AI Search

Historically, finding the right image often means remembering the exact keyword, tag, or filename it was saved under — and if you don't remember it precisely, it can be hard to find what you're looking for even when it's sitting right there in your archive. 

AI Search solves this by letting you search using natural, descriptive language instead of exact keywords, making it much easier to discover visual assets in your DAM based on what they actually show.

Examples:

• Searching "person smiling in an office" can surface relevant lifestyle shots even if they were never tagged that way
• Searching "product shot on a white background" can return matching images regardless of filename

Who can use it?

AI Search is available on request as part of our Beta programme. Please contact your Customer Success Manager if you'd like your organisation to take part.

Once enabled by your Customer Success Manager:

• AI Search is turned on for all users in your organisation — there's no separate permission or setup required at a user level
• AI Search applies across all searches in both Point and Place
• AI Search currently works with images only; other asset types are not yet supported
• As a Beta feature, functionality may evolve based on feedback from participating organizations

How AI Search works

AI Search is designed to make content discovery easier — helping you find visual assets based on what's actually in them, not just how they were tagged or named. It is available in Point, Place and Asset Selectors. It's especially useful when you're browsing for inspiration rather than retrieving one specific known asset.

• Search using natural, descriptive language rather than exact keywords or tags
• Combine AI Search with your usual filters (folder, date, metadata, etc.) to refine results further — it works alongside your existing search tools rather than replacing them
• Easily switch between AI Search and your traditional Exact search depending on what you need 
• AI search shows assets that are similar to your search and ranks them from highest to lowest match. Much like with a Google search, results will become less relevant the further down the page you go. 

Enabling AI Search 

AI Search is configured in two steps: first at company level, then per archive. Global Settings permission is required for the company-level step; Archive Settings permission is required for the archive-level step.

Step 1: Enable AI Search for your company

1. Navigate to Place Settings

   

2. Select Global Settings
3. Navigate to AI Features
4. Enable Visual Search

This turns the feature on across your organisation, but it won't be active in any archive until you complete step 2.

Step 2: Enable AI Search for each archive

1. Navigate to Place Settings
2. Select Archive Settings for the archive you want to enable
3. Enable AI Search

Note: AI Search works on image assets. If an archive contains mostly non-image files, enabling AI Search there won't add much value — you may want to limit it to image-heavy archives.

How this affects Search Views in Point

Once enabled on an archive, AI Search automatically applies to any Point Search View connected to that archive — no additional configuration is needed there.

• If a Search View is connected to a single archive with AI Search enabled, AI Search is available on that Search View.
• If a Search View is connected to multiple archives, AI Search only applies to the assets in the archives where it's enabled — assets in archives without it enabled won't be included in AI Search results.

Turning AI Search on or off

• Users can toggle AI Search on or off directly within the search view 
• Your last selection is remembered via cookies, so your preference is saved between sessions on the same browser

Getting the best results

AI Search works best with clear, descriptive language about what's actually in the image.

What AI Search is great for:

• Describing a clear subject or scene — "dog on a beach", "sunset over water", "team meeting in an office"
• Combining a couple of concepts together — "snow-capped mountain", "dark moody portrait"
• Describing mood or tone alongside a subject — "warm, inviting kitchen", "bright, energetic street scene"
• Describing how an image is composed — "overhead shot of a table", "close-up of a product"

Where it's less reliable:

• Very short or generic words (e.g. a single three- or four-letter word) can sometimes pull in unrelated results — adding more descriptive detail helps narrow this down
• Very specific sub-categories (e.g. an exact breed, model, or variant) may return the broader category instead of the exact match
• A style word used on its own (e.g. just "glowing" or "close-up" with nothing else) can surface a wide mix of unrelated images that share that visual quality — pairing it with a subject gives much better results

Tip: If your first search doesn't return what you expected, try adding more description rather than less — the more visually specific your query, the better AI Search can narrow in on the right assets.

Frequently Asked Questions

Do I need any special access to use AI Search?
No. Once you have enabled it for your organisation, it's available to all users — there's no additional permission required.

Does AI Search work for all asset types?
Not yet. AI Search currently supports images only. Support for other asset types may be added as the Beta progresses.

Can I use AI Search alongside my regular filters?
Yes. AI Search works in conjunction with your existing filters, so you can search descriptively and then refine further using folder, date, or metadata filters as usual.

If I turn AI Search off, will it stay off next time I search?
Yes. Your on/off preference is saved via cookies in your browser, so it will persist across sessions until you change it again.