diff --git a/3-research-approach/approach.tex b/3-research-approach/approach.tex
index 4d3c7fe..bc148ec 100644
--- a/3-research-approach/approach.tex
+++ b/3-research-approach/approach.tex
@@ -68,7 +68,10 @@ discrete level first, we transform the intractable problem of global hybrid
 verification into a collection of local verification problems with clear
 interfaces. Verification is performed per mode rather than on the full
 hybrid system, keeping the analysis tractable even for complex reactor
-operations.
+operations. \addedprose{Figure~\ref{fig:hybrid_automaton} illustrates this
+structure for a simplified reactor startup sequence, showing discrete modes
+connected by guard-triggered transitions with continuous dynamics governing
+behavior within each mode.}
 
 \begin{figure}
   \centering
@@ -203,7 +206,8 @@ objective.
 \end{figure}
 
 
-This operational control level is the main reason for the requirement of human
+\addedprose{As shown in Figure~\ref{fig:strat_op_tact}, nuclear plant control
+spans three levels: strategic, operational, and tactical.} This operational control level is the main reason for the requirement of human
 operators in nuclear control today. The hybrid nature of this control system
 makes it difficult to prove what the behavior of the combined hybrid system will
 do across the entire state-space, so human operators have been used as a