Ariella Zubarev
September 16, 2026 at 9:00:00 AM
After a summer break, Misha and Ivo return with a full episode on search. Ivo, a former LumApps search product owner, explains where the search engine came from and where it's likely heading. Misha shows admins how to tune search filters and how LumApps can now find content buried inside uploaded PDFs.
Back from summer, straight into search
We took about a month off over the summer. When we came back, we picked the topic people ask us about most: enterprise search.
This one is personal for Ivo. Before ADC, he was the product owner for search at LumApps, so he knows the engine better than almost anyone outside the search team.
Pro tip: you control the search filters
Many global admins don't know this, but you can configure the filters on the LumApps search page. In the back office, go to Search, and you'll see the default configuration. You can create your own from there.
A custom configuration lets you:
Choose your sources. LumApps content is the default, and Google Drive and SharePoint are both available natively.
Pick result types. Include or exclude contents, posts, users, spaces, communities, directory entries, videos, learning, and more.
Edit and reorder filters. Remove ones nobody uses, or move the important ones, like Author, to the top.
Before you add more filters, check your search analytics. Most people type a query, scan the results, and rarely touch the filters. Fewer, better-placed filters usually work better.
Ivo's tip for testing: don't edit your live configuration directly. Build a second one, activate it during off-hours in your time zones, test it, and switch back. As far as we know, you can't limit a test configuration to a specific group yet, so timing is your safety net.
Still excited about: Page Reader
Our Page Reader widget reads any LumApps page aloud, either the whole page or just a selection. It offers around 20 voices and works well for news articles and knowledge content. Because it doesn't use third-party AI, you don't need to go through an AI approval process to roll it out.
A short history of LumApps search
Ivo gave the backstory:
2019: LumApps rebuilt its search engine from scratch. It's more powerful than most people realize, and many of its capabilities have never been released as features.
Around 2021: The focus shifted to partners like Coveo, which was the best option at the time for connecting many content sources.
The Search SDK era: Customers could plug in any search engine, but it took a developer, and results appeared in a separate tab.
Today: AI, agents, and Ask AI are getting the spotlight, and search powers all of them behind the scenes.
The July 9 release notes showed the first signs of change, including semantic search for content admins.
Why native search has to catch up
Ivo used the Kano model from product management to explain the pressure on LumApps. Some features are basic expectations: if every product on the market can open a can and yours can't, customers won't even consider it. With so many search tools on the market, strong native search is now one of those basic expectations.
His predictions:
Synonyms will return.
Result clustering will arrive, something the team was already discussing in 2019.
Native connectors will let results from tools like Salesforce or ServiceNow appear in LumApps without the SDK or an extra tab.
Will a third-party engine come back as the answer? Ivo doesn't think so. Between Ask AI and the LumApps MCP, he expects LumApps to keep investing in its own engine.
Three ways LumApps search finds things
Exact match: finds the words you typed, as you typed them.
Semantic: looks at what you mean. Someone searching "time off" might really need your HR or expense tool.
Fuzzy: handles typos and small spelling variations.
The feature we use every day: searching inside documents
The July release also improved knowledge extraction for media documents, and this is the one Misha was most excited about.
Until now, connected sources like Drive or Box showed results in their own tab, so users had to click around to find files. The goal is one search feed with everything in it, and this update is a real step in that direction.
Upload a PDF to the LumApps media library, and search will index what's inside it, not just the title. In the demo, Misha searched for "Page Reader" and found a PDF bundle whose title had nothing to do with Page Reader. The match came from text inside the file.
There's one important limitation: you can't set visibility on media documents yet. Any file you upload will show up in search for everyone. For now, use this feature for organization-wide content, like benefits guides, general documentation, and design assets, and keep restricted files elsewhere.
Let's take a look at the replay: