Recently we ran a quick survey with over 350 Customer Success professionals to explore what we believe is a common challenge: struggling to come up with the right questions or phrasing them effectively when creating forms or surveys. Interestingly, during follow-up discussions, some Customer Success Managers (CSMs) also shared that they face similar difficulties when it comes to formulating the right questions during live calls with customers.
It is true that engaging conversations, getting insights, and providing customer support all start with the right questions and good follow-ups. But when you’re still building experience or are new to the operations of the company, it’s easy to feel stuck or unsure of what to say next.
To address this matter, we decided to experiment with AI and see how helpful it can be for Customer Success professionals in their day-to-day job. With Large Language Models (LLMs) like ChatGPT or Claude, you can get practical support in crafting the right questions — whether you’re prepping for an interview, managing a live call, or building confidence in spontaneous questioning.
As AI becomes more powerful, it opens up new creative ways to enhance customer interactions and problem-solving for CSMs. We’ve broken this process down into three key stages — preparing for the call, the call itself, and post-call debriefing. In this article, we’ll walk you through each stage with practical examples of how to leverage ChatGPT to improve your customer conversations.
Stage 1. Preparation for Customer Calls as a Customer Success Manager
Preparation is key to a successful customer interaction. Thanks to AI, you can now simulate a live conversation with users or ask ChatGPT to come up with multiple smart questions based on your context. Let’s dive deeper into these approaches and explore the best ways to implement them.
Use ChatGPT to Generate Smart, Contextual Questions
Sometimes, even if you know the key topics you need to cover on a call, figuring out how to phrase your questions in a way that gathers sincere responses and real insights can be tricky. That’s where ChatGPT can step in to help you craft well-thought-out questions tailored to each client’s situation.
To get the most relevant suggestions, provide ChatGPT with detailed information about the customer and their journey with your product. For instance, you could prompt:
“I’m a CSM at a SaaS company, and I have a call with a person who’s been using our product for six months. They’ve interacted with our analytics dashboard but haven’t engaged much with the collaboration features. What are some insightful questions I can ask to understand their challenges and potential churn risks?”
Based on the context you provide, ChatGPT will generate a list of thoughtful, targeted questions like:
- What’s been the most valuable feature for your team so far?
- Have you encountered any roadblocks that slowed down adoption?
- How does our platform compare to other tools you’ve used?
- What’s one improvement that would make your experience significantly better?
- We noticed you’ve been using the analytics dashboard regularly but haven’t explored the collaboration features—could you share if there’s anything preventing your team from using them?
This approach ensures you walk into the call with a strong set of questions, helping you feel more confident. Plus, it allows you to steer the conversation towards meaningful insights, discover potential problems, and better understand what technical support the customer needs. Practice Mock Interviews with ChatGPTOne of the most effective ways to improve your skills as a CSM is through mock calls. If you struggle with thinking on your feet during live conversation with clients, use ChatGPT as a practice tool — to simulate user interactions, train your answers and improve your confidence. This way, you can:
- Practice active listening and follow-up questions
- Learn how to navigate difficult conversations
- Improve your ability to respond naturally and with empathy
To get started, provide AI with some context and ask it to role-play the customer challenging you with technical questions or ambiguous feedback. You can either let ChatGPT create diverse customer profiles or simulate specific scenarios tailored to your needs. For example, if you want to get better at handling complaints, you might prompt the following:
Act as a customer who has been using our SaaS product for three months but is struggling with the reporting feature. You're considering downgrading your subscription. Let's have a conversation where I try to understand your pain points and offer solutions.
After you're done with the mock call, you can take your practice a step further by asking ChatGPT to analyse your responses. It might point out if your answers were too generic, suggest more targeted follow-up questions, or recommend ways to better handle objections. It can even highlight moments where you could have reinforced the value of your service more effectively or adapted your tone to match the customer’s mood. Practice Mock Calls with Interactive AvatarsIf you're looking for a more realistic experience when practicing conversations with users, you can turn to HeyGen's Interactive Avatar service. While we’re not the biggest fans of AI-generated videos for onboarding, workshops, or sales, there’s definitely value in training with a human-looking avatar instead of just text-based interactions with ChatGPT. It can add a layer of realism to your practice sessions, helping you better prepare for real-world conversations.
HeyGen allows users to create lifelike AI-generated avatars that can engage in real-time, two-way conversations. These avatars can be customised to represent different characters or personas and are designed to respond dynamically to user input. The service combines advanced AI with natural language processing to create realistic, human-like interactions. In addition, you can connect it to your ChatGPT or other LLM account to provide the platform with better context and enhance your conversation experience.
Stage 2. Assistance During the Customer Calls
Even during live interactions, AI can be a supportive tool to keep the conversation flowing smoothly. While you can't rely on it to listen to the call in real-time, you can input key details to generate follow-up questions or answers. This can be particularly useful if you hit a conversational roadblock, need help with interpreting of what you customer says, or want to dive deeper into a topic without sounding repetitive.
Get Real-Time Question Suggestions During Calls
If you’re on a live call and feeling stuck, you can discreetly use LLMs to generate follow-up questions in real-time. Unfortunately, at the moment of writing this article, ChatGPT is not capable of listening to live conversations and providing real-time suggestions based on what’s being said. Although this would be ideal, there is a workaround which will let you receive questions and follow-up suggestions based on what your customer says in real-time.
Before the call, give ChatGPT a detailed prompt explaining that you need help asking thoughtful, relevant questions during your client conversation. The more context you provide, the better ChatGPT can tailor its responses to fit the situation. For example, specify what your product or service does, how long the client has been using it, whether they’re a paying customer, and any other relevant details related to their use case or feedback so far. This way, ChatGPT can suggest the most effective questions to guide your conversation.
Step 1. Provide ChatGPT with some context about your call and explain how you want it to act.
In the first step, you not only share the context but also clearly explain how you want ChatGPT to assist you. Do you want responses that focus on technical details? Or ones that encourage the client to open up about their workflow? Setting this expectation upfront ensures the AI provides responses that match your conversation goals.
Step 2. Engage with ChatGPT during the call to ask better follow-up questions
When you’re on the call, use ChatGPT’s voice mode (activated by the black button in the prompt field) to record your customer talking. Once they’ve finished speaking, send their message to ChatGPT, and it will generate a text response you can use to keep the conversation going. Be sure to mention in your setup prompt that you’d like text output instead of audio, so the AI responds in the format you need during the call.
Pros and Cons of the Real-Time Assistance Approach
Of course, this approach isn’t flawless. You can’t just turn on the voice mode and expect instant, real-time recommendations based on the conversation with your client. It does require a bit of manual effort, and there will be some delay in receiving responses — ChatGPT needs a few seconds to generate output after you send the prompt, and you’ll need a few more moments to read and process its suggestions. However, it’s still worth trying AI assistance, especially if you're new to Customer success and looking to improve the quality of your conversations with clients.
While there may be a slight delay, the benefits of having a second brain focused on the technical details and deeper needs of your client can make a significant difference. As you get more familiar with ChatGPT’s suggestions, integrating its insights into the conversation will become more seamless, and the process will feel much more natural.
Stage 3. Post-Call Debrief
After the call, reflection and follow-up are vital for turning insights into actionable strategies. There are several ways you can use AI for the post-call stage. For example, it can review your performance during the conversation. And it can also help you create a list of vital questions and answers for your future interactions.
Use ChatGPT for Post-Call Analysis
After talking to a customer, it’s essential to reflect on the conversation and find areas for future improvement. You can transcribe the recording of your call into text and share it with ChatGPT, or you can provide with your own summary of the conversation. For example, you can use this prompt:
"Here’s a summary of my customer call. What additional questions should I have asked?"
ChatGPT can then evaluate your summary and suggest areas where your questioning might have been too narrow or missed critical details. It might identify potential follow-up questions that could collected better data. By receiving this feedback, you can adjust your approach for the next interactions. This process also helps you pin point patterns in the types of information you're gathering. For instance, it can confirm that you're consistently asking questions that drive real value in each user interaction. Over time, using ChatGPT for post-call analysis will help refine your questioning skills.
Build a Personalised Question Bank
As you continue to interact with customers and work with AI tools like ChatGPT, you’ll start noticing patterns in the types of questions that are most effective for different situations. Over time, you can compile the best questions into a personal question bank.
Pro tip: To get the best results, build your bank by categorising questions based on common scenarios, such as onboarding, renewal, churn risk, feature adoption, and more. For example, you might include questions specifically designed to help you discover a customer’s satisfaction level during onboarding, or questions that can help identify early signs of churn risk.
As you conduct more mock interactions with AI and real client calls, you’ll create a refined set of questions, including the most effective phrasings that get to the heart of what you're trying to learn. This resource will save you time and effort in future and help you build consistency in the customer interactions.
Final Thoughts
For CSMs, especially junior, developing strong questioning skills is crucial for building relationships, driving customer success, and reducing churn. ChatGPT can serve as a reliable assistant, helping you prepare, practice, and improve your ability to ask meaningful questions and provide thoughtful answers. By integrating AI into your workflow, you can accelerate growth in your job and gain confidence in customer conversations. Try experimenting with ChatGPT today and see how it transforms your approach to customer interviews!
