O&M Manual AI Search Demo

The AI reads and indexes your uploaded PDF using semantic search — it understands the meaning of your questions, not just the keywords. So asking 'how do I reset the high-pressure trip on the level 3 chiller?' will find content even if the manual phrases it as 'HP fault reset procedure' or 'compressor safety cutout reset sequence'. Every answer is grounded in the actual content of your uploaded document and references the section it came from.

This free demo supports PDF files up to 5MB — enough to trial the search on a single piece of equipment's O&M manual. The full PM Assist platform supports larger PDFs (up to 100MB per file), scanned PDFs with OCR, multiple files per building, and a wider set of document types including handover packs, drawings, fire risk assessments and compliance certificates.

The free demo allows 3 questions per session — enough to evaluate whether AI search delivers value on your specific kind of documentation. For unlimited AI questions across all your building documents, sign up for PM Assist; there is no per-question fee on paid plans, so FM teams use it as much as they need during fault-finding.

No. Your uploaded document is processed in memory for the duration of your session and is then discarded. It is not stored against an account, not used to train AI models, and not visible to anyone else. For secure persistent storage with per-building data separation, version control and role-based access, that's the paid PM Assist platform.

Accurate enough for FM and engineering decisions, because every answer is cited back to the section of your uploaded document it came from. You can verify any answer against the original PDF before acting on it. This is the deliberate difference from general AI like ChatGPT — the AI here only answers from your document, never invents content, and always shows you the source.

The free demo works best with born-digital PDFs. The full PM Assist platform runs OCR (optical character recognition) automatically when an uploaded PDF has no extractable text layer, so scanned paper documentation becomes searchable. A document that mixes scanned pages with digital ones is indexed from the text it already contains. Legacy O&M manuals from older buildings are a common reason teams move to the paid product.

Yes. The tool is a responsive web page that works on any modern browser including phones and tablets. Field engineers commonly use the paid PM Assist product on a tablet next to the equipment they're working on, with the manual already uploaded so they can ask questions in plain English without flipping through a 400-page PDF.

Ctrl-F only finds exact keyword matches. AI search understands the meaning of your question and finds relevant content even when the manual uses completely different wording. You can also ask multi-part questions ('what's the filter spec and how often does it need changing?') and get a single grounded answer that draws on multiple sections of the document.

ChatGPT can summarise or chat over a PDF but it can hallucinate — it may invent setpoints, fault codes or procedures that aren't in the document. PM Assist's AI is built for the FM use case where the wrong answer has real consequences (incorrect maintenance, voided warranties, safety risk). It only answers from the document, always shows the source, and refuses to speculate when the content isn't there.

Per-building document organisation, source-cited AI answers across every document in a building (not just one PDF at a time), team collaboration with role-based access control, OCR on scanned documents, support for drawings and compliance certificates, full audit logs, encryption in transit and at rest, and per-building data separation for multi-site portfolios. Paid plans include unlimited documents per building.

Most teams are searching their first building within an hour. Sign up takes a minute, create the building, drag and drop the O&M pack, and indexing typically completes in 5–15 minutes for a standard commercial property — even with scanned PDFs. No IT deployment, no installation, no plugin to integrate.

The application and database are hosted in Frankfurt, and document storage is in Stockholm. All customer data is stored in the EU. Data is held with encryption in transit and at rest, and per-building data separation. When you ask a question, the relevant excerpts from your documents are sent to OpenAI in the United States to generate the answer. This is the only routine transfer of your content outside the EU. It is covered by OpenAI's Data Processing Addendum, which incorporates the EU Standard Contractual Clauses and the UK International Data Transfer Addendum. PM Assist is built and supported in the UK. Documents are explicitly never used to train third-party AI models.