iGeneering

Since a while, I have been asking myself how I could use this shiny new technology, namely AI, to my advantage. As an engineer, I keep being bombarded with tasks that seemingly do not have much to do with engineering per se, and it feels more like something in between secretary work and a monkey job, with a sprinkle of unnecessary project management.

Let’s see which tasks I encounter daily that could be automated and kept out of my way to more exciting engineering challenges.

Interesting projects start with some hefty documentation that has to be understood and parsed through. As long as it is of internal origin, like product specifications and quotes, it would be nice to have an agent that could parse it through and highlight the most relevant information to me, but I can survive also without.

Much more interesting is when you deal with documents generated by clients. A nice example is a URS (User Requirement Specification). A URS is a compilation of all possible requirements regarding a machine or plant. What is challenging about it is that it is written in wording and form that is specific to each client; it is also kind of bloated with all the information that has to be there because of legal reasons or just because of the international nature of the business. It might also come packaged with layouts, other documents, and correspondence between project management and the client. To have an engineering AI tool that is able to parse through Word documents, Excel tables, and layouts, make sense of it, and highlight the most relevant information would be really lovely to have. It would really speed up the starting phase of any project. It would also make sure that any eventual special requirements and peculiarities do not remain overseen.

Another example where an engineering AI tool comes in very useful is just parsing, searching, and summarizing internal documentation. Every company accumulates, through time, experience and knowledge. A lot of it ends up in written form. From internal standards, process descriptions, and instructions to all kinds of notes in digital or digitized forms and project files, it is often fragmented between different departments. To be able to navigate through all of that just like chatting with a friend over morning coffee… That would be something.

With time, project information accumulates, and decisions are taken. Folders on drives are filled with documents and designs; so are my email inbox and chat history. The project is closed, my attention is already on the next challenge, and then… someone comes and asks: “Do you remember that project we had two years ago?”. Yes, I would really love to have an AI assistant that could parse through all that history and help me give a quick and clear answer.

Engineering encompasses many times also a lot of monkey jobs, as I love to call it. For example, you need to publish a new revision of a design. A lot of files do get generated and delivered in foreseen folders; a lot of people need to get notifications with personalized emails, workflows like approvals need to get started… I feel like I would need a whole secretariat just for me. It is something that should be easily managed by an AI agent without any issue.

Someone might argue that all of the above could be solved by better organization and a bit of discipline, and I agree, up to a certain point. My personal estimation is that with the AI technology that is available today, it might be easily possible to save something between 10 and 15 percent of engineering time without even optimizing existing processes.

But now, my casual reader kicks in my curiosity. Are you also involved in engineering?

Which are the tasks where AI agents might help you optimize your work? Tasks that you might find frequent, repetitive, time-consuming, and demanding.

Let me know in the comments below.