Softlandia background

Softlandia

Blog

How to 10x your productivity with onsite maintenance work - Private Generative AI example

See an example video of an engine maintenance assistant we created in 15 minutes with YOKOT.AI. Continue reading for a more detailed explanation of the workflow!

Industrial maintenance tasks can be complicated and take a lot of time

Automated manufacturing tools and processes are complicated to deploy and maintain. The productivity can be remarkable if your tuning & maintenance are on top level. And this is something where Private Generative AI can genuinely help. The challenges come from the complexity of tasks and the diversity of parameters, and it doesn't help that someone does the maintenance once a year.

  • If you do not do the tasks continuously, it's hard to remember all the details and dependencies

  • There are several risks and complications associated if you do not perform the tasks correctly

  • Risk to personal safety

  • Risk of malfunctioning or even breaking the equipment 

  • Productivity losses caused by poorly performed maintenance

  • Instructions are available, but documents are long, and information is scattered

Automated welding process boosts productivity

How Generative AI can help

You can use YOKOT.AI Generative AI as your assistant and tutor in the project. You can ask about the steps needed in the work, safety precautions, spare parts, and checklists.

How? You can achieve this in less than 30 minutes of preparation work with the following steps:

  1. Upload equipment user manuals, maintenance manuals, and other existing documents to YOKOT.AI

  2. Upload & crawl spare part catalogs (with text explanations)

  3. Upload safety guides and work instructions for the maintenance work

  4. Start asking for help interactively with your tablet, PC, or phone while you are onsite

Example case: Engine maintenance

Here is an example of what we did to create content for a maintenance assistant. First, we opened a new YOKOT.AI instance (available as a SaaS model; please book a demo from here) - the deployment takes a few hours.

Secondly, we searched for a few relevant documents from internet sources. In this case, we used a Briggs & Stratton lawn mower engine as an example. 

These are the sources we found in 15 minutes:

  • 2 x webshops selling spare parts for Briggs & Stratton engines

  • Briggs & Stratton original user manual (pdf)

  • Website having instructions and safety precautions for lawn mower maintenance tasks

  • Public web forum with chat & discussions talking about problems in the maintenance work and potential solutions 

YOKOT.AI source document listing

In an industrial case, for example, these materials could include notes from previous maintenance breaks and some emails/guidelines specific to the equipment in question.

YOKOT.AI will use these sources and the general knowledge of the OpenAI language models to answer your questions. The brilliance is that you can see the answer sources all the time. When you want to double-check the original document, you can quickly see where to look for the information.

What kind of help can I get?

Here are some examples of what we did with our Briggs & Stratton materials. We asked for basic steps & instructions on what to do overall in the annual maintenance.

Guidelines for maintenance task

We asked specific questions about carburetor replacement, including what parts we should have and where to look for them.

We also tried to get instructions in other languages - this worked like a charm. You do not need to worry about having mixed language in source documents; you can get the instructions and answers in your language.

You can book a demo with our experts if you want to learn more.