Using the Forage Records Section

Forage appears as a Records section inside your Squirrel portal, alongside Archive and Security. New to it? See how Forage fits together.

The menu has two groups.

• Records: Dashboard, Search, Ask AI a Question, Coverage, Map, Sites, Site Access, Exceptions, Readers, Queue, Activity, Log, Reports and Settings.
• Classification: Overview, Search by Label, Labels, Suggested Labels, Sensitivity Labels, Files to Protect and Content Rules. Where content classification is not switched on for a deployment, this group shows a padlock.

Dashboard

The overall picture:

• which site is being read now, how many of the included sites are done, and how much of the site on show is done;
• an Up to date message, with how many files will be checked later, when no included site has a file left to read;
• how many documents Forage is tracking, and how many are searchable, waiting, and could not be read;
• document libraries read and still to read;
• progress by source, live and archived, each with how much of it is searchable and how many files are still to read, with the files waiting for their next try shown as Will be checked later;
• documents made searchable per day and per hour, and an estimate of the time to finish at the rate actually achieved.

The estate total is only as current as Squirrel's last scan of each site, which the page states. Other figures refresh on a short cycle.

How to: follow Forage's progress.

Search

One full-text search across live and archived content. Results carry highlighted extracts, the file's name and location, and the retention and sensitivity labels Squirrel recorded for it. A file still in SharePoint links straight to it; a file in the archive links to the Archived Files page instead, ready to restore.

Search tips, as the page gives them:

• Put a phrase in quotes to find those words together and in order.
• Use | between words for OR.
• Put - in front of a word to exclude it.

Filter results by All, Live or Archived, and open any result to read its text and its details: where the file is, what Forage did with it, any identical copies, and any problems reading it (read a file's details).

Search covers only what Forage has read so far, which the Coverage page quantifies.

How to: search, and see why a file was not read.

Ask AI a Question

A Full Administrator can type a question in plain English on an Ask AI a Question page in the Records section, and receive an answer from the AI assistant that runs inside your own deployment. The page offers:

• a box headed "What would you like to know?";
• three cards to choose from: a quick answer; Read every matching file, with a choice of how many files to read (1,000, 2,000, 5,000 or all of them) and about how long each takes; and Find every matching file, which lists every matching file, however many, without reading them;
• the file types to search (all of them, or chosen groups such as Word documents and Emails) and the sites to search (all of them, or chosen sites), which the answer says it searched;
• an address for an email with a short summary of the answer, the whole answer as a PDF, and a link to it;
• a window over the page, opened as soon as a question is asked, with the steps while it is answered and, for Read every matching file, progress and an estimate of the time left. When reading every match would take a long time, the window asks how many files to read, with Read and Not now, and can cancel the question;
• the answer, with the number of the file behind each statement, each cited file opening from the answer, the files the answer used shown as cards (the cited ones first, the others marked Not cited), List every matching file, and a link to all the matching files in Search;
• for a list of files, how many match, the files a page at a time, each opening from the list, and the list exported as a spreadsheet file (CSV);
• an Export the findings choice under an answer: PDF, CSV or copy as text;
• the earlier questions, each with who asked it and when, which can be reopened, and Ask again at the end of a finished, stopped or cancelled question, which puts it back in the box with the same choices to change.

It finds files by the words in them, not their meaning, uses only files Forage has read, can misread a figure and has no follow-up questions. It is for Full Administrators only.

How to: ask AI a question about your files. See also the AI assistant.

Forage for Copilot

Microsoft Copilot, or another AI assistant that signs in through Microsoft, can search Forage and read the text of the files Forage has read, archived files and scans included, so a question asked in Copilot or Teams is answered from everything Forage has read. Records administrators see every file Forage holds, as they do on these pages; staff see only the files they can open in SharePoint. Every search and every file Copilot opens is recorded in the Log under the person's name.

See Forage for Copilot, and set up Forage for Copilot.

Coverage

Coverage keeps three answers apart: how much of the files in scope is searchable, how much of all of SharePoint is, and how much of what Forage has read so far is. It also shows whether what was read is being kept current.

It shows:

• how much of the files in scope is searchable, how many are still to read, how many have a problem now, and how many sites are listed;
• where every file in scope stands: searchable, waiting or being read, could not be read, listed but not read, or not listed yet;
• the reading rate and the files found in the last hour;
• the sites Forage has started but not finished, with the files found so far and the time so far;
• under tabs, coverage by file type, the problems now, what was listed but not read and why, how files were read, and all of SharePoint for reference.

Everything Forage does not read is still counted, with the reason. It is accounted for, not missing.

Map

Every SharePoint site on the Sites list, drawn as a tile. The bigger the tile, the more data the site holds, or the more files, if you size by files.

• sites sit in three blocks, Included, Not decided and Excluded, each with its biggest sites and a striped tile for the rest;
• colour the tiles by how far Forage has got, by whether a site is included, by its state or, where content classification is enabled, by the most sensitive label found in it;
• point at a tile to see its figures, click to keep it, double-click to open the site's libraries;
• from a site's panel, open Search on that site or open the site in SharePoint;
• where content classification is enabled, pick one or more labels and the map shows only the sites that hold files with them, sized by how many, with a link to every file with the label in Search;
• include a site that is not included, with a confirmation that is recorded in the Log.

The map reads its figures again every minute, and a Full screen view shows it on its own.

How to: see your SharePoint estate on the map.

Sites

Every site on Squirrel's list, read or not, with its file count and size, whether Forage reads it, its rank, and how far reading has got: not started, reading, checking again, done, empty, deleted, excluded or not decided. A site deleted from SharePoint stays on the list, marked Deleted: Forage has finished with it, and the files it had read stay searchable. This is also where you choose what Forage reads.

• Include or exclude sites, one at a time, in bulk, or by an address pattern.
• Rank them, so the sites that matter are read first. Rank 1 is read first, 2 second; an unranked site is read after every ranked one.
• Choose what happens to new sites, and to a site you exclude after Forage has already read it: keep what was read searchable, or remove it from search.

A new deployment reads nothing until sites are included. Forage follows each change from its next pass over the sites, every change is recorded with who made it and when, and excluding a site never touches the site itself.

How to: choose which sites Forage reads.

Site Access

Who can open each site Forage reads, as SharePoint's own permissions say. Forage checks every included site each night, reading only, and a newly included site soon after it is included.

• for each site, how many people and groups have access, how many of them only because a file, folder or library inside it was shared with them (Shared items only), and when it was checked;
• for a chosen site, who can open it, grouped by how they get in: site administrators, each SharePoint group, access given directly and shared items only, and whether the site is open to everyone;
• sites that could not be checked, which keep what the last check found, and sites deleted in SharePoint, which show as Deleted in SharePoint.

Forage for Copilot uses the same record to start from the sites each member of staff can reach.

How to: see who can open each site.

Exceptions

What could not be read, grouped by cause and ranked by documents lost.

• when each cause was first and last seen;
• which kind of cause each is, and so who can act on it;
• whether each cause will be retried or needs someone to act;
• which sites lost documents;
• the documents behind each cause, with filters, and an export of the current page of that list.

Failures that belong to the document itself, such as a password or encryption, or to a copy SharePoint held that was already damaged before Forage fetched it, are labelled as such so nobody chases them as faults.

How to: understand exceptions and unread files.

Scanned pages that could not be read are not yet listed here. Their documents are counted as partly read.

Activity

• what has been read;
• what an administrator changed;
• who searched, and for what;
• what would not open.

How to: see what Forage has been doing.

Queue

What is waiting to be read and what is being read right now, for administrators.

• tiles for the files waiting to be read, the files made searchable in the last hour, and an estimate of the days to go at the current speed;
• Changed files waiting: files that changed after Forage read them. Each is read again to see whether its content really changed, and search finds its earlier text until then;
• Large files waiting, with how many are being read now. Forage reads a limited number of large files at once, so when most of the waiting files are large, the other readers stand idle until one of them finishes;
• Unread scans: files whose scanned pages were not read. Their typed text can be found in search, but the words in their scans cannot yet. They are read again when the file next changes;
• Where the files are: for SharePoint and for the archive, how many files are waiting, being read, searchable and could not be read, with a bar of how far each has got;
• Being read now: the files being read this minute, longest first, with the site, where each is kept, its size and how long it has been read. A long scan can take many minutes, so a long time is not always a problem, but it is the first place to look if reading seems stuck.

If no files are being read the page says so, and says when the last file became searchable. When every waiting file is waiting for its next try, a notice at the top says that nothing is waiting that a reader can take right now, when the first of them falls due, and that reading carries on by itself as each try falls due; the days-to-go tile then reads "Nothing to read now" with the time of the next try under it. If reading pauses until there is room for more files, a notice says so and reading carries on by itself.

Readers

The readers that are running now, for administrators. Readers read your files, and where content classification is enabled, rule checkers check them against your rules.

• how many are running and what they are working on;
• for each one, its name, what it is doing (standard files, large files or rule checks), where its files come from (SharePoint, the archive, or both), how long it has been running and how many files it has done;
• a notice when SharePoint reading is paused, because Microsoft asked Forage to slow down or an administrator paused it, saying until when. The readers work on the archive meanwhile.

Forage chooses how many readers run and adjusts the number itself, so there is nothing to set. Readers that have stopped are not listed. If none are running, the page says so; if that does not change within a few minutes, contact your administrator.

Log

One timeline of what people did, newest first. The Log records every change to settings, content rules, labels and sites, every search, every document opened and every report exported, each with who did it, when, what it was done to and the reason given.

Open an entry to see the value before and after the change, and the detail recorded with it. Filter by kind, by person, by any text, and by date, then export to PDF or a spreadsheet.

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. Sign-ins to the portal are not part of the Log: they go through Microsoft Entra ID, which keeps its own record of each one.

How to: see who did what.

Reports

Five reports, exported as a PDF or a spreadsheet. The first four can be filtered by site and by date, and Duplicate files by site:

• All your files: files by site, by SharePoint or Archive, and by type, with how much was read in full, read in part, or not read, and why.
• Sensitive content, where content classification is enabled: the files carrying each label, by site and severity.
• Labels by file, where content classification is enabled: one row per file, with the labels on it and when it was last checked.
• Files not read: what was not read, and the reason.
• Duplicate files: the files whose contents are exactly the same, byte for byte, and the space the extra copies take, with the largest sets, where each copy is, and the sites whose extra copies take the most space. The copies of deleted files that the archive keeps so they can be recovered are left out.

Every export states the customer, when it was produced, and the filters used. Lists show up to 100 rows; the summary figures count everything the filters select.

How to: produce a report for a board or an auditor. Forage can also email the reports on a schedule.

Classification Overview

Where content classification is enabled, one page showing how far checking has got and what has been found.

• how many files have been checked out of all the files that can be checked, the share done, and whether checking is running;
• files checked, waiting, labelled, the active labels and rules, and files that could not be checked;
• how many files carry each label, and the sites holding the most labelled files;
• the files most recently labelled, and the most recent changes to rules.

A file is checked once Forage has read it, and Forage reads a file's pictures and scanned pages in the same pass as its typed text, so a file is checked on all of its text. Files with no text whose scanned pages were never read are shown apart: they have nothing to check unless they are read again. Checked means compared against the rules active now. The totals are prepared in the background every few minutes, and each says how old it is.

How to: see classification at a glance. You can also find files by label in Search.

Labels

Where content classification is enabled:

• how much of what Forage has read has been checked against the current rules;
• each label with its document count and severity;
• create, edit, retire, reinstate and delete your own labels, with a history of every change;
• the rules behind the labels, with the documents behind each.

Counts are within the documents checked, not the whole estate, and are counted in documents rather than findings. When the rules change, the checked figure falls and refills as Forage checks everything again in the background.

New to this? Start with labels and rules in plain words.

How to: create and edit labels, and see what a label or rule found.

Suggested Labels

Where content classification is enabled, the labels Forage suggests from the files it has read.

• choose the sites to look through, and follow the steps while suggestions are prepared;
• each suggestion with its file count, example files, keywords and sites, and how many files a rule on those keywords would find;
• edit a suggestion's name, category, severity and words before deciding, then accept it, which creates the label and a draft rule, use it for an existing label, merge it into another suggestion, or reject it.

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 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 Log. Reading slows down while suggestions are prepared, and returns to normal speed by itself when they are ready.

How to: get label suggestions.

Sensitivity Labels

Where content classification is enabled, a page that lists the sensitivity labels your own Microsoft 365 tenant publishes, such as Confidential or Public, and checks which files carry them. These are Microsoft's labels, not Forage's: a Forage label records what a document contains, and a sensitivity label records how your organisation has decided it should be handled.

• for each Forage label, choose which sensitivity label its files should carry, which other labels are good enough, and which may be replaced, such as General or Public;
• start a check: Forage looks at the sensitivity label each file carrying that Forage label has now, in Microsoft 365's record of labels;
• see, as a dry run, what a labelling run would do: how many files would get the label, how many would have a weaker label replaced, and how many it would leave alone because they carry another label, their label is not known, or they are no longer there;
• open any file in the list, and export the list as a spreadsheet.

Nothing is applied. The check only reads and reports, and its results are as of the check. Files in sites Forage does not read, and in OneDrive, are left out.

Files to Protect

Where content classification is enabled, an administrator can choose 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. On the page, an administrator can:

• choose the label and where Forage looks for files that carry it;
• see the files that look like the labelled ones but carry no label, or only General or Public, grouped by the kind of document they resemble;
• open a file, mark it Protect or Not sensitive, and export the list.

Forage never applies or changes a label. It only points files out, and a person decides what to label. It can list files that do not need the label and can miss files that do, so a person reviews what it finds.

How to: find files that may need a sensitivity label. See also the AI assistant.

Content Rules

Where content classification is enabled, this is where a rule is written, saved as a new version, test run, activated, retired, or deleted while still an unused draft.

Every change is a new version. Rules are not edited in place.

A test run before a rule goes live shows three figures:

• the most the rule could match;
• a sample checked with the real rule;
• a projected range across the documents read so far.

The test run is offered rather than required, and its projected range covers the documents read so far rather than the whole estate.

Check accuracy. Forage picks up to 25 of the files an active rule labelled, at random. You open each one and mark it Right, Wrong or Not sure, and the rule's card then shows the share that was right, with the range the true figure is likely in, because 25 files is a sample. Files marked not sure are left out. Every check is recorded in the Log, file by file, and belongs to the rule version it was made on.

Suggest words. Forage reads the opening of up to 25 files the rule labelled, picked at random, and lists the words that are common in those files and uncommon in the rest of the files it holds. Each word shows how many of the sampled files contain it, how common it is elsewhere and whether the rule already uses it. Tick the words you want and the rule editor opens with them added to a new version. Nothing changes until that version is saved and switched on, and you decide every word. The list is a set of ideas from a small random sample, so it differs a little each time.

How to: write, test and switch on a content rule. There is also a worked example, from a new label to its results.

Settings

• Read live content, archived content, or both.
• Choose which document types are read.
• Choose what happens to documents of a type that is switched off: Keep them, or Delete them permanently.

Delete them permanently requires a reason, which is recorded. It removes Forage's copy of the text, its search entries, and its record that those documents were examined. The documents themselves, in SharePoint and in the archive, are untouched.

Every change is recorded with who made it and when, and takes effect automatically.

Which document types a deployment can read is confirmed by SmiKar. A new deployment starts with the most common types.

How to: change what Forage reads.

---

Questions go to sales@smikar.com.