Anyone who truly works with IFC models knows this: the challenge is not “opening” a file, but understanding whether its contents are actually reliable.
In real-world projects, an IFC file is never a neutral object. It is the result of an export from an authoring tool, often influenced by BEP (BIM Execution Plan) conventions, custom properties, simplified hierarchies, or incomplete documentation.
The result? Operational questions — the ones needed for quantity takeoffs, checks, and decision-making — require time, experience, and a sequence of validations that few tools can handle in a structured way.
AskIFC: a tool to truly query IFC models
AskIFC was created to solve exactly this problem. It is a tool designed for advanced querying of IFC models, capable of transforming natural language questions into a structured and verifiable interpretation of the model’s information content.
It is not a simple chatbot that “describes the model.” Its approach is different: to provide a reliable answer, it must first reconstruct the conditions of truth within the model. Every answer therefore comes from a precise process that navigates entities, relationships, properties, and context, leading to final validation.
The result alone is not enough—the process that leads to it is just as important.
The real problem: IFC is not a list, but a graph
One of the most common mistakes is to think of IFC as a list of objects. In reality, an IFC model is a relational system.
A wall, for example, may exist as an IfcWallStandardCase, but it only becomes truly useful when it is correctly placed within the spatial structure of the project, linked to its type, enriched with coherent properties, and associated with reliable materials and quantities.
When even one of these relationships is missing, the outcome may appear correct but be operationally useless. This is often the case with surface calculations: if the data is not consistent or standardized, the result may be numerically precise but technically wrong.
A structured method for querying an IFC model
To avoid fragile answers, AskIFC adopts an approach based on macro-level validation areas. Instead of returning a direct result, the system breaks each request into a sequence of checks.
This means analyzing the model structure, verifying correct object classification, checking the presence and quality of properties, and evaluating the continuity of relationships between elements. Additional checks include geometry, positioning, quantity consistency, and proper material usage.
The outcome is a comprehensive reading of the model that turns a simple answer into a truly reliable one.
Real-world cases: what it means to properly query an IFC
How many spaces are there and what are their areas?
A seemingly simple question actually requires several steps. It is necessary to correctly identify IfcSpace entities, verify their containment within the model, and extract area data from the correct source. If the model contains duplicates, placeholders, or unassigned spaces, the result can be misleading. A robust system must also be able to flag such anomalies.
How many elements belong to multiple levels?
In this case, it is not enough to read the assigned storey. The logical containment must be compared with the geometry to distinguish between legitimately multi-level elements and actual modeling errors.
Do the doors have panic bars?
This information is not always located in the same place within the model. It may be associated with the instance, the type, or custom properties. To provide a reliable answer, the system must search across multiple levels, normalize conventions, and return a verifiable result while highlighting any exceptions.
Is it possible to extract a reliable BoQ from the model?
This question requires an overall assessment. It is not enough to extract data: completeness of quantities, consistency of units, and the presence of generic or proxy objects must all be verified. A correct answer, in this case, also includes the model’s limitations.
Why Properly Querying an IFC Model Is Essential
In real BIM workflows, models are used for concrete decisions: quantity takeoffs, validation, coordination, and asset management. For this reason, an effective querying system must not only provide useful answers, but also demonstrate their reliability.
AskIFC introduces exactly this discipline: reading the model as a complex system, breaking analyses into verifiable checks, and managing real-world exceptions. It is an approach that enables the practical use of artificial intelligence in BIM—transforming it from a generic aid into a truly operational tool.
