Better answers from complex documents
Helping users find the right answer – and fast – is the name of the game. This month, we rolled out two big improvements that significantly upgrade how our AI agents handle technical and complex documentation.
1. Smart document breakdown
Now automatically detects and flags “hard” parts of documents (think: dense tables, layered diagrams, scanned manuals). When a user asks a question that likely requires information from one of these tricky sections, we route it to a more powerful, reasoning-heavy AI - ensuring more accurate, thoughtful responses where it matters most.
2. Dual document indexing
Adds a second layer to our document understanding. It pairs semantic indexing (when AI searches for meaning in a document) with keyword indexing (when it’s looking for exact matches that don’t “mean” anything words-wise).
As a result, questions like “loose panel near exhaust” or “model 328-XL-A22” are both handled with precision, regardless of how complex the document is.
Smarter and faster voice conversations
Our Voice Assist experience just got another serious upgrade - on speech recognition and conversation handling fronts.
It now uses a faster, more capable AI conversation handling model that allows 3s or less response time even for complex, nuanced questions - key for keeping phone conversations fluid and human-like. This already resulted in higher resolution rates and lower escalations across our customer base.
We’ve also introduced a new model for turn-taking in voice. It’s more sensitive to human speaking patterns - whether users speak quickly, slowly, or with hesitation, our assistant adapts smoothly. This dramatically improves the fluency of conversations and reduces interruptions or awkward silences.
Automatic language detection
We added automated language detection for both generative and hybrid digital assistants.
If your assistant is set to French but the customer types their question in another language - like English, Spanish, or German, - it will automatically detect the language and respond accordingly. This makes Mavenoid an even stronger choice for multi-region, multi-language brands. It’s seamless, it’s smart, and it just works.
Reminder: Mavenoid is localized in 60+ languages for digital and 30+ languages for voice AI agents.
See it all in action: live in Amsterdam
We’re excited to be hosting the Product Support Summit in Amsterdam this March.
It’ll be a deep dive into multimodal support - voice, text, images, actions - all in one seamless conversation. We introduced multimodal in the November 2025 update and now we’re going further, with live demos, real use cases and customer showcases.
Register here if you’re interested in how AI is transforming support - from product selection and pre-purchase help to troubleshooting and upgrades.
We’d love to see you there!
