# searchit Enterprise Search — find what your organisation knows

One search across file servers, mailboxes, SharePoint, Teams, databases and 100+ sources. Full text, graphical filters, text recognition — and everyone sees only what they are allowed to see.

<https://searchit-agent.ai/enterprise-search/>

## Find what your organisation knows.

One search across file servers, mailboxes, SharePoint, Teams and databases: in the content of every document, without tagging it first. And everyone sees only what they are also allowed to see in the source system.

- [Book a live demo](#demo)

### This is what a search in searchit looks like

On the left the result list, on the right the document in the browser. The match is highlighted without opening the file.

- Highlighted in yellow: where the search term appears in the document.
- Tabs in the preview: text, original view, entities, metadata and annotations.
- With the searchit click service on the workstation, one click opens the file in its original program.

## Where was that again?

The quote from the spring, the contract with the supplier: everything is filed somewhere. Only it sits on drives, in mailboxes, in SharePoint, Teams and the business system, and each has its own search.

- Today: An employee sits at her desk between stacks of files and binders, head in her hand, searching two screens for a document.
- With searchit: The same employee at a tidy desk smiles as she holds the page she found; the screen shows a simple search window, and the small searchit AI mascot stands beside her.

**Find every document by its content.** searchit reads the text of every document, plus file name, path and metadata, and finds it across all connected sources with a single search, even when nobody knows any more where it is.

- **Found in the content** — What is found is what the document says. Your filing stays as it is.
- **All sources, one search** — Drive, mailbox, SharePoint and Teams in one result list.
- **Permissions from the source** — Every hit carries the permission from its system.

## Try a search.

Choose an example query. The hits come from all repositories at once, with the match highlighted.

*Example run with invented data*

### `termination` (one word)

- Drive · **Maintenance_contract_Kranich_2025.docx** — § 9 Termination: three months’ notice to the end of a contract year. (Contracts › Suppliers · 14 Mar 2025)
- Mailbox · **RE: Renewal of hosting agreement** — … please put the termination deadline of 30 Sept in the calendar. (Purchasing · 2 Jul 2025)
- SharePoint · **Lease_warehouse_Hamburg.pdf** — Termination requires six months’ notice to the end of a quarter. (Administration · scan, read via text recognition)
- Teams · **Purchasing channel** — We can end the framework agreement with three months’ lead time. (Chat · 11 Aug 2025) — *found by meaning*

### `"second reminder"` (phrase)

- Drive · **Reminder_RE-2023-1001.pdf** — Second reminder for invoice RE-2023-1001 of €29,712.65. (Accounting › Dunning · 15 Jun 2026)
- Mailbox · **Reminder AutoTeile Müller – follow-up** — … the second reminder has gone out, payment promised by the end of the month. (Accounting · 17 Jun 2026)
- SharePoint · **Open_items_2026-06.xlsx** — Status RE-2024-1012: second reminder sent. (Finance › Reports · 15 Jun 2026)
- Mailbox · **June dunning run** — Dunning level 2 is set for RE-2024-1012. (Accounting · 14 Jun 2026) — *found by meaning*

### `inspect*` (wildcard)

- Drive · **Inspection_report_forklift_07.pdf** — Inspection report from the annual test of forklift truck 07. (Warehouse › Vehicles · scan, read via text recognition)
- SharePoint · **Inspection_reports_2025.xlsx** — Overview of inspections by vehicle and date. (Fleet · 9 Jan 2026)
- Mailbox · **Insurance, Hall 2** — We need the inspector’s report number for the damage claim. (Administration · 21 Nov 2025)
- Teams · **Warehouse channel** — The test record for forklift 07 is in the workshop folder. (Chat · 4 Dec 2025) — *found by meaning*

Semantic search, where set up: also finds other words for the same thing.

### How does searchit find it?

Behind the single search line there are five steps. The first three run beforehand, in the background.

1. **Connect** — One crawler per source that picks up new, changed and deleted documents on a schedule.
2. **Read** — Text, file name, path and metadata of every document; scans and photos via text recognition.
3. **Capture permissions** — Every document carries the permissions from its source into the index.
4. **Search** — Keyword search with word stems and synonyms, combined with semantic search where set up.
5. **Filter and show** — With every search, only what the user is allowed to see, with the match highlighted in the preview.

**For the difficult question**

- Phrases in quotation marks, wildcards such as “maintenan*”
- AND, OR and NOT, also in groups
- Suggestions as you type, similarity search and “Did you mean?” for typos
- Word proximity, search in individual fields and regular expressions

## See only what you are allowed to see.

The same search, two colleagues: switch the role. What someone may not open in the source, they do not see in searchit either, with every search.

*Example run with invented data* — Signed in as: Payroll / Sales, „salary“

- Drive · **Salary_list_2025.xlsx** — Salary, allowances and tax code per employee. (Payroll)
- Mailbox · **October payroll: approval** — … the salaries go out on the 28th. (Payroll)
- SharePoint · **Travel_policy_2025.pdf** — Expenses are reimbursed with the next salary. (Payroll, Sales)
- Teams · **Sales channel** — Commission comes on top of the salary, see policy. (Sales)

Filtered from the permissions of the source, such as the file server’s file permissions with the groups from Active Directory, or the permissions in SharePoint. Check it in the live demo with a test account from your organisation.

### How filtering works

searchit adopts the permissions where you already maintain them.

1. **During indexing** — For every document, searchit stores who may open it in the source.
2. **At sign-in** — The user signs in via Active Directory, LDAP, Windows sign-in or an identity provider using SAML or OpenID Connect, with a second factor if required; searchit thus knows their groups.
3. **With every search** — The hits are filtered against these groups before the list is built.
4. **For your AI too** — AI assistants and the searchit AI Agent search with the permissions of the signed-in user.

## Organise entire result sets.

One tag on one hit or on all at once, and every tag becomes a filter.

- **By hand** — Public tags for everyone or private ones just for you, enabled per role, for one hit or the whole result set.
- **In hierarchies** — Tags break down into sub-terms across several levels, so you can map whole classification trees such as Contract › Supplier › Maintenance, each tag in its own colour.
- **During indexing** — Where set up, an AI assigns tags and categories and detects signals such as urgency or sentiment, all as filters.
- **On request** — Your AI assistant tags entire result sets, first as a preview, then on your word.

*Screenshot · test data*

- **Assign** — Apply a tag to all marked hits at once; the panel offers the existing tags.
- **Filter** — Every tag shows as a label on the hit and on the right as a filter; the tile chart shows how many hits carry it.

## Connect people, companies, cases.

Where configured, searchit extracts people, organisations, places and amounts from every document and links them across all sources into a knowledge graph, with the permissions of the source document.

- **In the document** — A dedicated tab in the preview shows the recognised entities and their relations; one click filters to all documents in which they occur.
- **One spelling** — Variants of the same company or person are merged. A curator confirms the suggestions instead of the system guessing.
- **Filter by value** — Amounts above a threshold or time ranges, not just exact strings.
- **Follow connections** — From the company via the relation to the other party and to every piece of evidence. Your AI assistant queries the same graph.

*Screenshot · test data*

- **In the document** — The “Entities” tab in the preview: companies, addresses, a discount and an amount from a supply contract, with the relations below.
- **Workspace** — The workspace shows each entity once, with the number of documents, type distribution and relations with their evidence.

- [Try the knowledge graph ›](https://searchit-agent.ai/#wissensgraph)

## Use every hit.

Look inside, narrow down, organise, analyse. Tap a tile and see how it works.

### Look inside instantly

*Preview with the match highlighted*

Documents, spreadsheets, images and emails appear in the preview, with the matches highlighted.

- With the searchit click service on the workstation, one click opens the file in its original program.
- On Windows, emails open in Outlook.
- On Windows and Linux, a file’s folder opens in the file manager.

### Read the summary

*AI summary, where set up*

Where a language model is set up, an AI reads along during indexing.

- A summary appears in the preview next to the document.
- Signals such as sentiment and urgency become filters.
- People, companies, places and amounts are recognised and become filterable.

- [From documents to knowledge ›](https://searchit-agent.ai/#gedaechtnis)

### Find text in scans

*Text recognition for scans and photos*

Text recognition makes scans, photos and image-only PDF files searchable.

- The scanned lease is found by its text.
- The recognised text appears in the text preview and can be copied from there.
- Here too, filtering follows the permissions of the source.

### Narrow down in one click

*Graphical filters and analyses*

Graphical filters by folder, author, file type, source and time: as a tree, bar, pie, topic map or timeline.

- Time ranges also relative, such as “the last 30 days”.
- Ad-hoc analyses across any filter and interactive dashboards.
- Search on from a hit, for example for all emails to the same recipient, in one click.

### Share searches and hits

*Saved searches, subscriptions, export*

What you have found once, you find again, and so do your colleagues.

- Save searches and publish them for colleagues; rerun earlier searches from the history.
- Subscribe to saved searches: new hits arrive by email.
- Result list as a CSV table, files as ZIP or straight to the clipboard.

### See what is stored twice

*Storage space, duplicates, history*

Whoever finds everything also sees what exists twice and what should have been deleted long ago.

- How much space each folder takes up, across file servers and mailboxes.
- searchit detects duplicate documents by their checksum.
- On request, earlier and deleted versions remain searchable per source.

## Keep every access request deadline in sight.

A GDPR access request concerns every repository. searchit collects the hits across all sources, tracks the deadline and has deletions approved by a second person.

1. **Create the request** — Record the requester and purpose. The one-month deadline runs from receipt; it can be extended once to three months with a justification.
2. **Collect hits** — Search for the person across drives, mailboxes and all other sources, and add all hits of a search to the request at once.
3. **Create the report** — The access report is created directly in the application, as a document to pass on. Passages in PDF and Office files can be permanently redacted beforehand.
4. **Approve deletion** — Deletion requests are made by whoever has the role for it. A second person must approve them: the four-eyes principle.
5. **Delete and document** — Deletion takes place in the file system and in Exchange mailboxes, with a confirmation prompt at the workstation. Every deletion request is kept in a history.

## Search every repository.

searchit comes with 100+ source connectors. Tap a source to see what searchit reads there.

### Files and drives

Network shares and file servers, local drives, HDFS, Google Drive and Nextcloud.

- Content, file name, path and metadata of every file.
- PDF, Office, text and CSV; images and scans via text recognition; archives such as ZIP, 7z or RAR.
- Filtered by the file server’s file permissions, with the groups from Active Directory.

### Email and calendar

Exchange, Exchange Online, IMAP, POP3 and PST archives.

- Emails including attachments, plus appointments and contacts.
- Attachments are read like any other file, scans included.
- On Windows, one click opens the email in Outlook.

### Microsoft 365 and collaboration

SharePoint, Microsoft Teams, Jira and wikis.

- SharePoint libraries with their documents and metadata.
- Teams channels and chats including attachments.
- Filtered by the permissions of the source.

### Document management

Alfresco, OpenText Livelink, IBM FileNet P8, Documentum, Meridio and systems with a CMIS interface.

- Documents including their metadata from the document management system.
- One search across document management, drives and mailboxes at once.
- Filtered by the permissions of the document management system.

### Databases

PostgreSQL, Microsoft SQL Server, Oracle, MariaDB and other SQL databases.

- Records from tables and queries become searchable like documents.
- A specialist database can be set up as a separate search module with its own filters.
- You decide which tables are included.

### Web and media

Websites, RSS feeds, Google News and YouTube.

- Websites with rules for which pages are included.
- Monitored sources as a saved search: new hits arrive by email as a subscription.
- A search box or an entire search page for your own website.

**File formats:** PDF, Office, emails, PST, text and CSV, images via text recognition, archives such as ZIP, 7z or RAR.

**And your business system?** We also connect web services and business systems, such as a specialist database as a separate search module with its own filters. Which of your systems are included, we check in the meeting.

## Choose your area.

The same search supports very different work. Choose an area: the image and example change with it.

*Example queries, invented*

### Service and IT: Find the fix from last time

Example query: „Error E-204 label printer“

- Earlier cases from the mailbox and from Jira.
- The manual from the drive, with the match highlighted.
- The tip from the Teams channel, with the permissions from Teams.

*Industries: IT · Retail*

### Law and finance: See the whole matter

Example query: „Liability matter Weber 2024“

- Briefs, files and emails on a matter, all together.
- Each file only for those allowed to open it.
- Tags per matter, on entire result sets at once.

*Industries: Lawyers · Tax advisers and auditors · Financial service providers · Healthcare · Human Resources*

### Public sector: Search paper too

Example query: „Zoning plan North Street“

- Scanned records spanning decades, via text recognition.
- A search for your own website, for citizens.
- GDPR access requests with deadline tracking.

*Industries: Public sector · Education · Archives and museums · Media and publishing*

### Industry and construction: Find plans and inspection reports

Example query: „Inspection report crane 3“

- Plans, contracts, inspection reports and delivery documents from all repositories.
- New inspection reports by email subscription.
- Overview of storage space and duplicates.

*Industries: Industry · Construction · Real estate · Energy and utilities · Transport and logistics*

In use at Allianz, Samsung, the Österreichische Gesundheitskasse and other organisations.

- [To the references ›](https://searchit-agent.ai/#referenz)

## Give your AI a memory.

searchit comes with an MCP server: ChatGPT, Copilot Studio or Claude Code use it to work in your collection, with the same permissions as the search. Choose a task.

- **With your permissions** — The assistant sees only what the signed-in user sees.
- **Enabled per role** — Roles in searchit decide whether an assistant has access and what it may change, such as applying tags.
- **More than search** — Analyse, look things up in the knowledge graph, tag, and open files on the workstation.

For your own applications, the search is also available as a REST interface (Representational State Transfer).

*Example run with invented data*

### Summarise

- You: Summarise the inspection report for Hall 3.
  - `search` “Inspection report Hall 3”: 3 hits, filtered by your permissions
  - `get_document` Inspection_report_Hall3_2026.pdf
- AI assistant: The report of 12 March 2026 lists two defects: a roller door closes with a delay, an escape route is blocked. Deadline for rectification: 30 April. (Source: Drive › Engineering › Inspection reports)

### Tag

- You: Tag all 2026 invoices from Müller GmbH with “Review”.
  - `search` “Invoice Müller GmbH 2026”: 14 hits
  - `bulk_tag` Preview: 14 hits would get “Review”, nothing changed yet
  - `bulk_tag` applied: 14 hits carry “Review”
- AI assistant: 14 invoices would receive “Review”. Shall I apply it? — Yes, apply
- AI assistant: Done. In searchit, “Review” now appears as a filter above the result list.

### Find connections

- You: Who handles our maintenance at Müller GmbH?
  - `graph_lookup` “Müller GmbH”: company in the knowledge graph
  - `graph_neighbours` Person Anna Berger, contract Maintenance contract 2026
- AI assistant: Anna Berger. She is named as the contact in the Maintenance contract 2026 and has been writing the invoice emails since May. (Sources: SharePoint › Contracts, Purchasing mailbox)

- [searchit AI Agent: the digital employee ›](https://searchit-agent.ai/)
- [searchit Chatbot for your website ›](https://searchit-agent.ai/chatbot/)
- [Try the knowledge graph ›](https://searchit-agent.ai/#wissensgraph)

## Pay by collection size.

The licence follows the number of objects in your index, not the number of your users or sources. Set your size.

| Objects in the index | Monthly subscription | Annual subscription |
|---:|---:|---:|
| 500,000 | €285 | €257 |
| 1,000,000 | €330 | €297 |
| 2,500,000 | €401 | €361 |
| 5,000,000 | €802 | €722 |
| 7,500,000 | €1,203 | €1,083 |
| 10,000,000 | €1,604 | €1,443 |
| 15,000,000 | €2,138 | €1,925 |
| 20,000,000 | €2,713 | €2,441 |
| 25,000,000 | €3,391 | €3,052 |
| more than 25,000,000 | on request | on request |

*Licence per month* — searchit indexes larger collections too; we quote the price for them on request.

- Any number of users and search queries
- All source connectors, no per-source fee
- Search, preview, filters, access checks
- Software maintenance
- Same price on-premises or in the cloud
- Monthly subscription: cancel monthly

Semantic search and AI enrichment: the feature is part of the licence; the provider’s model costs are shown separately in the calculator. Guide values, net plus VAT; the offer after measuring your collection is binding.

### Run the numbers for your collection

1. Enter objects and text volume per object
2. Choose semantic search, AI enrichment and deployment
3. Read off licence and model costs, optionally as an email request

- [Calculate the price for your collection](https://searchit-agent.ai/pricing/)

## Frequently asked questions

### Do we have to tag our documents first?

No. searchit reads the text of every document, plus file name, path and metadata. You assign tags in addition, by hand or, where set up, by the AI during indexing.

### Does every employee see everything?

No. Every hit is filtered by the permissions of its source, with every search. Anyone who may not open a file on the drive will not find it in searchit either.

### Where does searchit run?

searchit is a separate installation for your organisation: in your own data centre or on your own virtual machine in the cloud. Searching happens in the browser.

### How up to date are the hits?

The crawlers run on a schedule you set per source and pick up new, changed and deleted documents. Whether a document was picked up, and if not why, your IT sees per document in a dedicated search module.

### Does searchit also find text in scanned documents?

Yes, with text recognition. It makes scans, photos and image-only PDF files searchable.

### Do we get answers instead of result lists?

Where a language model is set up, searchit summarises the top hits into a short answer and shows a summary in the preview of every document. Detailed answers with sources come from your AI assistant via the MCP server or from the searchit AI Agent.

### Can we start small?

Yes. You first connect the sources that are searched most and add more whenever you want. Semantic search and AI enrichment can initially be limited to one department, one project or one document type.

## See your search live.

In about 45 minutes we show searchit on sources like yours and answer your questions.

- Name
- Company
- Phone (optional)
- Preferred date (optional)
- Which sources should become searchable? (optional)

- The session takes around 45 minutes and is held online.
- A person from sales gets back to you, usually the next working day.
- We use your details to arrange the appointment and for nothing else.
- The button opens your email program with the finished message. It is only sent when you send it.

---

- searchit AI Agent: <https://searchit-agent.ai/>

This version is generated from the page's language files (scripts/gen-enterprise-search-textfassung.py). The page itself is authoritative.