Maintenance assistant
Plain-language questions about your assets, your plans and your history. It answers based on your documentation and your data, not on industry generalities.
Artificial intelligence
AI that answers about your maintenance, not about the world
A general assistant knows a lot about maintenance and nothing about your installation. GMAO CLOUD's is the other way round: it answers about your assets, your manuals, your preventive plans and your history, because that's the only thing it has read.
The difference is in the corpus
What makes AI useful inside a CMMS isn't the model, which is the same one anyone can access: it's what it knows about you. GMAO CLOUD indexes your equipment manuals, your checklist templates, your preventive plans and your intervention history, and searches there before answering. That's why a question like "what was done to the chiller at the Valencia site last year" has an answer, and why that answer doesn't exist outside your installation.
What it does today
Concrete features, not a promise that AI will fix everything.
The assistant, in the field too
The same assistant inside the technicians app, to check on the go what a piece of equipment's manual says or what was done last time.
Document reading
The report a supplier sends you as a PDF gets read and turned into data: supplier, asset, hours, materials. Stop typing what's already written.
Help with order text
A technician writes the solution in a hurry, on site. The AI suggests a clearer wording based on what he wrote, and he decides whether to use it.
Translation of your catalogues
Your own terms — asset types, statuses, families — translated into the languages you use, for review and saving.
Questions about your figures
Orders, hours and costs aggregated over the period you ask for, written up as an answer instead of a table you have to interpret.
Switched on by you, switched off by you
AI is not a layer forced onto you with the product.
Each feature, its own switch
The assistant, indexing, document reading and text help are each turned on separately. Most come switched off by default.
Scope by profile
The assistant respects the permissions of whoever is asking: it doesn't show what that person couldn't see browsing the system.
Your corpus, in your database
The indexed documents, their search indexes and the conversation threads are stored in your installation's database. It's an architecture rule, not a configuration option.
Proposes, doesn't execute
The AI features are for queries: they answer and suggest. They don't close orders, don't assign work and don't change anything on their own.
What the system decides and what you decide
It's the question worth asking any vendor talking about AI, so we answer it plainly. GMAO CLOUD's artificial intelligence features don't make decisions: they read, answer and propose, and whatever gets applied is applied by a person. There are automations that write on their own — assigning a job to a technician following rules you've configured, for example — but those are deterministic rules you decide, not a model: they can be switched off, they log every decision with its reason, and they are reversible. Telling those two things apart matters, and blending them to sound more modern would do you a disservice.
Shall we look at it with your way of working?
Leave your details and we will get in touch to see whether we fit. No commitment, no lock-in period.
What you need to answer to your audit
We don't sell a seal of approval. We give you what it takes to document it.
Inventory of AI features
What features exist, what data each one handles and for what. Maintained as product documentation, not rebuilt whenever someone asks.
Identified as AI
The assistant appears labelled as such in the interface, in the product's languages. No one thinks they're talking to a person.
Usage log
Queries and consumption by day, user and origin, so you know what's actually being used and can scope it.
Reversible
Whatever an automation writes is flagged for review, with its authorship and an action to undo it.
Where the limit is
GMAO CLOUD's AI is only as good as what it has read. If your manuals haven't been uploaded and your reports are closed with "fixed", the assistant has nothing to draw a useful answer from, and no model fixes that. It also doesn't plan routes or decide your maintenance plan: it doesn't do that, and we'd rather say so than hint otherwise. What it does solve well is the question that today forces someone to dig through folders and someone else to remember something that happened two years ago.
Frequently asked questions
What happens if the AI misreads a document?
Someone corrects it, because whatever it proposes always stays subject to review by a person before it becomes a system record. It's set up that way on purpose: the point of these features is to save the work of typing and searching, not to replace the judgement of whoever is responsible for maintenance.
Do I have to use AI to use GMAO CLOUD?
No. The artificial intelligence features are optional and each one is switched on separately; most come off by default. The product works exactly the same with none of them turned on.
Where are my documents and my conversations with the assistant stored?
In your own installation's database: the indexed corpus, its search indexes and the conversation threads. It's a product architecture rule, not an option you have to configure.
Can the AI close an order or assign work on its own?
No. The AI features are for queries: they answer and propose, and whatever gets applied is applied by a person. There are automations that write on their own, such as rule-based assignment, but those are deterministic rules you configure, which can be disabled and leave a reversible log.
Does it evaluate my technicians?
No. The assistant can group orders and hours by technician when asked, just as a report would, and the answer warns that it's aggregated activity data, not a performance evaluation. There is no automatic scoring of people.
Is it useful if our documentation is incomplete?
Less than it should be, and it's worth knowing that up front. The assistant answers based on what's indexed: with manuals not uploaded and reports closed without a description, the answers will be poor. It usually pays to start with critical equipment rather than everything at once.
Try it with your documentation
In the demo we index a couple of your manuals and ask about your equipment, which is the only way to see if this works for you.