“Just Put Some AI On It” => Why That’s Killing Projects Before They Start
- Anne Werkmeister
- Jun 30
- 2 min read
“Can we automate this with AI?”
“Let’s just plug in GPT and make it smart.”
“AI is the future, we need to show we’re using it.”
If you work in project or process management, you've heard these lines before.
Probably too many times. The buzzword “AI” is flying around boardrooms and pitch decks like fairy dust, a magical cure for anything vaguely inefficient.
But let me give you the short, honest answer:
No. AI is not the solution. Not yet. Not alone.
Here’s What Actually Happens
Recently, an entrepreneur approached us asking for a “full AI workflow.”
No clear process, no real data model, just the expectation that GPT would take care of everything.
Their vision? Automated emails, predictive logic, chatbot support, reporting, all hands-free, all AI.
Our first question:
“What’s the current process?”
Their answer:
“Well... that’s what we’re hoping AI will define.”
🚩 Red flag 1: You can’t automate what you don’t understand.
🚩 Red flag 2: You can’t enhance what doesn’t exist.
Before AI, You Need Data Clarity
You want AI? Great. But here’s the precondition no one wants to hear:
Start with clean, consistent, meaningful data.
AI learns from your data. Automations depend on clear inputs. If your customer database is a mess, your tagging inconsistent, or your workflows undocumented, AI will only scale your chaos.
In short:
Bad data in → worse decisions out.
Vague process → unreliable automation.
No feedback loop → no improvement.
This is why Intelligent Automation (IA), based on clear, structured workflows, should often come first. It forces you to understand your process before throwing complexity at it.
Start Here: Build Intelligently, Not Just Automatically
Understand your needs and current state: Don’t measure everything “just in case.” Data without purpose is noise. Map your workflows, talk to your users, and identify the real pain points. Be strategic: what are you trying to improve and why?
Clean and structure your data: No automation will work if the data feeding it is broken. Remove duplicates, define naming conventions, agree on formats, and ensure consistency across tools.
Put a process in place to protect your data: That could be automation, training, or just clear rules.The goal is to make sure your data stays clean over time, not just for one project.
Observe the results and take action: Once your data and workflows are under control, start measuring outcomes. Look for patterns, identify what’s working, and refine your process.This is how you improve your business, with clarity, not guesswork.
Only then, and maybe, bring in AI If your logic is too complex, if decisions require dynamic learning, or if human processing becomes a bottleneck, AI might help.
But don’t underestimate the cost: AI is resource-intensive. Even with strong ROI potential, it demands serious investment, in data, infrastructure, and design.
Sorry if it’s less sexy…
But this is the reality.
AI isn’t a silver bullet. It’s a scaling tool, not a foundation.It makes great things better, and bad things worse, faster. If you want your business to truly benefit from digital transformation, start with clarity, structure, and strategy.
At Romulus Technology, we help companies do the unsexy but essential work first, so when you do add AI, it actually works.
Comments