diff --git a/docs/notebooks/pulse_efficient_transpilation_qsim.ipynb b/docs/notebooks/pulse_efficient_transpilation_qsim.ipynb
index a182055..b26f378 100644
--- a/docs/notebooks/pulse_efficient_transpilation_qsim.ipynb
+++ b/docs/notebooks/pulse_efficient_transpilation_qsim.ipynb
@@ -14,13 +14,6 @@
     "In this notebook, we show readers how to transpile circuits in a pulse-efficient way by transpiling the two-qubit gates into the [`RZXGate`](https://qiskit.org/documentation/stubs/qiskit.circuit.library.RZXGate.html). The 'RZXGate' is a continuous gate set native to the hardware. We show examples where this method results in shorter circuit depth and duration. Since qubits have a limited coherence time, a shorter total circuit duration is desirable. We demonstrate how to implement this technique to transpile a relevant circuit in simulating quantum dynamics. And the Runtime result from the hardware suggests that the `RZXGate` type of transpilation produces more accurate results."
    ]
   },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "# Start"
-   ]
-  },
   {
    "cell_type": "code",
    "execution_count": 56,
@@ -74,7 +67,6 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "# Build\n",
     "When running a generic quantum circuit on the hardware, the circuits need to be transpired into a basic set of gates so that it is compatible with a target backend. This process is handled by the [`qiskit.transpiler`](https://qiskit.org/documentation/apidoc/transpiler.html), and in the how-to guide, we will introduce two ways of defining the basis gate set."
    ]
   },
@@ -107,8 +99,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## Example: Ising model\n",
-    "\n",
+    "# Qiskit Patterns Step 1: build trotterized evolution     \n",
     "Let's start by looking at a simple example circuit on two qubits. The circuit example we use here simulates the dynamics generated by the Hamiltonian $H = - J \\sum_{i,j}Z_i\\otimes Z_j + h\\sum_{i}X_i$. To approximate the dynamics, we implement its trotterized evolution using [`PauliEvolutionGate`](https://qiskit.org/documentation/stubs/qiskit.circuit.library.PauliEvolutionGate.html) pametrized by $J$, $h$ and the trotterized time step $t$."
    ]
   },
@@ -156,7 +147,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "# Transpile"
+    "# Qiskit Patterns Step 2: Transpile"
    ]
   },
   {
@@ -579,7 +570,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "# Run"
+    "# Qiskit Patterns Step 3: execute"
    ]
   },
   {
@@ -630,6 +621,13 @@
     "    ecr_job_zz = estimator.run([(ecr_circ, ecr_obs)])"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Qiskit Patterns Step 4: compare results"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,