diff --git a/en/reference/bm25.html b/en/reference/bm25.html
index a3246865c0..f47640e762 100644
--- a/en/reference/bm25.html
+++ b/en/reference/bm25.html
@@ -362,22 +362,17 @@ <h2 id="example">Example</h2>
 </p>
 
 <p>
-If the <em>enable-bm25</em> flag is turned on after documents are already fed,
-some extra steps must be executed to prepare the posting lists in the memory and disk indexes for this field.
-For each content node do the following:
-<ol>
-  <li>Perform <a href="../proton.html#memory-index-flush">memory index flush</a> using
-      <a href="/en/operations-selfhosted/vespa-cmdline-tools.html#vespa-proton-cmd">vespa-proton-cmd</a>:
-<pre>
-vespa-proton-cmd --local triggerFlush
-</pre>
-  </li>
-  <li>Perform <a href="../proton.html#disk-index-fusion">disk index fusion</a>,
-      preparing the posting lists for use with <em>bm25</em>:
+If the <em>enable-bm25</em> flag is turned on after documents are
+already fed then <a href="../proton.html">proton</a> performs
+a <a href="../proton.html#memory-index-flush">memory index flush</a>
+followed by a <a href="../proton.html#disk-index-fusion">disk index
+fusion</a> to prepare the posting lists for use with <em>bm25</em>.
+
+Use the <a href="../proton.html#custom-component-state-api">custom
+component state API</a> on each content node and
+examine <code>pending_urgent_flush</code> to determine if the
+preparation is still ongoing:
 <pre>
-vespa-proton-cmd --local triggerFlush
+/state/v1/custom/component/documentdb/mydoctype/subdb/ready/index
 </pre>
-  </li>
-</ol>
 </p>
-