Frequently Asked Questions About INTERGATOR
How does INTERGATOR work? What role do hybrid search and AI play? How are corporate knowledge, permissions, and data sources integrated? Here you’ll find answers to frequently asked questions about INTERGATOR, its editions, and typical use cases.
INTERGATOR in general
What is INTERGATOR?
INTERGATOR is a platform for enterprise-wide search and the AI-supported use of enterprise knowledge. It makes information from connected data sources searchable and usable through a central interface.
Depending on the edition, users can work with traditional full-text search, semantic and hybrid search, as well as additional AI features such as Agentic Chat and AI tasks. INTERGATOR respects existing permissions and provides access to the underlying original documents.
INTERGATOR does not automatically replace existing business systems. Instead, it adds an overarching search and knowledge layer.
What problem does INTERGATOR solve?
In many organizations, relevant information is distributed across numerous systems, repositories, and applications. Employees therefore often need to know where a document is stored, which system contains it, or which search terms were used.
INTERGATOR brings research across connected data sources together in a common interface. This allows information to be found, narrowed down, reviewed, and reused across system boundaries.
The AI editions also add semantic and hybrid search methods. This means users can find information even when they do not know the exact term or formulate their information need as a natural-language question.
Which companies and organizations is INTERGATOR suitable for?
INTERGATOR is particularly suitable for organizations in which knowledge and documents are distributed across different systems and data sources.
Typical use cases arise in companies, public administrations, and other organizations with extensive information holdings, different business applications, and differentiated access rights.
INTERGATOR is especially relevant when information needs to be researched across systems or when traditional search is to be enhanced with semantic research and controlled AI features. The exact scope of functions depends on the edition, data sources, and configuration.
What distinguishes INTERGATOR from a traditional search engine?
Traditional full-text search is particularly strong when searching for specific terms, names, numbers, or known wording.
Depending on the edition, INTERGATOR extends this type of research with semantic and hybrid search methods. Hybrid search combines exact word matches with semantic similarity. This allows both short, precise search terms and longer natural-language or multilingual queries to be taken into account.
INTERGATOR also combines search with features such as facets, document preview, source access and, in INTERGATOR AI+, Agentic Chat and AI tasks.
Is INTERGATOR a chatbot?
No. INTERGATOR is not just a chatbot solution, but a platform for search, knowledge access, and AI-supported research.
INTERGATOR AI+ also provides Agentic Chat. It can interpret a request, use INTERGATOR search when needed, ask follow-up questions, and—depending on the language model—perform multiple research steps.
Chat is therefore one function within INTERGATOR. The foundation remains the combination of connected data sources, a search index, permissions, document access, and research capabilities.
Does INTERGATOR replace existing systems such as DMS, intranets, or business applications?
No. INTERGATOR generally does not replace existing source systems.
Document management systems, file repositories, intranets, business applications, and other systems remain the leading systems for their respective content and processes. INTERGATOR makes approved information from these sources accessible through a shared search and knowledge layer.
The original documents generally remain in their source systems. INTERGATOR enables cross-system access without automatically replacing those systems.
Search
What is enterprise-wide search?
Enterprise-wide search enables centralized research across multiple connected data sources. Instead of searching individual systems separately, users can start their research from a common interface.
INTERGATOR makes approved content and metadata from connected sources searchable and provides it in a common search index. Which information can be found depends on the connected data sources, access rights, and configuration.
This makes it possible to research information across systems without users necessarily having to know in which system or storage location it is stored.
What is the difference between full-text search and semantic search?
Full-text search focuses on specific words and their occurrence in documents and metadata. It is particularly suitable for known terms, names, numbers, identifiers, or precise wording.
Semantic search additionally considers the meaning of a query. This means documents can also be found when they describe a topic using different terms or wording.
Both approaches have different strengths. INTERGATOR therefore combines them in a hybrid search in the AI editions.
What is hybrid search?
Hybrid search combines lexical and semantic search methods.
INTERGATOR takes both exact word matches and semantic similarities between a query and documents into account. The results of both methods are evaluated together and merged into a single result list.
This makes hybrid search suitable for individual search terms, names, and identifiers as well as complete questions, longer descriptions, and multilingual queries. It is available in INTERGATOR AI and INTERGATOR AI+.
Why does INTERGATOR combine lexical and semantic search?
Both search methods are particularly well suited to different types of queries.
Lexical search is strong for precise terms, names, numbers, and known wording. Semantic search is useful when users describe their information need in more detail or do not know the exact terms used in the document they are looking for.
INTERGATOR therefore combines both methods. Users do not have to decide before every search which method is better suited. Short, precise inputs and longer natural-language queries can be processed within the same search logic.
Can I enter natural-language questions in INTERGATOR?
Yes. In INTERGATOR AI and INTERGATOR AI+, search queries can also be formulated in natural language.
Instead of entering only individual keywords, users can describe their information need as a complete question or a more detailed description. The semantic component of hybrid search takes the context and meaning of the query into account.
For conversational research with natural-language answers, INTERGATOR AI+ additionally provides Agentic Chat.
Does INTERGATOR also work with short search terms, names, and numbers?
Yes. The improved hybrid search in INTERGATOR is designed to support both short and detailed queries.
Specific names, identifiers, invoice numbers, or individual technical terms benefit from the lexical component of the search. Longer and more topic-oriented queries can additionally be evaluated semantically.
This makes the same search interface suitable for very different types of research.
Does INTERGATOR support multilingual search queries?
Yes. INTERGATOR AI and INTERGATOR AI+ support semantic and hybrid search methods that can also handle multilingual queries.
A query therefore does not necessarily have to be formulated in the same language as the relevant document. Semantic search considers the meaning of the query and can identify content in another language as relevant.
The actual quality depends, among other things, on the models used, the indexed content, and the configuration.
How can I narrow down a large result set?
INTERGATOR provides facets, filters, and perspectives for this purpose.
Facets can narrow results by data source, file type, time period, category, author, or other criteria. Which facets are available depends on the selected perspective and project configuration.
Perspectives can also define which data sources are searched and which filters or result information are displayed. This allows large result sets to be progressively narrowed down to the relevant content.
How does INTERGATOR decide which results are shown first?
Search results are sorted by their calculated relevance by default.
With hybrid search, results from lexical and semantic search are merged and evaluated together. Both exact word matches and semantic proximity therefore influence the result order.
Depending on the configuration, additional sorting options may be available, for example by date, name, or size.
Why can a document be found even if my exact search terms do not appear in it?
This is a typical effect of semantic search.
It considers not only identical words but also semantically related wording. A document can therefore be relevant even if it uses different terms from the query.
In INTERGATOR AI+, an LLM-supported passage display can additionally help make the relationship between the query and the result visible. The displayed text passages still come from the actual document content.
AI and Agentic Chat
What is Agentic Chat in INTERGATOR?
Agentic Chat is the advanced chat function in INTERGATOR AI+. It analyzes the user’s request and, depending on the context, decides whether to answer directly, use INTERGATOR search, or first ask a follow-up question.
Depending on the language model used, the chat can also perform several search steps, compare results, and create an answer from them. This makes it particularly suitable for more complex research tasks in which information is distributed across multiple documents.
How does Agentic Chat differ from a traditional RAG chat?
A traditional RAG chat often follows a fixed process: first a search is performed, then the retrieved content is passed to a language model and used to generate an answer.
Agentic Chat in INTERGATOR works more flexibly. The language model first decides whether a search is required at all. It can also ask follow-up questions and, depending on the model, perform multiple research steps.
This means search is no longer necessarily the first step in every interaction, but a tool the chat uses when needed.
Does Agentic Chat access enterprise knowledge?
Yes. Agentic Chat can use INTERGATOR search to access connected enterprise information that is available to the user.
If a request requires research in enterprise knowledge, the chat can use search to identify relevant documents and evaluate their content for the answer.
Which information can actually be used depends on the connected data sources, permissions, and configuration.
Can Agentic Chat also use general model knowledge?
Yes. Not every request requires a search in enterprise knowledge.
For general questions, simple instructions, or other suitable situations, the language model can also answer without retrieval. Enterprise knowledge nevertheless remains a central part of INTERGATOR and can be incorporated into research via search when appropriate.
Whether a search is performed depends on the request, context, model, and configuration.
Does Agentic Chat decide whether a search is necessary?
Yes.
The search function is available to the language model as a tool. The model decides based on the context whether a search is required and which query should be used.
This can avoid unnecessary searches and allows follow-up questions to be handled more effectively within the context of an ongoing conversation.
Can Agentic Chat perform multiple research steps?
Yes, if the language model used supports this capability.
Reasoning models in particular can break a more complex question into several steps, perform multiple searches, and then combine or compare the results.
Multi-step research can therefore enable more comprehensive answers, but it generally requires more time and computing resources.
What are reasoning models?
Reasoning models are language models that can break complex tasks down into several processing steps.
In Agentic Chat, they can first interpret a question, then perform one or more searches, evaluate the results, and derive additional research steps from them.
This can support more demanding research tasks. At the same time, response time and resource requirements may be higher than with simpler language models.
Can Agentic Chat ask follow-up questions?
Yes.
If a request is not sufficiently clear, the chat can first ask a follow-up question before generating an answer or performing a search. This allows users to narrow down, for example, the relevant time period, topic, or research context.
How often follow-up questions are asked also depends on the language model used and its configuration.
What are chat scopes?
Chat scopes limit the knowledge space that the chat should use for a conversation.
Depending on the configuration, selected documents, lists, favorites, or other defined document sets can serve as a scope. This makes it possible to deliberately restrict a conversation to a specific information set.
Selecting a chat scope is not mandatory. It remains available when research should be intentionally limited.
How are sources displayed in chat answers?
When the chat uses content from INTERGATOR search, the underlying documents can be displayed directly in the answer as inline references.
In addition, a document list can show the sources that were used. Users can move from the answer to the document preview or original document and verify relevant statements.
These source references improve the transparency and traceability of generated answers.
Does an answer without sources mean that no enterprise information was used?
Not necessarily in every case. However, an answer without references should not automatically be interpreted as source-based.
The chat can also use general model knowledge when no search is necessary. Source references, by contrast, transparently show which documents were used for an answer.
For professionally important statements, it is therefore advisable to review the underlying original documents.
Is INTERGATOR an autonomous AI agent?
No. INTERGATOR is not an autonomous business-process agent.
Agentic Chat can organize its research flexibly, use INTERGATOR search as a tool, ask follow-up questions, and—depending on the model—perform several search steps.
In the current scope, INTERGATOR search is essentially the main tool available to the chat. This does not imply arbitrary external actions, changes in third-party systems, or autonomous business decisions.
Which tools can Agentic Chat use?
INTERGATOR search is the main tool used by Agentic Chat.
Using this search tool, the chat can identify relevant documents and use their content for its answer. Additional tools for facets, metadata, external systems, or actions are not generally part of the standard feature set at present.
Such extensions should therefore not be presented as generally available standard functionality.
Are chat answers displayed in full immediately?
No. INTERGATOR supports streaming.
The answer is displayed while it is being generated and grows progressively. Users can start reading before the complete answer has been finished.
This primarily improves perceived response time for longer answers and more complex research.
Data Sources and Permissions
Which data sources can INTERGATOR search?
INTERGATOR can make information from different connected data sources searchable and accessible through a common search interface.
Depending on the project, these sources can include file repositories, document management systems, email systems, intranets, Microsoft 365 environments, collaboration platforms, databases, or business applications.
Which sources are actually available depends on the specific installation, connectors, and project configuration. INTERGATOR does not automatically access every system in an organization.
Can INTERGATOR search information across multiple systems?
Yes. This is one of the central benefits of INTERGATOR.
Information from different connected sources can be researched in a shared search context. Users do not necessarily need to know whether a relevant document is stored on a file server, in an intranet, or in another business system.
The original systems remain in place. INTERGATOR adds an overarching search and knowledge layer.
Do the original documents remain in their existing systems?
Yes. Original documents generally remain in their respective source systems.
For search and analysis, INTERGATOR transfers relevant content and metadata into its own search index. Original documents are not automatically moved, changed, or deleted as part of indexing.
From the result list or document preview, users can return to the original document, depending on the data source and their permissions.
Does INTERGATOR respect existing access permissions?
Yes. INTERGATOR is designed for permission-aware use of enterprise information.
Users should only be able to find and open content for which they have the required rights. The specific implementation depends on the source system, identity and permission structure, and project configuration.
As a result, different users can receive different results for the same query.
Do all users see the same search results?
No.
Which results a user sees can depend, among other things, on their access rights. Selected perspectives, active filters, connected data sources, and the configuration can also influence the result.
A document for which a user lacks sufficient permissions should not become accessible merely because another user found or linked it through INTERGATOR.
Does a shared INTERGATOR link extend access permissions?
No.
Sharing a search or document does not change existing access rights. A recipient can only open content if they independently have the required permission.
The same principle applies to favorites, lists, and other saved references within INTERGATOR.
What information is indexed for search?
INTERGATOR can prepare both document content and metadata for search.
This can include, for example:
– title and file name
– author or creator
– storage location
– document type
– creation or modification date
Which metadata are actually available depends on the data source and configuration.
Are new or changed documents immediately available in INTERGATOR?
Not necessarily.
Connected data sources are updated at configured intervals. New, changed, or moved content may therefore only appear in the search index after a subsequent update cycle.
How quickly changes become visible depends on the data source, connector, and project configuration.
Why might an expected document be missing from INTERGATOR?
There can be several reasons.
For example:
– The relevant data source is not connected to INTERGATOR.
– The document has not yet been re-indexed.
– The user does not have the required permissions.
– A filter or perspective excludes the content.
– The content is not approved for the relevant search.
A missing result therefore does not automatically mean that the document does not exist in the organization.
Can INTERGATOR connect Microsoft 365, file servers, intranets, or DMS platforms?
Such systems are among the typical data sources that can play a role in INTERGATOR projects.
However, the specific systems, versions, and functions supported must be checked for each project. INTERGATOR includes, for example, a modernized SharePoint connector based on the SharePoint REST API. Other connectors and their feature scope should only be committed according to the currently confirmed technical documentation.
Can Agentic Chat access data for which the user has no permission?
INTERGATOR is designed to take the user’s permission context into account when using AI-supported research.
Agentic Chat uses INTERGATOR search to research enterprise information. It should therefore only be able to include content that is available to the respective user through the underlying search.
The specific permission integration depends on the source system and project.
Are enterprise data automatically transferred to external AI services?
No, this cannot be stated as a general rule.
Which language models and providers are used, and which data are processed or transferred, depends on the specific operating model and project configuration.
Different model and deployment variants may be used depending on the project. Data flows, providers, and data-protection requirements should therefore be clarified specifically for each customer scenario.
Editions and AI features
Which INTERGATOR editions are available?
INTERGATOR is available in three editions:
- INTERGATOR Core
- INTERGATOR AI
- INTERGATOR AI+
The editions build on one another functionally. INTERGATOR Core provides traditional lexical search, INTERGATOR AI adds semantic and hybrid search, and INTERGATOR AI+ additionally provides LLM-based features such as chat and AI tasks.
What is the difference between INTERGATOR Core, AI, and AI+?
INTERGATOR Core provides the core functions for enterprise-wide search, with a focus on lexical full-text search for terms, names, numbers, and known wording.
INTERGATOR AI adds semantic and hybrid search. This allows natural-language, longer, and multilingual queries to be taken into account.
INTERGATOR AI+ additionally includes LLM-based features such as Agentic Chat and AI tasks. The exact scope depends on the configuration and models used.
Which edition includes hybrid search?
Hybrid search is available in INTERGATOR AI and INTERGATOR AI+.
It combines lexical and semantic search methods, allowing both short, precise terms and longer natural-language or multilingual queries to be processed.
INTERGATOR Core uses lexical full-text search.
Which edition includes Agentic Chat?
Agentic Chat is part of INTERGATOR AI+.
It can interpret a request, use INTERGATOR search when needed, ask follow-up questions, and—depending on the language model—perform multiple research steps.
Agentic Chat is not part of INTERGATOR Core or INTERGATOR AI.
Which AI features are available in INTERGATOR AI+?
Core LLM-based features in INTERGATOR AI+ include, among others:
– Agentic Chat
– chat scopes
– sources and inline references
– questions about an open document
– AI tasks in the result list and document preview
– free-form questions and instructions as an AI task
– reusable pinned AI tasks
Which features are actually available depends on configuration, data sources, permissions, and the language model used.
What are AI tasks in INTERGATOR?
AI tasks are predefined or freely formulated instructions to a language model that can be applied directly to documents.
Typical tasks include:
– summarizing documents
– identifying main topics
– extracting facts
– checking for specific statements
– asking a free-form question about a document
Tasks can be executed in the result list or document preview.
How do AI tasks work in the result list?
In the result list, an AI task is applied to the individual results being processed.
A separate result is generated for each document. This makes it possible, for example, to review several documents against the same criteria or summarize each one separately.
An AI task does not automatically create a single combined analysis of the entire result list.
How do AI tasks work in the document preview?
In the document preview, the selected AI task is applied to the currently open document.
Users can, for example, generate a summary, extract facts, or ask a specific question about the content. Depending on the task, existing comments can also be taken into account.
Can custom questions or instructions be used as an AI task?
Yes.
A free-form instruction or question allows users to formulate their own task and apply it to the relevant document context.
In the result list, the question is applied to individual results. In the document preview, it refers to the currently open document.
Can frequently used AI tasks be saved?
Yes.
Frequently used tasks can be pinned and then reused directly at the relevant point of use.
Free-form questions or instructions can also be pinned together with their text. This allows recurring review or analysis tasks to be run again more quickly.
Can AI tasks be filtered?
Yes.
The list of available AI tasks can be filtered by title and description. This makes selection easier, especially when many different tasks are configured in an installation.
What is the difference between AI tasks and INTERGATOR Assist?
AI tasks and INTERGATOR Assist follow different approaches.
AI tasks are applied directly to search results or an open document. They are suitable for individual analyses, checks, summaries, or free-form instructions.
INTERGATOR Assist goes further by representing defined business workflows with structured inputs, specified data sources, and concrete output formats.
An AI task is therefore not automatically an INTERGATOR Assist.
Are all AI features available in every installation?
No.
The exact scope of features depends, among other things, on:
– the edition used
– the configuration
– the language models used
– the connected data sources
– permissions
– project-specific settings
Available AI tasks, chat features, or models can therefore differ between INTERGATOR installations.
Operations, Integration and Security
Can INTERGATOR be operated on premises?
Yes. INTERGATOR can be operated in on-premises scenarios.
The specific architecture, infrastructure, and components used depend on the edition, operating model, and project requirements. For binding statements about system requirements, hardware, or deployment, the current technical documentation should be consulted.
Can INTERGATOR be integrated into existing IT environments?
Yes. INTERGATOR is designed to connect existing information sources and systems within a shared search and knowledge environment.
Depending on the system, integration is implemented using connectors, crawlers, APIs, or other import mechanisms. Which integrations are specifically possible must be checked for the individual project.
INTERGATOR does not automatically replace existing systems, but adds an overarching search and knowledge layer.
What role do connectors play in integration?
Connectors establish the link between INTERGATOR and the respective data sources.
They can provide content, metadata, and—depending on the source system—permission information for processing in INTERGATOR.
Which connectors are available and suitable for a particular system version depends on the product status and project. Specific integration commitments should therefore always be checked against the current technical documentation.
Can INTERGATOR work with different language models?
Yes. INTERGATOR can use different large language models for different tasks.
For example, a model that supports more complex multi-step research can be used for chat, while a faster model may be more suitable for other tasks.
This makes it possible to balance requirements for quality, speed, resource consumption, and cost according to the use case. Which models can specifically be used depends on the technical environment and configuration.
Do enterprise data have to be transferred to an external AI provider?
Not necessarily.
Which language models and providers are used, and which data are processed or transferred, depends on the operating model and project configuration.
Different model and deployment variants can be used depending on the project. Data flows, providers, and requirements for data protection and information security should therefore be assessed specifically for each deployment scenario.
Do all data always remain in the organization's own data center with INTERGATOR?
This cannot be stated as a general rule for every installation.
INTERGATOR can be used in different deployment and model scenarios. Whether data are processed exclusively in an organization’s own infrastructure or whether external services are involved depends on the specific configuration.
A project should therefore transparently define where components are operated, which models are used, and which data flows result from this.
How does INTERGATOR address information security?
INTERGATOR is designed for controlled access to enterprise information. This includes, in particular, taking permissions into account when accessing connected content.
Technical components, base images, and libraries are also regularly updated and known vulnerabilities are evaluated. Logging, error messages, and operational robustness are also continuously improved.
The specific security architecture depends on the operating model and customer project.
How does INTERGATOR handle known security vulnerabilities?
Known vulnerabilities in relevant components and images are monitored and evaluated.
When new vulnerabilities are identified, they can be assessed for their actual relevance to the deployed environment and appropriate measures can be derived.
The aim is not only to improve technical currency but also to enable a traceable assessment of security issues during operation.
Does INTERGATOR also respect existing user permissions for AI features?
INTERGATOR is designed to respect the user’s permission context during AI-supported research as well.
Agentic Chat uses INTERGATOR search to research enterprise information. It should therefore only include content that the respective user is allowed to access through the underlying search.
How permissions are technically connected and evaluated depends on the source system and project configuration.
How does INTERGATOR support the safe use of AI-generated results?
INTERGATOR links AI-generated answers to the underlying information sources.
Inline references, document lists, passages, and access to original documents help users trace and verify statements.
These mechanisms improve transparency and verifiability, but they do not guarantee error-free AI answers. Professionally or commercially relevant statements should still be checked against the original sources.
What are the technical requirements for INTERGATOR?
The technical requirements depend on the edition, architecture, operating model, models used, and project requirements.
These can include requirements for servers, operating systems, deployment, networks, storage, or additional resources for AI models.
Because these requirements can evolve technically, specific values should not be fixed on a general FAQ page. The current technical documentation is authoritative for project planning and operation.
Is INTERGATOR suitable for Kubernetes environments?
Kubernetes can be part of corresponding INTERGATOR deployment scenarios.
However, which platforms, versions, and deployment variants are specifically supported should always be checked against the current technical documentation and project status.
For a general website FAQ, specific Kubernetes versions or binding platform commitments should therefore be avoided.
How are errors and technical issues made traceable during operation?
INTERGATOR includes functions and improvements for logging, error messages, and operational diagnostics.
The aim is to make technical issues easier to understand and to support analysis by administrators, project teams, and support staff.
The specific scope of available diagnostic and operational functions depends on the installation and system environment.
What does digital sovereignty mean in the context of INTERGATOR?
Digital sovereignty in the context of INTERGATOR primarily means that organizations can retain influence over key technical decisions—for example the operating model, data flows, model usage, permissions, and infrastructure.
It does not automatically mean that every installation is operated entirely locally or that external services are never involved.
The appropriate setup is determined according to the project’s functional, technical, data-protection, and organizational requirements.