Everyone's talking about AI and automation. How it can save time, reduce manual work, and boost efficiency. But many SMEs start out by treating AI automation like a traditional IT project.
They buy tools, train staff... but the processes stay the same, bottlenecks remain, and the impact is minimal. So what's going wrong?
Quick note on what we mean by AI + Automation: Automation includes rule-based workflows inside your CRM, Google Apps Script, simple JavaScript, and no-code tools like Zapier, ActivePieces, or N8N. AI covers functions like OCR, summarization, or classification, connected into most no-code tools nowadays.
1. IT team is essential, but they can't do it alone
It's common for SMEs to hand off automation projects to the IT team. And while IT teams are great at building and deploying tools, they don't always have full visibility into how every business process works on the ground.
- Many business processes rely on judgment calls and experience that aren't easy to translate into code.
- They bring the technical skills but they're often not the ones using the workflows day-to-day.
- So even when tools are built, they may not match the actual needs of the business teams.
AI automation isn't just "adding a new tool."
2. Can business teams drive automation themselves? Not so simple
Business teams know their process the best, but asking them to lead automation projects on their own doesn't usually work either:
- They're already busy with operations.
- The tools can be technical: writing prompts, setting up APIs, managing logic.
- Coordinating with IT and other teams is time-consuming.
Some companies think: "We've got someone good with Excel and computers, let's have them figure out automation." It sounds reasonable, but often ends in frustration. That person is already overloaded, there's no clear goal, no support, no space to experiment. When it doesn't work, the blame lands on them.
AI Automation needs dedicated effort and clear direction. It's not just a side task for whoever's tech-savvy.
3. AI automation isn't just "doing the same work faster"
Many assume AI automation means taking the current process and doing it quicker & cheaper. But most processes were built around manual work. They may not be a good fit for AI automation as-is.
- AI might allow for entirely different (and better) steps.
- What used to be A → B → C → D could become A → B → straight to D.
- If you stick too closely to old ways, you miss the real benefits of automation.
The best results come when automation and process redesign go hand in hand.
4. An experiment: build something AI-first
Sometimes, the answer isn't fixing the old. It's starting something new, with AI and automation at the core from day one.
That's what being "AI-native" really means:
- Workflows are optimized for machines, not adapted from manual steps.
- People are only involved where they add real value: judgment, insight, or creativity.
- Testing and iteration come fast, using no-code tools to build real working versions in days, not months.
You will need someone who understands business workflows, has some hands-on ability with tools, communicates well across teams, and isn't afraid to test and learn. Their job is to spot repetitive steps, quickly test ideas using no-code tools like Zapier, Make, N8N, or Activepieces, and partner with business users and IT to create real, working prototypes.
5. Real-world examples
Example 1: Reviving a stalled project with AI-first thinking. A lead generation project was shelved two years ago. Too manual, too expensive, results were poor. Instead of picking up where they left off, the team rebuilt it from scratch. AI scores and prioritizes leads, outreach messages are automatically generated, sales only steps in when a lead shows real engagement.
Example 2: Launching a new service with limited resources. Want to offer trend analysis but don't have the budget for a big team? Use AI to gather and summarize data, auto-generate a draft report, let a human reviewer add insights before delivery. That's a service you can launch lean and scale later.
Final thoughts
Most failed automation efforts aren't caused by bad tools or lack of talent. They stall because:
- Business teams don't have time or support to rethink workflows.
- IT teams can't redesign what they don't fully live day to day.
- And no one is owning the space in between, where real change happens.
The most effective teams don't treat automation as an upgrade. They treat it as an opportunity to rebuild, from the ground up, with AI and automation as core building blocks.
Originally published on LinkedIn. Want to discuss this further? Get in touch.
