Have you ever tried to explain something to someone and even though you didn’t phrase it perfectly, they still got what you meant? Someone “just getting it” is something we’re all used to experiencing, and end up relying on in our daily communication. We know humans will take our words and infer their meaning to understand what we’re getting at. This behavior is also something we tend to expect from our interactions with technology and search engines, but it’s not your everyday search engine that can do this well.
This expectation is why our Search and Discovery team at Guru has invested in adding new machine learning technology to augment our semantic search capabilities: to allow Guru’s search to consider keywords and meaning when pointing you to the most relevant search results. Our goal at Guru is to make it as easy as possible to access the information you need to do your job, and incorporating machine learning into our search offering is an exciting step forward.
What does adding machine learning to search mean?
By adding new machine learning capabilities to our search secret sauce, we can layer semantic search on top of our existing keyword search to leverage both the meaning of your search and the actual words of it. For example, if you’re looking for info on your company’s insurance policy for your furry friend and you call it “dog insurance” in your search input and your People Ops team calls it “pet insurance” in their documentation, Guru’s search will infer the meaning of your search term, and bump up the more relevant card towards the top of your search results list. If you’re interested in the technical details of how this technology works, check out our post about it in the Guru Community.
By getting to the meaning of what our users are looking for, we can materially improve search results for customers who have employees in different regions (we in Philly know that it’s “hoagie,” but still want our remote employees to be able to find the best “subs” in town when they visit) and have even set the stage for support across languages. Not only does this update help bridge subtle differences in vocabulary, but it enables Guru to learn from how your team searches for information, and use that context to make life a little easier for newer members of your team. Newer members who aren’t as familiar with your company jargon benefit from Guru having seen which Cards tend to get chosen for certain search terms and then adjusting the results list accordingly.
What else can Guru do with this new technology?
Tons! One of the reasons we’re so excited about this update is because it opens the door for updates beyond search. Our Search and Discovery team has expounded significant effort to get this technology in place, and the rest of our development teams and customers will reap the benefits:
With this new ability to relate content based on meaning, someday Guru could do things like form a picture of what content you tend to look at and provide suggestions for content that is similar in meaning. That’s right, personalized suggestions for you based on content you care about! Guru could also visually map how content is related across your company’s entire knowledge base and display it in different ways to different users.
We already referenced support for searching across languages earlier, but to expand on that: imagine being bilingual and searching for the Spanish “seguro para mi perro” in Guru (to keep with the dog insurance example) and getting back the most relevant “pet insurance” content, that happens to be in your second language, English! Because meaning is consistent across languages (for the most part), Guru could draw relationships between content in different languages—no Google Translate required.
This new ability to match based on meaning could help Guru’s AI Suggest Text feature become even better at bridging the gap between how your customers refer to things and how your company refers to them internally.
In the future, we can further train this model to include more industry-specific (or domain-specific) terminology to support the various verticals represented by you, our customers. We have doctors’ offices, mortgage brokers, teachers and professors, mattress makers, podcast curators, and more using Guru—each with its own unique jargon and acronyms. This technology could help us customize search in Guru to your company’s language quirks; no company is quite like your company!
Continued investments in making information accessible
Our mission at Guru is to make sure that the information our customers need to do their jobs is easily accessible, whether that’s via search, improved UX, or new functionality. We continue to prioritize effortless access to knowledge in everything we build to make Guru the best tool to help you do your best work. Keep an eye out for future updates to search and other areas of Guru where we’ll build upon this industry-leading machine learning technology.
Have you ever tried to explain something to someone and even though you didn’t phrase it perfectly, they still got what you meant? Someone “just getting it” is something we’re all used to experiencing, and end up relying on in our daily communication. We know humans will take our words and infer their meaning to understand what we’re getting at. This behavior is also something we tend to expect from our interactions with technology and search engines, but it’s not your everyday search engine that can do this well.
This expectation is why our Search and Discovery team at Guru has invested in adding new machine learning technology to augment our semantic search capabilities: to allow Guru’s search to consider keywords and meaning when pointing you to the most relevant search results. Our goal at Guru is to make it as easy as possible to access the information you need to do your job, and incorporating machine learning into our search offering is an exciting step forward.
What does adding machine learning to search mean?
By adding new machine learning capabilities to our search secret sauce, we can layer semantic search on top of our existing keyword search to leverage both the meaning of your search and the actual words of it. For example, if you’re looking for info on your company’s insurance policy for your furry friend and you call it “dog insurance” in your search input and your People Ops team calls it “pet insurance” in their documentation, Guru’s search will infer the meaning of your search term, and bump up the more relevant card towards the top of your search results list. If you’re interested in the technical details of how this technology works, check out our post about it in the Guru Community.
By getting to the meaning of what our users are looking for, we can materially improve search results for customers who have employees in different regions (we in Philly know that it’s “hoagie,” but still want our remote employees to be able to find the best “subs” in town when they visit) and have even set the stage for support across languages. Not only does this update help bridge subtle differences in vocabulary, but it enables Guru to learn from how your team searches for information, and use that context to make life a little easier for newer members of your team. Newer members who aren’t as familiar with your company jargon benefit from Guru having seen which Cards tend to get chosen for certain search terms and then adjusting the results list accordingly.
What else can Guru do with this new technology?
Tons! One of the reasons we’re so excited about this update is because it opens the door for updates beyond search. Our Search and Discovery team has expounded significant effort to get this technology in place, and the rest of our development teams and customers will reap the benefits:
With this new ability to relate content based on meaning, someday Guru could do things like form a picture of what content you tend to look at and provide suggestions for content that is similar in meaning. That’s right, personalized suggestions for you based on content you care about! Guru could also visually map how content is related across your company’s entire knowledge base and display it in different ways to different users.
We already referenced support for searching across languages earlier, but to expand on that: imagine being bilingual and searching for the Spanish “seguro para mi perro” in Guru (to keep with the dog insurance example) and getting back the most relevant “pet insurance” content, that happens to be in your second language, English! Because meaning is consistent across languages (for the most part), Guru could draw relationships between content in different languages—no Google Translate required.
This new ability to match based on meaning could help Guru’s AI Suggest Text feature become even better at bridging the gap between how your customers refer to things and how your company refers to them internally.
In the future, we can further train this model to include more industry-specific (or domain-specific) terminology to support the various verticals represented by you, our customers. We have doctors’ offices, mortgage brokers, teachers and professors, mattress makers, podcast curators, and more using Guru—each with its own unique jargon and acronyms. This technology could help us customize search in Guru to your company’s language quirks; no company is quite like your company!
Continued investments in making information accessible
Our mission at Guru is to make sure that the information our customers need to do their jobs is easily accessible, whether that’s via search, improved UX, or new functionality. We continue to prioritize effortless access to knowledge in everything we build to make Guru the best tool to help you do your best work. Keep an eye out for future updates to search and other areas of Guru where we’ll build upon this industry-leading machine learning technology.
Experience the power of the Guru platform firsthand – take our interactive product tour
