I want to get quick AI audio response from the user voice

I want to get quick AI audio response from the user voice
I heard that Deepgram API is the best solution for this, so I integrated Deepgram API into our system.
It has long delay time to generate AI audio responses.

const audioData = event.target.result;
await axios.post("https://api.deepgram.com/v1/listen", audioData, { headers: headers })
  .then((response) => {
    if (response.data) {
      transcript = response.data.results.channels[0].alternatives[0].transcript;
    }
  }).catch((error) => {
    console.error("Error while transcripting:", error); // Handle errors
  });

  // Get response from OpenAI GPT-4o
  const completion = await openai.chat.completions.create({
    messages: messages,
    model: "gpt-4o",
    temperature: 0.2,
  });
  ai_response = completion.choices[0].message.content;

  // Generate audio from text and play
  const config = {
    headers: {
      Authorization: `Token ${process.env.DEEPGRAM_API_KEY}`,
      "Content-Type": "application/json",
    },
  };
  const data = {
    text: ai_response,
  };
          
  const response = await fetch("https://api.deepgram.com/v1/speak?model=aura-zeus-en",
  { 
    method: "POST", 
    headers: { ...config.headers }, 
    body: JSON.stringify(data)
  }
);
if (!response.ok) {
  throw new Error(`HTTP error! Status: ${response.status}`);
}

const audioContext = new (window.AudioContext || window.webkitAudioContext)();
const source = audioContext.createBufferSource();
// Fetch the audio data as an ArrayBuffer
const arrayBuffer = await response.arrayBuffer();
audioContext.decodeAudioData(
  arrayBuffer,
  (buffer) => {
    source.buffer = buffer;
    source.connect(audioContext.destination);
    source.start(0);
  },
  (e) => {
    console.log("Error with decoding audio data" + e.err);
  }
);

This is my current code using Deepgram API, but I am not sure what is the issue.
I integrated Deepgram API and OpenAI API in React.js-frontend for quick speed, but yet it has long delay.

What are the best solutions for quick AI audio response?