Frequently Asked Questions

Engineering organisations sit on thousands (sometimes millions) of legacy drawings that are hard to search, inconsistent, or poorly structured. DrawingHub extracts structured metadata, reduces duplication, and improves findability so your DMS becomes usable again — without replacing it.

No. DrawingHub is designed to augment existing systems like ProjectWise or similar DMS platforms. The goal is to improve data quality and search outcomes, not force a migration.

No. DrawingHub runs alongside existing processes. Most teams start with a limited dataset or engagement to validate outcomes before scaling.

It highlights the data quality issues that affect finding the right drawing at the right time (e.g. invalid or inconsistent titles) so you understand why drawings are hard to locate today.

Security and data sovereignty are core design principles. Processing environments, access controls, and auditability are aligned to meet enterprise engineering requirements.

Yes. Multi-model AI ensures trusted extraction of data, even on inconsistent title blocks, scanned PDFs, and decades of historical drawings.

Most customers begin seeing insights during the initial assessment phase, often through the findability analysis before full implementation.

We can detect when an AI extraction isn't quite right and flag it for human review.

We built this model while operationalising AI inside a real product company. It helped us make sense of what was working, what wasn't, and where effort actually mattered. We're sharing it publicly because many teams are asking similar questions — not because we're offering assessments or advisory services.

No. Organisations rarely move in a straight line. Some teams will experiment deeply before leadership fully aligns, while others will establish governance early. Think of this as an orientation map rather than a linear playbook.

That's normal. Most businesses are navigating conflicting advice and rapid change. The early stages of the model focus on literacy, curiosity, and small experiments because progress usually starts with understanding — not large programmes.

Not necessarily. Many organisations learn faster by running focused pilots tied to real work. Strategy tends to become clearer once teams see where AI actually improves outcomes rather than just generating excitement.

Mastery isn't a finish line. It simply describes a state where AI becomes a normal part of how work gets done — supported by leadership, culture, and governance. The real goal is sustainable progress, not perfection.

DrawingHub exists because we believe operational AI should solve practical problems, not create more complexity. The framework reflects the same philosophy — start with real workflows, build trust gradually, and focus on outcomes people can see.

We've learned that trust grows when governance evolves alongside experimentation. Waiting until later stages often creates friction or slows adoption, especially in security-conscious industries and infrastructure.

The ideas apply regardless of size. Smaller teams often move through stages faster because decisions are closer to the work. The key difference isn't scale — it's intentional learning.

That's common. A company might be advanced in experimentation but still early in leadership alignment or workforce adoption. The framework is meant to show patterns, not assign a single score.

Some teams use it as a conversation starter during planning sessions or retrospectives. Others use it to explain AI progress to leadership or to ground discussions that might otherwise feel abstract.

DrawingHub uses AI-based metadata extraction — not OCR. Multiple models work together to read title-block fields such as drawing numbers, titles, and revisions across scanned PDFs, legacy prints, and inconsistent formats.

Our pricing isn't one-size-fits-all because every project is different. Engineering drawing environments vary widely in data quality, file types and formats, volume and scale, and level of organisation (or disorganisation). Instead of offering generic pricing, we tailor solutions based on your specific situation to ensure accuracy and real value.

Pricing is based on the actual scope of work required, not just the number of files. Before proposing anything, we understand your workflow and challenges, review a sample of your data, identify which files are actual engineering drawings vs. irrelevant documents, and assess complexity (e.g. scan quality, revisions, formats). This allows us to give you a fair, transparent quote aligned with the real effort required.

No. Many datasets contain a mix of true engineering drawings, duplicate files, and unrelated documents. We focus on processing only the relevant drawings, so you're not paying for unnecessary work.

Yes. We typically start with a pre-engagement analysis or pilot using a subset of your data. This helps demonstrate real outcomes, validate accuracy, and refine scope before full deployment.

During this phase, we analyse your 'messy' dataset, classify and isolate relevant drawings, evaluate data quality and structure, and identify potential challenges early. This ensures the full project is well-scoped, predictable, and successful.

Yes — we can provide a high-level estimate based on number of files, expected drawing ratio, and complexity indicators. A precise quote is provided after reviewing sample data.

Simply reach out and share a sample of your dataset or describe your challenge. We'll review your data, run an initial analysis, and provide a tailored proposal. Ask us anything if you're not sure.