We Automated 40% of Our Operations — Not to Move Faster, but to Think Better
Automation promises order—silence where there was chaos, structure where there was scramble. But in a services startup, automation without intent can quietly erode human judgment and the culture that holds teams together.
At OSDB, we automated close to 40% of our operational workflows. The goal wasn’t speed or scale. It was sanity. What mattered wasn’t how much we automated, but how intentionally we drew the line between systems and people.
When Nothing Breaks, but Everything Feels Loud
We didn’t automate because things were failing. Tasks were tracked, updates existed, and work moved forward. Yet ownership was fuzzy, information was scattered, leaders became default escalation points, and senior team members spent more time coordinating than thinking. Everything worked—but it was noisy. That noise was the signal.
The 40% We Chose to Automate
We focused only on repeatable, low-judgment work. Odoo was already our task backbone, but adding intelligence changed its role. Instead of acting as a tracker, it became a visibility layer—surfacing daily and weekly summaries, flagging stalled work, and nudging follow-ups when thresholds were crossed. Ownership didn’t disappear; it became clearer.
Knowledge was the next friction point. By pairing XWiki with AI, repeated explanations turned into living documentation. SOPs evolved as workflows evolved. Tasks naturally linked to relevant knowledge, and documentation stopped being something we postponed until something broke.
We also built a lightweight internal AI layer that quietly connected the dots. It pulled signals from Odoo and XWiki, generated meaningful summaries, and flagged risks early. Automation stopped being about execution and started being about awareness.
What Stayed Human, On Purpose
Client conversations stayed human. Trust, tone, and nuance don’t translate well to systems. AI helped with context, but humans led the conversation.
Decision-making—especially where judgment mattered—remained human-led. Trade-offs, prioritization, and long-term impact required context that systems can’t fully grasp.
Critical task execution also stayed human. When work involved high impact, irreversible outcomes, or real-world consequences, humans stayed in the loop end to end. Automation could assist, but never decide or execute alone.
Community and creative work—from WomenForData to City based Meetups —retained its human soul. Energy, empathy, and intent can’t be automated
The Real Outcome: Less Noise, Better Thinking
Automation didn’t reduce people.
It reduced noise.
That gave humans the space to do what they do best—apply judgment, show empathy, and solve meaningful problems.
Closing Thought: Intent Over Scale
The goal isn’t maximum automation.
It’s intentional automation, with a clear boundary around what makes your business human.