Clouded Judgement: Managing AI as Living Systems vs. Static Software
The article argues that AI products require fundamentally different operational and architectural approaches than traditional deterministic software, treating them as living systems with continuous evolution rather than stable releases. The author emphasizes that successful AI product companies must build infrastructure for observability, continuous evaluation, controlled rollouts, and customer communication to manage model drift and probabilistic behavior. A secondary section provides SaaS valuation multiples analysis and M&A landscape observations across the software industry.
Metrics in this report
32months
median
Public SaaS companies (gross profit-adjusted)
5.1multiple
median
Overall public SaaS companies
24.2multiple
median
Top 10 public SaaS companies
28.4multiple
median
High-growth SaaS (>25% NTM growth)
8.8multiple
median
Mid-growth SaaS (15-25% NTM growth)
4.1multiple
median
Low-growth SaaS (<15% NTM growth)
18percent
median
Public SaaS companies
15percent
median
Public SaaS companies
76percent
median
Public SaaS companies
14percent
median
Public SaaS companies
12percent
median
Public SaaS companies
108percent
median
Public SaaS companies
-2percent
median
Public SaaS companies
24percent
median
Public SaaS companies
37percent
median
Public SaaS companies