Forage Overview

What Forage does

• Estate inventory. Forage builds its inventory of SharePoint Online files and Squirrel-archived files from the records Squirrel already keeps, so SharePoint is not crawled a second time.
• Reading the archive. Forage reads the content Squirrel has archived, opening the encrypted, compressed archive copies inside your own deployment with your own archive key, and reads the current version of each file.
• Document types. Every cell of an Excel workbook is read, including cells that hold only a number or a date, as the workbook shows them. PDF; Word, Excel and PowerPoint files, including templates and macro-enabled files; older Office formats; Outlook and standard email messages; OpenDocument files, as LibreOffice and OpenOffice save them; the words on drawings (.dwg) saved in any version of the format from 1984 on and on exchange drawings (.dxf) (their text, title block values and notes, and the properties where the version keeps them, not the drawing itself); and rich text, plain text, CSV, HTML and XML, without needing Microsoft Office.
• Scanned documents. Scanned PDFs and image files are read with optical character recognition, so scanned contracts and forms become searchable.
• A record of how each document was read. For every document it reads, Forage records how it was read, page by page, and keeps a cryptographic fingerprint of the file.
• Coverage accounting. Every file in the estate is accounted for, whether read, excluded by type or by policy, or failed with a recorded reason.
• Search. One full-text search across live and archived content, with highlighted extracts, filters by source and document type, English word-stem matching, and each file's name and location on every result.
• Keeping current. Forage keeps its picture of the estate current by asking Squirrel what was added, changed or deleted, and reads a document again only if its content really changed.
• Content classification, where it is enabled on a deployment. Forage recognises what documents contain, such as bank details, payment card numbers, tax and company identifiers, passport numbers, health information and contract terms, using transparent rules. Every finding keeps the words that produced it.
• Writing and testing rules, where content classification is enabled. Administrators write, version, test and switch on content rules in the portal, and a test run before a rule goes live shows the most it could match, a sample checked with the real rule, and a projected range across the documents read so far.
• Labels and the documents behind them, where content classification is enabled. Administrators see, label by label, how many documents contain each kind of content, how serious it is and in which sites, and can open the documents behind any label or rule and read the matched text in context.
• Managing your own labels, where content classification is enabled. Administrators create, edit, retire, reinstate and delete their own labels in the portal, and every change is kept in that label's history. A label a rule is using is retired rather than deleted, so the documents already found under it, and its history, are kept.
• Suggested labels, where content classification is enabled. Where content classification is enabled, an administrator can ask Forage to suggest labels for the sites they choose. The AI assistant, which runs inside your own deployment, reads the files Forage has read there, groups them by what they contain and suggests a label for each kind of record it finds, such as invoices, packing lists or test reports, with how many files it covers, where they are, example files to open, the keywords a rule would look for and how many files in those sites such a rule would find. Nothing is labelled until a person accepts a suggestion and switches its rule on, a rejected suggestion is not made again, and every decision is recorded in the audit log.
• Ask AI a Question. A Full Administrator types a question in plain English on the Ask AI a Question page in the Records section and receives an answer from the AI assistant that runs inside your own deployment. The AI assistant searches the files Forage has read, reads the matching passages of the best-matching files, and writes an answer in which each statement gives the number of the file it came from. Each cited file can be opened from the answer. An administrator can limit a question to chosen types of file, such as Word documents and emails, and to chosen sites. For a bigger question, a Read every matching file choice reads the matching files in the background, as many as the administrator chooses (1,000, 2,000, 5,000 or all of them), and combines what they say into one cited answer. When an administrator wants the files rather than an answer, a Find every matching file choice has the AI assistant turn the question into a search and lists every file it finds, however many, without reading them, with a spreadsheet export. The answer can be emailed as a PDF, and exported as a PDF, a spreadsheet file or text. Every question, who asked it, when, the answer and the files it used are kept, and administrators can reopen them. See the AI assistant and its limits.
• Files to Protect, where content classification is enabled. An administrator chooses one of their Microsoft 365 sensitivity labels, such as Confidential, and the AI assistant that runs inside your own deployment looks at the files carrying it and lists the files Forage has read that look like them but carry no label, or only General or Public, for a person to review. Forage never applies or changes a label: it only points files out, and a person decides what to label. See the AI assistant and its limits.
• Checking how often a rule is right, where content classification is enabled. An administrator marks a random sample of the files an active rule labelled as right or wrong, and Forage shows the share that was right with the range the true figure is likely in.
• Suggested words for a rule, where content classification is enabled. An administrator asks for words that stand out in a random sample of the files a rule labelled, ticks the ones they want, and adds them to a new version of the rule. Nothing changes until that version is saved and switched on.
• Checking sensitivity labels, with a dry run of labelling, where content classification is enabled. An administrator chooses which of their Microsoft 365 sensitivity labels the files carrying each Forage label should carry, and Forage shows which files already do and what a labelling run would do with the rest. Nothing is applied.
• The SharePoint map. Every site on the Sites list drawn as a tile sized by its data or its files and coloured by how far it has been read, whether it is included, its state or, where content classification is enabled, the most sensitive label found in it. An administrator can pick labels to see which sites hold their files, open a site's libraries and go straight to its files in Search.
• Scripts and orientation on scanned pages. Forage also reads rotated scanned pages, accented Latin-script pages and Cyrillic-script pages. Documents already read are not read again automatically.
• Labels Squirrel recorded, on results. Search results show the retention label and the sensitivity label that Squirrel recorded for each file.
• What could not be read. Forage shows what could not be read, grouped by cause and ranked by the documents affected, and says whether the cause lies with the document itself (for example a password), the copy SharePoint holds (damaged before Forage fetched it), the source, the infrastructure or Forage, and whether it will be retried or needs someone to act.
• Choosing what is read. Administrators choose live content, archived content or both, and which document types are read. Every change is recorded with who made it and when.
• Progress reporting. The Dashboard and Coverage pages show how much of the estate is searchable, which sites have been read, whether read sites are being kept current, the processing rate, and an estimated time to finish at the rate actually achieved.
• The Log. The Log records changes to settings, content rules, labels and sites, data resets, searches (the words used and who searched), documents opened in the portal, report exports and emailed reports, whether sent or not, and every test email asked for, each with who did it, when, why, and the before and after where that applies. The Log is tamper-proof: no one using Forage, administrators included, can change or remove an entry, and clearing Forage's data to read the estate again keeps them. It exports to PDF or to a spreadsheet, up to 1,000 entries at a time. It does not record search results, by design.
• Access. The Records section is available only to the Squirrel portal's Full Administrators.
• Per-customer deployment. Each customer has its own Forage deployment, with its own database and its own search index. There is no shared index and no path for one customer's query to reach another customer's data.
• The source is left alone. Forage only reads SharePoint, the archive and Squirrel's records. It does not change, move or delete any source document, and it writes nothing to Squirrel.
• No document content leaves the deployment. The text inside your files is never sent outside your own deployment, and no document content goes to any external AI service. Forage's AI assistant runs inside your own deployment and suggests labels; every label is applied by a transparent rule that a person has switched on, and the AI assistant does not learn from or train on your content, and scanned pages are read by optical character recognition that runs inside the deployment.
• Nothing is deleted. Forage does not delete anything.
• Recoverable index. Forage's search index can be rebuilt from the text Forage already holds, without reading SharePoint or the archive again.
• Forage keeps a copy of the text. Forage keeps the text it reads, and a search index of that text, inside your deployment.

Content classification is included on the deployments where it is enabled. Confirm with SmiKar what a given deployment includes.

See product features for how each feature behaves.

Questions go to sales@smikar.com.