Product Blog Post #1 – What’s new with Textkernel

Hi there!

Welcome to Textkernel’s first ever product blog post. This is the outcome of one of our innovation week projects. We’ve been organising innovation weeks for the past six years. During this week, the entire company works together in teams getting their own innovative ideas into action.

The winning project – Customer communications – focussed on getting product updates to our customer and user base. They created in-product notifications for our TK Portal (we still need a better name for this) so users of the products can actually see what has been enhanced. We are currently building this out further to our other products.

Aside from the big feature releases, we haven’t always been updating you with all our progress. This does not mean we did not do any work! We’re often busy with things behind the scene. Like speed improvements, bug fixes, backend optimisations and AI research. In order to be more transparent, the team also created this product blog. Here you can read what we are working on, what improvements have been made and what you can expect in the (near) future.

Coming up: exciting big features

We have several releases coming up in the next months. Think about Deep Learning for Job parsing, Semantic search in Simplified Chinese and Japanese resume parsing. We’re expecting this end of Q3, early Q4, so keep an eye out on the blog for more details. Contact us if you want us to personally update you on the status.

What updates have we done last month(s)?

Improvements for our Bullhorn integration
We recently started getting reports regarding the time it takes to import a candidate from one of the external sources we offer into Bullhorn. Users were waiting up to a minute for this process to finish and their day-to-day work was severely impacted. Given this situation, we got busy and started by identifying quick wins. As a result, we completely rebuilt the way in which we call the Bullhorn API and we can now offer:

  • Faster import of candidates from external sources, email parsing or manual upload into Bullhorn.
  • Better deduplication and automatic update of existing candidates.

Users are now be able to insert candidates into Bullhorn and return success to the user once the basic candidate record has been inserted. This takes approx 2-3 seconds for the basic information, while the rest of the update happens in the background. This allows the user to continue with their work right away!

In addition, we made improvements in our deduplication algorithm and are now able to automatically update a candidate when there is only one potential duplicate found in the Bullhorn database. We hope this will help our users be more effective while making sure that their candidate database remains free from duplicates.

We are still busy with rolling these enhancements out to all our customers. An update will follow once all customers are able to enjoy from these improvements.

User activity reporting
You may have already seen our announcement: we released an update for our reporting functionality in Search!. This allows you to dive deeper into your user’s activities, providing valuable insights into user adoption and monitoring roll-out of strategic plans. For detailed information, read the release article.

Talk to us

We would also like to get the conversation going. We are working on some features for our Search! product and would like to get your input. Please fill out our poll:

[poll id=”3″]

If you have additional feedback, you can leave this with us in the comment section below.

That’s it for this month’s blog post! I’m handing the stick over to my colleague Andrea who will update you next month.


P.S. Do let us know what you think of this post; what you’re missing and would like to see next time!