Goemaat uses AI to make calculation capacity scalable

‍Introduction

Strengthening craftsmanship with technology

Goemaat is a leading manufacturer of PVC and aluminum window frames and glass. With over a hundred employees, the organization works daily on a wide range of projects for professional clients.

Behind every project lies an extensive estimation process. For projects alone, multiple estimators work on assessing requests, drawings, and specification documents. In addition, Goemaat has a separate estimation team for dealers.

Drawings, in particular, require a great deal of attention. Floor plans and elevations must be carefully reviewed to determine which window frames are needed, their corresponding dimensions, and how they are ultimately incorporated into the estimate.

For smaller projects, this can easily take several hours. For large projects, simply detailing the window frames can take several days.

Goemaat therefore identified an interesting question:

What if AI could help perform this specialized work faster and more scalably?

‍

‍

Ambition

More estimation capacity without compromising on quality

For Goemaat, the use of AI is not just about saving time.

The organization primarily sees opportunities to better utilize the knowledge and capacity within the estimation department. AI can support experienced staff with large-scale projects while simultaneously helping new and junior colleagues reach high-quality results faster.

This is relevant because a large part of the department needs to be able to work with the solution eventually.

The ambition is to:

• analyze drawings faster and more consistently,
• automatically generate window schedules from floor plans and elevations,
• link information from drawings and specifications,
• provide better support for junior estimators,
• handle more projects with the same estimating capacity,
• keep human oversight as an integral part of the process.

Goemaat estimates that this way of working could create the capacity to handle many more projects annually.

The greatest opportunity lies not in document analysis alone. For Goemaat, AI must primarily excel at understanding the content of construction drawings.

‍

‍

Solution

From construction drawing to window schedule with AI

Within the Flawless Platform , the focus for Goemaat is therefore on the Drawing Agent.

The Drawing Agent analyzes floor plans and elevations, consolidating recognized window units into a structured overview. The ambition is to further develop this into a complete window schedule.

The AI assists in identifying and structuring information such as:

• window units and positions,
• quantities,
• dimensions,
• relevant properties,
• and ultimately details like opening directions.

An important next step involves combining information from different drawings. After all, a floor plan provides different information than an elevation. By linking this data, an increasingly complete picture of every window frame can be formed.

In addition to drawings, specifications and other project documents can also be analyzed within the same platform. This ultimately creates a single environment where information from various sources comes together.

The estimator remains firmly in the lead throughout this process. AI performs the preliminary analysis, while the expert checks, evaluates, and decides what is ultimately included in the estimate.

‍

‍

Results

From manual drafting to targeted verification

The benefits are most visible in the preparatory drafting work. The larger the project, the more repetitive work the Drawing Agent can take over.

‍

1. Significantly less manual drafting

Creating a window schedule can be a time-consuming process.

For a small project, manually processing drawings can take about two hours. By automating part of that work, Goemaat sees opportunities to quickly shave off half an hour to an hour.

For large projects, the difference becomes much greater.

In a recent project, for example, an employee spent about four days drafting window frames. For projects with hundreds or thousands of frames, the amount of repetitive work increases even further.

That is precisely where the strength of the Drawing Agent lies: AI performs an initial analysis, and the estimator shifts from manual drafting to verifying, correcting, and evaluating.

‍

2. Supporting junior estimators faster

Goemaat is not just looking at how many hours AI can save.

The organization also views the Flawless Platform as a tool to help less experienced colleagues deliver high-quality work more quickly.

A junior estimator doesn't have to extract all the information from extensive drawings and documents independently from day one. The AI helps surface relevant information and presents it in a consistent structure.

The expertise and oversight of experienced staff remain essential, but can be applied more strategically.

In this way, AI can contribute to an estimating department where knowledge is easier to transfer and new employees can get up to speed faster.

‍

3. Comparing drawings, specifications, and requirements

In addition to drawings, information from specifications remains relevant.

The Flawless Platform makes it possible to analyze both information sources within the same request. This allows for verification of whether information from drawings aligns with the requirements and specifications in the contract documents.

For Goemaat, this also represents a clear quality ambition.

AI output must be reliable and verifiable. When information is open to multiple interpretations, the estimator must be able to refer back to the source and judge for themselves what is correct.

‍

4. Handling more projects with the same capacity

When the preparatory work per project decreases, it creates new capacity within the estimating team.

Goemaat expects the platform to eventually be relevant for about 90% of the department. Based on initial estimates, the organization sees the potential to handle 10 to 50 additional projects annually with the same capacity.

This makes the business case broader than just saving hours.

More estimating capacity also means:

• responding to requests faster,
• being able to evaluate more interesting projects,
• losing less capacity to repetitive drawing work,
• making better use of commercial opportunities.

‍

‍

Impact

AI-powered support from inquiry to quote

For Goemaat, the value of AI lies in smarter support throughout the entire process, from inquiry to quote.

The Drawing Agent is a key application for Goemaat, but it doesn't stand alone. Within the Flawless Platform, multiple AI Agents can work together to support the entire process from inquiry to quote.

For instance, the Bid/No Bid Agent helps determine which inquiries are worth pursuing, the Risk Agent identifies key points of attention early on, and the Drawing Agent analyzes the drawings and window frames.

Further along in the process, the Quote Agent assists in drafting the proposal, while the Comparison Agent highlights differences and inconsistencies between documents, quotes, and contract materials.

This creates a workflow where AI supports the estimation team at multiple stages—from selecting the right projects to analyzing, quoting, and verifying.

For Goemaat, this provides the ability to process more inquiries, better support junior staff, and make the entire estimation process more consistent and scalable.

‍

‍

Next steps

From AI support to a connected estimation process

For Goemaat, the ambition doesn't stop at reading drawings.

The ideal scenario is for verified information from the Flawless Platform to flow directly into systems such as Pritex or Matrix. Integrations with other parts of the existing IT landscape can also play a role in this.

This creates a step-by-step workflow where information doesn't need to be manually re-entered every time: from request and drawing to a verified estimation basis and, ultimately, into the existing ERP landscape.

The Flawless Platform serves as the central foundation for this, where AI Agents run and new applications can be added.

In this way, Goemaat is exploring how AI can evolve from a smart assistant for drawings into a structural component of a scalable estimation process.