Onboarding and Psychological Safety
Joining the team felt like being thrown into the deep end without a life vest. Because tenured team members were often stretched thin or taking leave in succession, new hires were expected to jump straight into high-stakes tasks with almost no shadowing or training time. The environment lacked psychological safety. Instead of a supportive ramp-up period, there was an immediate, heavy pressure on performance that felt based more on managerial perception than actual data-backed output.
The Documentation Struggle
Most institutional knowledge was tribal, meaning it lived in people's heads rather than in a shared database. Documentation was either missing entirely or out of date. If the person who last worked on a project was still around, you could prompt them for answers, but otherwise, you were left to infer context from whatever fragmented data you could find. This made it incredibly difficult to work efficiently or scale projects.
Visibility Over Results
The performance review process was heavily dependent on constant self-promotion. If you didn’t frequently log your wins into the internal tracking system, or if stakeholders didn't proactively celebrate your effort, it was as if the work never happened. This created a culture where being loud about your achievements was often more important than the actual quality of the work. Furthermore, the ability for management to manually adjust AI-generated summaries made the process feel highly subjective.
Strategic and Technical Blind Spots
There was a noticeable gap in fundamental market knowledge during planning. Results were often compared across different global regions without accounting for basic economic realities like currency strength or local cost-of-living. The company also tended to overstate its market position. While bolstering confidence is important, there was a lack of realism regarding the shifting competitive landscape and the time that had passed since the last major funding milestone.
The AI Integration
The company was eager to adopt AI but often did so without enough human oversight. Simple AI-generated summaries were sometimes trusted over deep technical research, leading to avoidable errors in judgment. There was also a lack of cultural sensitivity in AI-generated marketing materials (see SG60 video, 1993), which risked negative reception in diverse markets. Finally, the approach to tools was inconsistent, with an odd expectation for employees to navigate their own subscriptions rather than the company providing professional-grade access for work tasks.