Every document you have, searchable. Every sensitive one, found.
Forage reads the documents in SharePoint Online and in your Squirrel archive: the actual words inside PDFs, Office files, emails and scanned pages. It makes all of them searchable in one place. Switch on classification and it finds the ones holding card numbers, bank details, passport numbers or contract terms, and tells you which sites they are in. Then it gives you what an auditor asks for: what was read, what was not, and why.
Forage is the records half of Squirrel. Squirrel archives what is old and leaves a stub behind; Forage makes sure that nothing, archived or live, falls outside what you can search, classify and account for. A records tool that only sees live SharePoint loses sight of content the moment it is archived. Forage was built so that it does not.
One search. Your data. No gaps.
One search for everything. Live SharePoint content and archived content in a single full-text search, with highlighted extracts, filters by location and document type, and every file's name and where it lives. The retention and sensitivity labels Squirrel recorded appear on each result.
Your data stays yours. The text Forage reads, and its search index of that text, stay inside your own deployment. No document content is sent anywhere outside it. Each customer has a deployment of their own, with its own database and its own search index.
Nothing goes missing quietly. Every file is accounted for: read, excluded by type or by policy, or failed with the reason recorded. Coverage is reported against the sites you chose, so the figure you give an auditor is one you can stand behind.
The dashboard answers the first question anyone asks: how much of the estate is searchable, and how long until the rest is.
What you get, in two parts.
Search and coverage. The foundation. Forage reads your documents, makes them searchable across live and archived content, accounts for every file, reports what could not be read and why, tracks progress, emails reports on a schedule, and keeps an audit log of who did what. A Full Administrator can also ask a question in plain English and get an answer from the AI assistant, with the number of the file behind each statement.
Labels and rules. An add-on to the foundation. Content classification: rules that recognise what documents contain, the labels they apply, and search by label. Your administrators write and test the rules themselves, and every finding keeps the words that produced it, so an auditor can see why a file carries a label. Forage can also suggest labels: the AI assistant, which runs inside your own deployment, reads the files Forage has read in the sites your administrators choose, groups them by what they contain and suggests a label for each kind of record it finds, for your administrators to accept or reject. Files to Protect uses the same AI assistant to point out files that may need one of your Microsoft 365 sensitivity labels, for a person to review.
Features: what it does.
Reads the documents, not just the file names. 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); and rich text, plain text, CSV, HTML and XML. No Microsoft Office or AutoCAD needed on the server. Every cell of an Excel workbook is read, including cells holding only a number or a date.
Scanned documents become searchable. Scanned PDFs and image files are read with optical character recognition, so scanned contracts and forms are findable by their contents. Words in pictures inside Office documents are read too. A page read with low confidence is never left out: its text is kept and marked as a low-confidence reading, so nothing is quietly dropped.
Search that tells you what it could not read. Results carry highlighted extracts, the file's name and full location, when it was modified and its size. A file Forage has listed but not read is still found by its name, and says plainly why it was not read.
A file that could not be read is still found by its name, and the result says why rather than leaving a silent gap.
Ask a question, get the files behind the answer. A Full Administrator types a question in plain English on the Ask AI a Question page in the Records section. The AI assistant, which runs inside your own deployment, 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 opens from the answer, and the files the answer used are listed with the cited ones first. It answers only from what those files say, and says plainly when they do not answer the question. It finds files by the words in them, not by their meaning, and it can misread a figure, so check an important figure against the cited file. Each question stands on its own, with no follow-up questions, and only Full Administrators can ask.
Ask in your own words, then choose a quick answer, every matching file read, or every matching file listed.
Bigger questions: read every matching file. For a bigger question, an administrator chooses Read every matching file and how many files it reads: 1,000, 2,000, 5,000 or all of them, with about how long each takes shown first. The AI assistant reads each file in turn in the background, notes what each says about the question and combines the notes into one cited answer. When fewer files are chosen than match, it reads the ones that match best, and the answer says how many it read of how many matched. The page shows progress and the time left, the administrator can leave and come back, and a question that is waiting or being answered can be cancelled. Reading many files takes a long time, hours for thousands, and Forage reads new files more slowly meanwhile.
Each statement gives the number of the file it came from, and each cited file opens from the answer.
Every matching file, listed. When an administrator wants the files rather than an answer, Find every matching file lists every file the search finds, however many there are, without reading them: name, site, folder and date changed, each one openable, and the list exportable as a CSV file. It lists the files whose words or names match the search, so a file that uses other words for the subject is not listed.
Every question kept, and the answer by email. Forage keeps every question, who asked it, when, the answer and the files it used, and administrators can reopen them. The administrator who asks gives an email address, and when the answer is ready Forage emails a short summary, the whole answer as an attached PDF with a table of the files it came from, and a link to the answer in the portal. The findings can also be exported as a PDF or a CSV file, or copied as text. The Log lists each question asked, opened, cancelled, answered, listed or stopped, each email sent and each export, with who and when, and holds neither the answer nor anything the files say.
Microsoft Copilot can use it. Forage for Copilot lets Microsoft Copilot, or another AI assistant that signs in through Microsoft, search Forage and read the text of the files Forage has read, archived files and scans included. A question asked in Copilot or Teams is answered from everything Forage has read, and not only from the pointer page Squirrel leaves in SharePoint where an archived file was. Copilot does the asking and the writing; Forage finds the files and gives their text. Everything is read-only and each person signs in with their own Microsoft account. Records administrators see every file Forage holds, and staff see only the files they can open in SharePoint, checked with SharePoint as that person before each file is shown. Every search and every file Copilot opens is recorded in the Log under the person's name. See Forage for Copilot.
Content classification, as an add-on. Forage recognises what documents contain, such as bank details, payment card numbers, tax and company identifiers, passport numbers, health information and contract terms. Administrators write, version and test their own rules, and a test run before a rule goes live shows the most it could match and a sample checked with the real rule.
Rules you can explain to an auditor. Rules use phrases, patterns, checksum checks and words that must appear near each other. A rule infers nothing: the same document always gives the same result, and every finding keeps the words that produced it.
Your own labels. Administrators create, edit, retire and reinstate their own content labels, each with a name, a category, a severity and a description. A label in use is retired rather than deleted, so what it found stays listed and explained, and every change is kept in the label's history.
Each label counts files, not findings, so a document with four card numbers in it counts once.
Labels suggested, people decide. 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.
Files to Protect: files that may need a sensitivity label. Where content classification is enabled, an administrator chooses one of their Microsoft 365 sensitivity labels, such as Confidential. The AI assistant, which 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. It can list files that do not need the label and can miss files that do, and Forage never applies or changes a label: it points files out, and a person decides what to label.
Files that look like the labelled ones but carry no label, grouped by the kind of document, for a person to review.
An audit log you can hand over. Settings changes, rule and label changes, site changes, data resets, searches and who made them, documents opened in the portal, reports exported and emailed. No one using Forage, administrators included, can change or remove an entry. It exports to PDF or a spreadsheet.
Reports for a board or an auditor. All your files, sensitive content, labels by file, and files not read with the reason. Each filtered by site and date, exported as a PDF or a spreadsheet, or emailed to the people who need them on a schedule you set.
How Forage works: three moving parts.
Step 1: You choose what's read.
Administrators choose live content, archived content or both, and which document types are read. Every change is recorded with who made it, when and why.
Step 2: Forage reads it.
Archived copies are opened inside your own deployment with your own archive key. For every document it reads, Forage records how it was read, page by page, and keeps a cryptographic fingerprint of the file.
Step 3: It reports what it did.
The dashboard and Coverage page show how much of the estate is searchable, which sites have been read and the rate it is working at. What could not be read is grouped by cause and ranked by how many documents it cost, and says whether the cause lies with the document itself, the copy SharePoint holds, the source or Forage.
FAQ: Questions, answered.
Does Forage work on content we have already archived?
Yes, and that is the point of it. Forage reads the content Squirrel has archived, opening the encrypted, compressed archive copies inside your own deployment with your own archive key. Live SharePoint content and archived content appear in one inventory and one search.
Does any of our document content leave our environment?
Forage sends no document content to any external AI service, and the text Forage reads, and its search index of that text, stay inside your own deployment. The one exception is this: when an administrator who asks a question gives an email address, Forage emails them the answer, as a PDF with the names of the files it came from, and a link to it. The files themselves are never emailed. If you connect Forage for Copilot, the text of the files Microsoft Copilot asks for goes to Copilot, in your own Microsoft 365. Where an administrator switches on emailed reports, the reports themselves go to the addresses chosen, and they carry figures, file names and site names, never the text inside a file.
How does classification decide what a document contains?
Transparent rules that you can read and change: phrases, patterns, checksum checks such as the Luhn check on a card number, and words that must appear within a set distance of each other. A rule infers nothing. The same document always gives the same result, and every finding keeps the words that produced it, so an auditor can see why a file carries a label. Where you ask for them, the AI assistant, inside your own deployment, suggests labels for the kinds of record it finds; it suggests, and a person decides, so no AI decides which label a document gets.
Does Forage use AI?
Yes, through one AI assistant that runs inside your own deployment, and Forage sends no document content to any external AI service. Forage for Copilot is separate: if you connect it, the text of the files Microsoft Copilot asks for goes to Copilot, in your own Microsoft 365. It helps in three places. It suggests labels when an administrator asks for them. It answers a Full Administrator's questions, with the number of the file behind each statement (Ask AI a Question). And it points out files that may need one of your Microsoft 365 sensitivity labels, for a person to review (Files to Protect). Every label is applied by a transparent rule that a person has switched on, and the AI assistant never decides which label a file gets. It does not learn from or train on your content. Ask AI a Question finds files by the words in them, not their meaning, uses only files Forage has read and can misread a figure. Files to Protect can list files that do not need a label and can miss files that do. See the AI assistant.
Will it slow down SharePoint?
Forage takes its inventory from Squirrel's existing records rather than crawling SharePoint again, and reads documents at a rate its administrators control. When SharePoint asks it to slow down, it backs off and returns gradually.
What does it run on?
Forage is operated by SmiKar as part of the Squirrel platform, with its own deployment per customer: its own database, its own search index, and no path for one customer's query to reach another customer's data.
More on Forage.
- Forage documentation
- The AI assistant in Forage
- Ask AI a question about your files
- Forage for Copilot
- Find files that may need a sensitivity label
- SharePoint data classification: the tools, their limits and a path for a tenant
- Purview auto-labeling failure codes for SharePoint files
- How Forage works, end to end
- Forage overview
- Using the Records section
- Forage product features
- Forage questions and answers
- Squirrel archiving (the other half of the product)
- Security and detection (Burrow)
Ready when you are.
Talk to the Squirrel team | Read the Forage docs
Or email sales@smikar.com.