# AI in Manufacturing: Where to Start Without an IT Department > Manufacturing company owners without IT staff can kickstart AI adoption by focusing on routine operations with clear algorithms. The first step is an audit of time-consuming tasks, selecting three… - Author: Yaroslav Maxymovych (Founder & CEO, AI Advisory Board) - Published: 2026-09-18 - Updated: 2026-11-02 - Source: https://aiadvisoryboard.me/blog/ai-in-manufacturing-where-to-start-without-an As a manufacturing company owner without an in-house IT department, you're likely hearing the buzz around AI and asking a simple question: Where do I start to avoid throwing money at something that won't work? The answer is straightforward: begin with routine operations that consume valuable time from operators, foremen, or logistics staff, but follow a clearly defined algorithm. This isn't about futuristic factory visions; it's about concrete actions you can take as early as next week. ## TL;DR - Start with a routine audit: Identify three daily operations that follow a consistent script. - Choose processes with clear inputs and outputs. If you can evaluate the result in 5 minutes, it's a good candidate. - Initial automation should be launched by a non-coder, otherwise you risk getting stuck at the demo stage. So, how do you find these processes? The easiest way is to ask each shift supervisor to name five tasks that consume most of their working time. You can do this in a 15-minute meeting or via a simple form. Once you have a list of 15–20 items, work with your supervisors to select three that: - Occur at least three times a week. - Have clear input data (form, spreadsheet, photo, voice message). - Produce an output that can be used further (request, report, calculation). Here's a real-world example: In one manufacturing company, an operator spent up to an hour on each commercial proposal. This involved gathering data from three systems, formatting text, and verifying prices. After just one session, they learned to articulate the logic verbally, and AI generated a draft in minutes. This was a turning point: the operator now completes five proposals in the time it previously took to do one. ### Checklist for First Steps 1. Conduct a 15-minute session with supervisors: Each names 5 time-consuming tasks. 2. Consolidate lists and identify common items. 3. Select three tasks with a clear algorithm (if you can't explain it to a new hire in 2 minutes, it's not suitable). 4. Run a test with non-sensitive data (sample spreadsheet, test photo). 5. Document the results: Compare time spent before and after automation. If you're unsure whether AI can process your data, start with the simplest: converting text from one format to another. For example: - Voice message from a driver about a delay → text notification for the dispatcher. - Photo of a faulty label → extraction of batch number and date. - Excel spreadsheet with a shift schedule → raw material procurement request. These are all examples where access to ERP or MES isn't required. A browser and access to data your employee already uses are sufficient. > **Definition:** Routine is an operation performed according to a fixed scenario more than twice a week, requiring no creative decisions. > **Definition:** Working automation is a tool that executes an agreed-upon scenario using company data within its existing tools, without continuous vendor support. ## FAQ **Do we need programmers for initial automation?** No. In our program, participants describe the logic in plain language, and AI writes the code. The first micro-automation is launched in the browser as early as the second session. **Is it safe to use production data with external services?** For training, we use test or anonymized data. When working with real processes, an NDA can be implemented upon request. **How will we know if the automation is working?** Before starting, we formally define the criteria for "working": what scenario, what data, what result. This isn't a demo; it's a tool that performs on your data within your systems. **What's the cost for a 30-person company?** A corporate intensive for two groups of up to 20 people costs 99,999 UAH. This includes training, video course, template specifications, and the first month of support for the created automations. **Do we own the tools created?** Yes. The code and created automations are the full property of your company. They operate within your tools, with no vendor lock-in. **How does this relate to our organizational chart?** The first step is to build your company's organizational chart. It reveals where routines are concentrated and helps the owner see the company from above: what can already be delegated to AI agents. A free tool for this is available via the link. **How do we measure the results of the first automation?** Compare the time taken to complete the task before and after. For instance, if an operator spent 40 minutes on a proposal and now completes it in 5 minutes, that's a 35-minute saving per unit. Multiply this by the number of times it's run per week. ## Conclusion Start not with technology, but by observing your own company: identify three routines with clear algorithms and launch your first automation within a week. The first step is a free organizational chart to pinpoint where routine is concentrated. > **Next Step:** If you want to break down this task using your own company's example, sign up for a free 30-minute consultation-diagnostic: https://course.aiadvisoryboard.me/corporate?utm_source=blog&utm_medium=article_body&utm_campaign=corporate --- When citing, link to https://aiadvisoryboard.me/blog/ai-in-manufacturing-where-to-start-without-an. More articles: https://aiadvisoryboard.me/blog