes. It still works. You’ll just need to guide tone in your prompt until your brand settings are in place. If you would like to set up your Brand Voice, check out this article.

How long does generation take?

Most content generates within a few minutes. Short prompts and lighter assets usually finish quickly. Larger themes, long documents, or heavy attachments can take a bit more time because the AI needs to process more context before it writes.

If something feels slower than usual, it’s typically because the system is handling a bigger load of information behind the scenes. The tradeoff is worthwhile: more context usually means better accuracy and stronger output.

Are there any Knak features that aren’t supported in AI Email Generation yet?

Landing pages aren’t supported today, and themes that use dynamic content will only generate with the default variant.

Will Knak AI automate my campaigns?

No. Personalization is definitely a dream that marketers have had for a long time. To be able to customize messages for each recipient. It’s definitely possible to achieve this now and made much simpler with AI.

However, there is a ton of risk associated to setting AI up to simply automate emails to go out to your audience. One wrong email could cause huge damages for a company. AI is not perfect.

We do not recommend fully automating your personalization.

Instead AI will help you make creating many content variations or tests much quicker and easier. Given how quickly AI can generate an entirely new email for you, image how fast it can generate a section of the email to be geared towards a buyer persona or industry (or any other dynamic content you can imagine).

What are some best practices for prompts?

Be clear about your goal, your audience, and must include content. For a deeper walkthrough, read our prompt best practices article.

What if my results feel generic?

Be more specific. Add your audience, tone, angle, and must-have messages. The clearer the direction, the stronger and less generic the output. For more guidance, check out our best practices.

Why did the AI include things I didn’t ask for?

When information is missing, AI fills in gaps. If something appears that you do not want, tighten your prompt or attach a source so there is less guessing.

Why did the AI miss something I expected it to know?

AI Email Generation doesn’t assume context. If a detail is important, mention it or include it in an attachment.

Can I further prompt once an email has been generated?

Yes. While you can’t edit the original prompt yet, you can refine the email directly in the builder. Update copy, swap images, adjust spacing, or rearrange modules. You can also use Knak AI Agents to rewrite or improve specific sections. Here’s a help article with more detail.

How can I review the prompt used?

To view your prompt, head into the Info tab in the builder and then navigate to Audit Trail.

Can I give a prompt in a different language?

Yes. Email Generation works with prompts written in other languages. If you want the output in a specific language, make sure to say so in your prompt. For the full list of languages supported in Knak Translations, check out this article.

How does Brand Voice impact AI output?

Brand Voice shapes the tone, wording, and overall personality of everything the AI generates. It tells the model how you speak, what you value, and which terms or phrases matter most. When your Brand Voice and Brand Voice Rules are set up properly, the AI mirrors your style instead of producing something generic.

Think of it as the difference between “write something” and “write something the way we write.” It keeps your content consistent, recognizable, and on brand every time.

If you haven’t set Brand Voice up, but want to, check out this article.

If I update my Brand Voice after generating an email, will the changes apply automatically?

Not automatically. You can generate a new email with the updated Brand Voice or use Knak AI rewrite to refresh the copy.

What if my regions or product lines have different writing styles?

Set up Brand Voice and Brand Voice Rules for each variation you need. This keeps tone, wording, and terminology consistent across all generated emails, no matter the region or product line. You can create separate profiles for each style and switch between them as needed. To set these up, check out this article.

Does the AI understand our product names?

If your product names are added to your Brand Voice Rules, the AI will use them correctly and consistently. If they are not included, make sure to spell them out in your prompt. Product names carry weight, so the best way to guarantee accuracy every time is to add them to your Brand Voice Rules.

Why doesn’t my email look how I intended?

Make sure you include a layout reference in your attachments. Without one, the AI will create a basic layout based on your prompt and your Brand kit (font, colors and default Brand Voice.)

Can I edit the generated email?

You are always in control. You can update the copy, swap images, adjust spacing, or rearrange modules right in the builder. You can also use Knak’s AI Agents to refine specific sections and tighten anything that needs polish.

Can I regenerate the email?

You can rewrite individual sections using Knak AI Agents in the builder or generate a brand new draft. Full email regeneration from the existing prompt is coming soon.

How can I give the AI stronger guardrails when working from a screenshot?

When you generate from a screenshot, you are giving the AI a lot of open space to interpret. If you need something more specific, use a theme, add images, or include a bit of direction in your prompt. These cues tighten the guardrails and keep the output closer to what you want.

How long does generation take?

Most content generates within a few minutes. Short prompts and lighter assets usually finish quickly. Larger themes, long documents, or heavy attachments can take a bit more time because the AI needs to process more context before it writes.

If something feels slower than usual, it’s typically because the system is handling a bigger load of information behind the scenes. The tradeoff is worthwhile: more context usually means better accuracy and stronger output.

Does Knak AI learn from our emails over time?

Knak AI does not learn from your individual emails or content over time. It responds to your prompt, follows your brand settings, and uses any attachments you provide to shape its output.

AI learning is on our roadmap. We are actively working on ways for Knak to adapt to your content patterns.

Can the AI handle technical or complex content?

Yes. It can work with technical, specialized, or complex subject matter as long as it has enough direction. If you need it to cover something nuanced like product architecture, integrations, compliance requirements, security concepts, or deep industry topics, give it a brief, a URL, or supporting material. Even a short summary of what matters most will sharpen the output.

Clear inputs lead to more accurate results, and there should always be a human in the loop to verify details and make sure the content lands correctly.

Is my data safe?

Yes. Knak AI only uses the content you provide for your specific generation. Your inputs are not used to train external models, shared with other customers, or fed into any public dataset.

We take data security seriously. Protecting your content and your brand is a core part of how Knak is built. If you want a deeper dive into how everything works behind the scenes. For more info, check out this article.

What is Knak’s AI Policy?

To understand how Knak uses artificial intelligence while keeping your data secure and private, check out this article.

What AI model are we on?

Knak uses Gemini 2.5 Pro as its primary AI model. It is fast, reliable, and strong at the kinds of work marketing teams need, from first drafts to smart suggestions that improve your assets. We also test new models on an ongoing basis and will bring them in when they deliver better results.

Can I bring my own AI tools to Knak?

Not yet. Knak does not integrate with external AI tools at the moment. You will still get the strongest results by creating inside Knak where your brand guardrails, themes, and voice rules are already built in.

Model choice is coming soon. You will be able to select the intelligence layer behind your experience whether that is a major foundation model, a specialized domain model, or your own internal LLM.

Do I have to use AI if I want to use Knak?

No - if you have security or company limitations that prevent you from using Knak, you can absolutely use our core creation platform without AI. We have the ability to turn AI off from your account so that no AI is used while you are using Knak. Should things change down the road, it is a simple change to enable it on your account if/when you are ready.

Can Knak AI generate images?

No. Knak AI doesn’t create images. Once brand visuals enter the picture, the risks climb fast. Off brand styles, inaccurate product shots, and copyright questions can all sneak in. Rather than gamble with your brand identity, we keep image generation out of the workflow.

You can still upload approved visuals or link to what you want the AI to use. That way, every image stays intentional and on brand.

Should I properly name the images I attach?

It helps. Clear file names like “speaker1,” “productA,” or “heroImage” make it easier for the AI to understand what’s what, especially when you attach multiple visuals. It won’t break anything if your images have messy names, but clean labels give you more predictable references in the output and make your own workflow smoother.

Can I generate an email with DAM assets referenced?

Not yet. Knak AI can use any images you upload or link to, but it cannot browse your DAM or pull assets from it during generation. It also can’t follow open-ended requests like “find an image of a beach” or “grab our latest product screenshots.” The AI only works with what you explicitly provide.

You can still upload visuals from your DAM manually or share public URLs if you want the AI to include or reference specific images. This gives you control over exactly what it uses while keeping your brand assets intact.

Will Knak be able to upload newly generated images back into an external DAM?

Image generation is not supported yet, and Knak does not push assets to external DAM platforms. Any images you add or update stay inside Knak unless you upload them manually.

Do I need to attach files?

No. A clear prompt is enough. Attachments improve accuracy when you want the email to follow specific messaging or structure.

What format of files are accepted to upload?

Email Generation accepts PDF attachments. You can upload up to 20 files, with a maximum size of 5MB each.

What if my PDF is too large?

Compress it, split it, or share a URL instead. Smaller inputs work best.

Is the AI pulling content from the internet?

Only when you give it a URL. If you don’t provide one, it writes based on your prompt and attachments only.

Does file naming matter when I attach PDFs or assets

Yes. Clear and simple file names help the AI understand what each attachment is and reference it correctly. Use names that make the content obvious, such as SpeakerName Headshot or Template One. Avoid long or vague file names. The clearer the label, the more accurate the output.