AI-powered Product Assistant is now available as a full-screen experience

Hardware companies love integrating Mavenoid Assistant widgets across their websites, mobile apps, and other channels. We've received numerous requests to make this a full-screen format so it can be shared via a link or embedded into other tools – and now it's possible! Why is this exciting?

1. You can create multiple support sites where customers can effortlessly find the answers they need and resolve issues in just a few clicks. These sites are custom-branded, AI-guided, and dynamic, just like our Assistant widgets. They enable you to automatically address between 45% and 80% of all incoming support requests.
2. You can proactively capture customer support queries in search engines and guide your customers to dynamic help pages that lead them to a resolution for the problems they've been searching for. This enhancement was made possible thanks to the redesign of our Assistant Webpages solution, which now supports the new full-screen mode.

The full-screen experience retains all the functionalities of the Assistant widget. It offers AI-guided self-service, searches for answers in product manuals, seamlessly escalates to a live agent or generates a ticket in your CRM, and captures detailed feedback about customers' interactions with your support. Additionally, our in-depth analytics provide valuable insights into customer journeys, including topics of interest, resolution rates, escalation rates, and satisfaction levels.Here’s a sneak peek at the new full-screen experience:

And this is just the beginning! Soon, we will announce an even bigger update in this direction, so please stay tuned.

Available for: beta customers.

Use the power of Generative AI  to go live quickly and broaden your product coverage

One major challenge in building an effective self-service assistant is that it requires time and expertise, especially when covering hundreds of products. This typically involves reading numerous manuals, reviewing support tickets, scraping online forums, and organizing that data into categories like errors, symptoms, solutions, and more.

To streamline this process for our customers, Mavenoid began experimenting with Generative AI in 2021 to create draft self-service flows. We were among the earliest adopters of the then-novel OpenAI’s GPT-3, utilizing it for the initial version of AI Auto-Generation solution that allowed modelers to create first drafts of self-service flows with AI's help.

Since that time, there has been significant progress in the AI field, including the recent release of GPT-4, which many of you know as their best friend ChatGPT ;) . We have harnessed this new technology, combined with our proprietary AI model, to rebuild the AI Auto-Generation from the ground up, making it even more powerful.

Modelers can now easily upload product manuals, highlight popular issues to prioritize, and indicate the types of content and escalation scenarios they’d like to see in the flow. Within ~20-30 minutes, they will receive a draft flow, saving massive amounts of time. This can be fine-tuned and pushed live on the same day to begin addressing support inquiries.Want to see the demo? Here it is!

Available for: all customers.

Get insights from your support data with Analytics API

Companies that want to process and analyse their support data in a centralised space, e.g. data warehouse or business intelligence tool, can now get access to their Mavenoid data using the new Analytics API.This API allows access to both self-service and live support conversations, as well as agent availability in live support. We expect it to be useful for:

1. Analysis – analyse data using your in-house business intelligence platforms, such as Tableau or Power BI.
2. Compliance – archive conversations securely within your system.
3. Accessibility – promptly provide conversation details upon customer requests.

We will be rolling out Analytics API gradually, starting with a few beta customers in September. If you’re interested in using it, please reach out to your Mavenoid customer success representative.

Available for: beta customers.

Honorable mentions

Apply translation automatically to all matching keys

When approving a manually-entered translation, you can now automatically apply it for the same language in all other keys with the same base text. This saves translators a time on working with duplicate translations.

Unpublished flow changes are much faster to review

We've added a new filter to the canvas - "Since last publish" - that shows all the nodes added or edited since the flow was last published. You can use it on its own or in conjunction with other filters and search. This improvement makes the review process much faster, especially when multiple people collaborate on the same flow.

Did you know that...

Mavenoid can efficiently manage parts sales at scale

‍Check this example from Husqvarna Construction: our assistant dynamically draws content and data from various systems (ERP, CMS, and more) to recommend the most suitable spare part tailored to a customer's specific needs. Which leads to increased sales and happier customers!

Fixes and polishes

Our team has pushed a lot of fixes, mainly focusing on the quality of translation management experience. Here’s the full list:

• If you scrolled extremely fast and then stopped, in rare cases, the TMS (Translation Management System) could get stuck in a loading state until you scrolled again. This doesn't happen anymore.
• We have improve scrolling and pagination in TMS. It's now smoother and the performance is substantially better.
• There was a glitch when non-translatable glossary terms translated to <x-0/>, <x-1/>. We have fixed that.
• In some cases, TMS failed to pick up a glossary term and use it for translation. It was fixed.
• Long texts in the TMS used to break the layout, making it hard to use. We’ve polished that and now it works nicely.
• Import to CSV sometimes missed translations and exported too many keys - but now it works as expected.
• TMS search used to be a bit vague, sometimes disregarding the filters. We’ve fixed it to only search the source text and translations that match the active filter.

Apart from the translation management improvements, we have slashed these bugs:

• AI Retrieval sometimes displayed odd matches that were not relevant to what customer is searching for. Now fixed!
• When admin changed settings on the Appearance page, they would not persist. Now they do.
• Flow Builder canvas was getting a bit slow in loading for super-large flows - so we made it much faster.
• Opening step by step guides on iOS used to open the select dropdown by default. It no longer does that.
• An error was thrown when publishing some flows - we’ve eliminated the root cause so you won’t see it anymore.

We welcome your ideas on how to improve Mavenoid!

Use this form to submit your pain points, ideas, and feedback straight to our product team. As always, we are all ears and want to hear your thoughts! Thank you and stay tuned for more updates ;)