Scribd
Upload
574 views24 pages

How Glean Search Works

Explains how Glean works.

Uploaded by

charliedotcrews
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
574 views24 pages

How Glean Search Works

Explains how Glean works.

Uploaded by

charliedotcrews
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
role a) How Glean Search Works Providing better results through innovative machine learning and indexing Table of contents The sate of search Glean search architecture Unboxing and setup Building the knowledge graph ‘The search journey 4 The state of search Mos:orusbossn ou aiatatiourey tough a hum search ba. “Tran, sare engines ae stil round 083 and 6 beter han ever ore. Homevey ost sean sokuton or buses and Saas ‘sppleatons, especie ntveslusons, a alarming being in fumetonlity compared the consumer options we use on 3 daly basis cute of wrk {Win the volume ofnermaton, 7d appcaions we use each aay only growing, having a eedly avaliable and personalized solution or scoverng existing knowledge Ino longer a ury~ its essential to ‘enabling your business work optimaty without interuption, 25 wel as onboaraing now enpiyees and tols a seamiassy a8 poss be, (ten s slung the proiem of workplace search, once and fr al. Giean iste enterprise search platform that helps evry employee ind the Information that thoy need, Instant. takes everyday queres at work and processes them in way that proses accurate an persoraized resus according to Ne way you work, andthe people you werk wih, dean ‘Our founders -incucing Arvind Jin, CEO of Gisan, who came from over @ decade at Google 35 a Distingushod Engnoor als recognized hata ‘eerie earch tool would never be good enough forthe jb. requies 3 ‘unquely personalized end constant Improving soutlon tat Naps search ‘work the way you need 2. Our latfoem uses the latest advancements in ‘machine learning to understand each business unique way of working, ‘ad provide users with the most personalized resus for thr queries. ‘Tis eBook wil gue you tough that unique Seah process also Ines dtalls on how Glan bul its Index on xing enterprise dats, ‘ong with how ensues the results you receiv are exactly what you roed 0 work without intorupion or ly. Late got states Glean search architecture TOR ee ay Cee eee eee en ee ee Cee eee ee ee ee eee cae ee cee Ce meer een) eee ee ec Sees cry Unboxing and setup Trptementng great search soon canbe hata work Formast optone cout ther, tinvoves + Monua seareh tuning + Dedicates develope support to ensure proper ranking + Leberious tweaking wth permissions to ensure securty Its ot of an commitment ram the user ena bth to got he system up ans runing 38 wal 38 constantly iterating to gett to work the Way you ‘lean takes that hare work ou ofthe equation. Glan search slaton ‘uly custortzabe, buts but to requis minimal operational overhead to ‘set up ana needs no thid-party engagements proessenal services Invests, dean Other solutions sraRT Gareaoy ‘Te pany engagements Pemislon waking glean sraRT Greaoy esas ‘Te deployment process bogins through the creation ofa naw Project n| Google oud Pattern (GCP), Each orgoizatin ratains thelr enterprise data within is exclusive oud intastutue. This sngle-tenant srentectue model, wien delegates separate GCP projet per customer, also guarantees complete ana secure dat isolation Your entaroise informations axclisively Usd to tran your own deployments machine Teaming models = no ore ose, Ciean aso wor wth the prrple of east privieg, ana iret enforces data access rls nat xt at multiple evel. “Then, once Gani given management access and is connected to pectic data sources (Google Dive, Sox, SFDC, et) athe customers raquest, ou software bagins to take the nacessary steps to construct ine foundational enterprise knowedgegranh tat powers Gleans unique search system. dean Level of data acces Product (eg, Ssesoe, Gage Drv, Fa rat Object. Record ius pce jc th str ent tes ect Rete wpa poet ‘ars who ese e-grioe perma Building the knowledge graph Geieon's sau search engine operates based ona reltine model of you entero indexedntrmation~the enterprise knowledge raph “Ts system involves thee ey pitas Content, “People “Activity ‘ Inawigual assets + idamtee end roles + Content creston + Documents ams + Eating history + Messages + Departments + Comments + Tekete + Groups + Searches dean Content Filing the krowiedge graph wih content begins wit ou easy 100+ connectors. Each connectors talr-made foreach appicaon’s Unique data made and APl endpoints, equ no son professions sccesabaty services to use, nds uly permissions-aware - ensuing and shating protocols ee srty flowed for aac source. -Communleton. Documents and wikis Mee ROEG “cae Cloud nernge ° 6 sebu 2 e0z gow dean Through these connectors, Gian’ content crawler searches over every part of apiece of con ‘content (ies, body opy, comments, meds ete), 8 we 38 met te erestor,tme of cestion,uodate story, le ype, toler structure, te). Allo tse elements, store with the knowledge graph nae, become racy searchable, Customizable weights can algo beset for specie etogores to inuence search esuks Facet soar esl can be tuned based on melas lds specie to each 9p, net us the tle, Ths nvoves relevant tern Permissions can also be Invidaly st fr specific ems. Fes canbe selective ince or axcluses from Gears crawl system by epectying| ast IDs or bya broader categorization, such stem containers (lders oratives Content cram strategy i also completely unable according to your ‘orgonization’s preference Aue raw! cadence, assign blackout haus to _2v0id peak work hous, an sift botwoen diferent ral methodoogios to ensue your knawedge graph contin te most relevant and upto-

You might also like