Custify Knowledge

teach custifyai how your team actually works by feeding it the policies, processes, and operating context it needs to reason about your accounts overview custify knowledge is the account level ingestion layer that turns the materials your cs team already uses into a structured context for custifyai instead of relying only on generic product behavior or scattered notes across tools, you can upload the documents, recordings, urls, and answers that describe how your team handles onboarding, escalations, renewals, handoffs, and everything in between each input becomes a knowledge record that custify processes into clean, searchable content that other modules can use knowledge is the ai layer it's where you ingest sources — documents, urls, past conversations, product data — so that custify's ai can draw on them when answering questions or surfacing insights think of it as feeding the engine your team doesn't browse knowledge records; the ai reads them the internal kb is the human layer it's a structured article module your team can actually open, search, and read — like an internal wiki it's designed for colleagues who need quick access to documented processes, playbooks, or reference material the two can connect when a knowledge source is well structured enough to double as documentation, custify can generate a draft internal kb article from it but that's a one way assist — not a sync knowledge doesn't automatically become a kb article, and an internal kb article isn't automatically ingested as a knowledge source a simple way to remember it knowledge feeds the ai the internal kb feeds your colleagues knowledge module capabilities the module gives you four core capabilities add knowledge from files, pasted text, urls, audio recordings, or guided cs operating model questions process raw input into formatted, searchable knowledge with a type, classification, and source metadata extract draft rules and draft internal kb articles from suitable sources, with full review and approval before anything goes live index ready knowledge so custifyai can use it as context inside playbooks, customer 360 questions, ai generated playbooks, and future agents the module can be accessed directly from the app menu → knowledge each row in the library represents one knowledge record, showing its type (file, text, url, or recording), classification (such as process, support material, or general knowledge), processing status, and the date it was added custify tip the more knowledge you add, the more account specific custifyai becomes a team that uploads its onboarding playbook, escalation policy, and renewal process gets noticeably more accurate ai summaries, customer objectives suggestions, and playbook drafts than a team that relies only on generic product behavior adding a knowledge source you can add knowledge in five ways each input type goes through the same processing pipeline and produces a knowledge record once it is ready to start, go to app menu → knowledge the upload options are available directly on the landing page the five input types and what each is for cs operating model a guided questionnaire is available at the top of the page if you are setting up knowledge for the first time, start here — it walks you through 28 questions covering your customer journey, lifecycle stages, escalation rules, handoff process, and team specific decision logic, and turns your answers into a structured knowledge baseline automatically paste text free form text you type or paste in use this for short policies, escalation rules, or notes you want custifyai to reason with but have not formalized into a document upload file pdfs, word documents, slide decks, and similar formats use this for existing process documents, training decks, and customer facing materials your team already maintains from url any public web page use this when the source of truth lives on your help center, public docs, intranet wiki, or a vendor's site record audio upload or record an audio file use this for verbal walkthroughs of how your team handles a specific scenario, as writing it down would take longer than recording it for each source, you fill in the following fields before submission title a short name for the record this is how the source appears in your knowledge library and how custifyai references it in answers classification the category that describes the source options include general knowledge, process, rule/policy, support material, reference, training, or operating model the type guides what custify can extract from the source later description optional a 1 2 sentence summary that helps your teammates recognize the record without opening it click submit when you are ready the record enters the processing queue with a status of processing and moves to ready once custify finishes parsing and structuring the content observation processing time depends on source length and type audio recordings and large pdfs take longer than pasted text or short urls if a source fails processing, the record shows a failed status, and you can retry from the record's action menu reviewing and editing processed knowledge once a source moves to ready , you should open the record and review the formatted content before relying on it custify cleans up the raw input, removes irrelevant boilerplate, and structures the result into sections custifyai can search efficiently the processed version often differs from the original, especially for audio and pdf sources, where the parsing is more involved to review a record, click its title in the library the detail view shows the formatted content, the original source metadata (file name, url, recording length, etc ), the classification, and the generated articles menu inside the detail view, you can edit the formatted content directly use this to fix typos, remove information that should not be shared with custifyai, or clarify wording that the parser misinterpreted update the classification download the source file (for file uploads and audio recordings) retry processing if the result is wrong or incomplete delete the record deleted knowledge is removed from the index immediately observation edits to a knowledge record take effect the next time custifyai references the source existing ai outputs that already used the previous version (such as saved summaries, generated drafts, or extracted rules) are not retroactively updated the processing status field has four values you should know processing the source is in the queue or currently being parsed ready processing finished and the record is indexed for custifyai failed processing did not complete use retry from the action menu, and if it fails again, check that the source is readable (audio is clear, the url is publicly accessible, the file is not corrupted) excluded you manually excluded the record from the index the record stays in the library, but custifyai does not use it as context excluding rather than deleting is useful when you want to retire a policy temporarily without losing the source generating rules and kb articles from knowledge some knowledge sources contain operating rules or are good candidates for documentation custify knowledge automatically extracts draft outputs from these sources for you to review there are two output types draft rules logical conditions custify extracts from rules and polic y documents and from the operating model (questionnaire filled in) for example, an escalation policy that says "if the customer has not responded for 7 business days, the csm should hand off to the team lead" becomes a draft rule draft internal kb articles long form drafts suitable for your team's internal kb when a source is dense and explanatory rather than rule based, custify produces a draft article instead you review the draft, edit it, and publish it into the internal kb module draft articles are generated by documents classified as general knowledge , process , support material , reference material , and training or context observation! the operating model is the only type that does both (rules and kb drafts) rule or policy is the only type that generates rules, but not kb drafts for each draft, the action options are review open the draft in the editor edit modify the draft before approving it publish for draft rules, this activates the rule so it is used by custifyai and rule driven features for draft articles, this publishes the article to the internal kb archive discard the draft rejected drafts are removed, and the source itself is not affected custify tip drafts are starting points, not finished work you stay in control read every draft rule before approving it, and check every draft article against your team's voice before publishing publishing is the moment a draft becomes active, so treat it as a deliberate step managing your knowledge library the library view in app menu custify knowledge is where you organize, audit, and prune your sources over time the library supports only filtering by type for now observation deleting a knowledge record is permanent the source content, the formatted version, and any draft outputs that have not yet been approved are all removed approved rules and published internal kb articles continue to exist independently of the record they were generated from, so deleting the source does not retract a published article or deactivate an approved rule for sensitive sources, use the classification field to mark them clearly (for example, internal only or executive ) and set up settings roles & permissions to restrict who can view, edit, or generate outputs from those records treat your knowledge library the same way you would treat any other place where customer related context lives review it on a cadence, retire what is out of date, and keep what is active and accurate