diff --git a/packages/api/src/routers/article_router.ts b/packages/api/src/routers/article_router.ts
index 691c3c293..7915abc4f 100644
--- a/packages/api/src/routers/article_router.ts
+++ b/packages/api/src/routers/article_router.ts
@@ -150,10 +150,7 @@ export function articleRouter() {
         user: { id: uid },
         elasticPageId: articleId,
         state: SpeechState.INITIALIZED,
-        voice:
-          voice ||
-          userPersonalization?.speechVoice ||
-          'en-US-JennyMultilingualNeural',
+        voice: voice || userPersonalization?.speechVoice || 'en-US-JennyNeural',
       })
       // enqueue a task to convert text to speech
       const taskName = await enqueueTextToSpeech(uid, speech.id)
diff --git a/packages/api/src/utils/textToSpeech.ts b/packages/api/src/utils/textToSpeech.ts
index 556f0b128..b75265743 100644
--- a/packages/api/src/utils/textToSpeech.ts
+++ b/packages/api/src/utils/textToSpeech.ts
@@ -20,6 +20,7 @@ export interface TextToSpeechInput {
   textType?: 'text' | 'ssml'
   rate?: number
   volume?: number
+  complimentaryVoice?: string
 }
 
 export interface TextToSpeechOutput {
@@ -177,7 +178,7 @@ export const synthesizeTextToSpeech = async (
         // use complimentary voice for blockquote, hardcoded for now
         const voice =
           htmlElement.tagName.toLowerCase() === 'blockquote'
-            ? 'en-US-AriaNeural'
+            ? input.complimentaryVoice || 'en-US-AriaNeural'
             : input.voice
         const ssml = htmlElementToSsml({
           htmlElement: e,
@@ -218,7 +219,7 @@ export const synthesizeTextToSpeech = async (
 export const htmlElementToSsml = ({
   htmlElement,
   language = 'en-US',
-  voice = 'en-US-JennyMultilingualNeural',
+  voice = 'en-US-JennyNeural',
   rate = 1,
   volume = 100,
 }: {
diff --git a/packages/api/test/utils/textToSpeech.test.ts b/packages/api/test/utils/textToSpeech.test.ts
index 2276b2ac4..e78492cca 100644
--- a/packages/api/test/utils/textToSpeech.test.ts
+++ b/packages/api/test/utils/textToSpeech.test.ts
@@ -35,7 +35,7 @@ describe('textToSpeech', () => {
       const htmlElement = parseHTML(
         `<p data-omnivore-anchor-idx="1">Marry had a little lamb</p>`
       ).document.documentElement
-      const ssml = htmlElementToSsml(htmlElement)
+      const ssml = htmlElementToSsml({ htmlElement })
       expect(ssml).to.equal(
         `<speak xml:lang="en-US" xmlns="http://www.w3.org/2001/10/synthesis" version="1.0"><voice name="en-US-JennyNeural"><prosody volume="100" rate="1"><bookmark mark="data-omnivore-anchor-idx-1"></bookmark><p data-omnivore-anchor-idx="1">Marry had a little lamb</p></prosody></voice></speak>`
       )