Head of CS breaks down the workflows that result in saved time and happier customers
Ajay Balamurugadas shares how AI workflows save time across customer research, support, and documentation while keeping relationships, nuance, and judgment human.
Ajay Balamurugadas
Head of Customer Success and Strategy at PostQode
Research: AI-driven customer research cuts preparation time from nearly a day to one or two hours.
Troubleshooting: AI-assisted troubleshooting speeds issue resolution by analyzing logs, errors, tests, and historical problems.
Human Trust: AI supports customer relationships, but empathy, judgment, and emotional connection still require direct human involvement.
Safeguards: Confident AI outputs require human validation, while fallback processes protect teams during tool outages.
Strategy: CX leaders achieve better results by matching models to tasks, breaking workflows down, and continuously refining usage.
Ajay Balamurugadas is Head of Customer Success and Strategy at PostQode, an AI-native software-development product.
We caught up with Ajay to understand how he's using AI to create happier customers. Here's what he told us.
Autonomous AI Agents in Software Development
My name is Ajay Balamurugadas, and I'm an experienced quality engineering leader. For the last two years, I have served as Head of Customer Success and Strategy at PostQode, a product company that brings autonomous AI agents to every stage of the software-development lifecycle.
Continue Reading for Free
Create a free account to finish this article, plus get ongoing access to timely insights and practical resources.
Have an account? Log In
Our customers include startups, enterprises, and service companies serving their own customers.
How AI Creates Happier Customers
Over the past year, we've made two concrete changes to how we deliver customer experience using AI.
First, we use AI extensively for customer research. Instead of relying only on manual analysis of meetings, feedback, and documents, we use AI to summarize conversations, identify recurring pain points, and extract expectations. AI also researches the customer’s company, their business activities, competitors, and possible solutions, highlighting gaps between what customers say and what they actually need.
AI-assisted analysis and summarization have reduced the time for customer and market research activities from nearly a full working day to typically 1-2 hours.
Second, we introduced AI-driven troubleshooting support for issue analysis and resolution. AI helps analyze logs, error patterns, test results, and historical issues to quickly narrow down possible root causes and suggest next steps. This has improved response times, reduced investigation effort, and enabled teams to resolve customer issues faster and with better context.
AI also generates and maintains documentation, such as onboarding guides, troubleshooting FAQs, and comparison documents — again, reducing the time and effort to create these from days to hours.
And most importantly, all of this has led to happier customers.
How AI can augment CX research workflows
Here's the CX research workflow we use before the first customer conversation even happens.
We begin with research using Perplexity AI to understand the prospect’s company, industry domain, competitors, market positioning, and recent activities. This helps us build strong contextual awareness before engaging with the customer.
Next, we use Claude or ChatGPT to combine that research with detailed information about our organization and identify how our platform can best help the specific prospect. We ask AI to suggest positioning strategies, value propositions, relevant use cases, and tailored ways to present the solution.
We then use ChatGPT or Gemini to simulate probable customer objections, concerns, and questions. This helps the team prepare responses in advance and improve the quality of customer conversations.
During customer meetings, we record and transcribe calls. We use ChatGPT to summarize discussions, generate structured meeting minutes, identify action items, and capture customer expectations and risks.
Post-meeting, we use Claude and ChatGPT to generate supporting artifacts such as solution documents, onboarding guides, comparison documents, diagrams, workflows, and troubleshooting materials.
Why Relationships and Nuance Cannot Be Outsourced to AI
But we don't use AI for everything.
Relationship building and emotional connection remain explicitly human. We believe trust, empathy, and long-term relationships cannot be outsourced to AI. People directly handle personal interactions like checking in on customer goals, understanding their progress, wishing them well on special occasions, and maintaining genuine human connection.
AI supports the relationship, but humans build and sustain it.
Additionally, in many cases, critical customer context resides across emails, CRM systems, meeting notes, call recordings, support tickets, and internal discussions, causing AI to underdeliver. In these areas where contextual continuity is missing, human intervention is still required to interpret nuance, validate recommendations, and connect information across sources.
Why Fallback Processes are Necessary with AI
Integrating AI into customer experience activities resulted in significant time and effort savings, improved consistency. But challenges have also emerged. This highlights the importance of having fallback processes and ensuring teams still retain core domain understanding and decision-making capability.
So, integrating AI into customer experience activities resulted in significant time and effort savings, improved consistency, and converting previously individual-dependent activities into more scalable and repeatable processes.
But challenges have also emerged. One concern is the “too good to be true” effect, where AI-generated outputs sometimes appear highly confident, even when deeper validation is required. This means human review and judgment remain important for critical decisions and customer-facing recommendations.
Another challenge is operational dependency. Once teams become accustomed to AI-assisted workflows, they find it difficult to maintain the same speed and consistency during AI downtime or tool unavailability. This highlights the importance of having fallback processes and ensuring teams still retain core domain understanding and decision-making capability.
Why CX Leaders Must Know Desired Outcomes Before Choosing Tools
It's important to clearly understand the exact desired outcome before choosing tools or models.
We initially assumed a single model could effectively handle every type of task, but we quickly learned that different models perform better for different use cases. Some are better at research, some at reasoning, some at structured documentation, and others at summarization or conversational tasks. Choosing the right model for the right activity can significantly improve quality and reduce cost and effort.
Another important lesson: Do not approach larger problems as one massive AI task. Breaking complex workflows into smaller, well-defined tasks and orchestrating them step-by-step produces far better and more reliable outcomes.
How to guide teams in effective AI use
We ensure effective adoption by fostering collaborative AI learning and usage across the team. We regularly train team members on different AI tools, use cases, prompting approaches, and workflows relevant to customer experience activities.
Additionally, we continuously share discoveries, examples, articles, prompts, and successful workflows internally so that knowledge spreads quickly across the organization instead of remaining with a few individuals. We also review outputs together, guide team members to improve context and prompting quality, and demonstrate artifacts created at different stages of customer engagement.
This continuous cycle of training, sharing, reviewing, and demonstrating helps the team build confidence and use AI more effectively and consistently in real customer scenarios.
How AI is Shifting Where Expertise Must Be Applied in CX
Traditionally, tasks like deep customer research, competitive analysis, documentation creation, meeting summarization, or solution preparation depended on specific experts with years of domain knowledge and experience. AI can now accelerate and support many of these activities more broadly, enabling more team members to contribute effectively with the right guidance and context.
This does not eliminate the need for expertise, but it changes how we apply it. Instead of spending most of their time on information gathering and repetitive preparation, specialists can focus more on strategy, relationships, decision-making, and customer outcomes.
How CX Leaders Should Approach AI
I advise CX leaders to treat AI as an assistant, not a replacement…Most importantly, experiment continuously. The organizations seeing the most value from AI are the ones actively testing ideas, refining workflows, learning from failures, and adapting quickly. AI adoption is not a one-time implementation — it is an evolving capability that improves with usage, feedback, and iteration.
I advise CX leaders to treat AI as an assistant, not a replacement. The quality of outcomes depends heavily on the quality of context you provide. Keep feeding AI with relevant business context, customer history, workflows, and even “context about the context” so it can generate more accurate and meaningful outputs.
Also, understand that different AI tools suit different use cases. Spend time learning the strengths and limitations of various tools across research, support, analytics, documentation, automation, and personalization.
Most importantly, experiment continuously. The organizations seeing the most value from AI are the ones actively testing ideas, refining workflows, learning from failures, and adapting quickly. AI adoption is not a one-time implementation — it is an evolving capability that improves with usage, feedback, and iteration.
Follow along
You can follow Ajay Balamurugadas' work on LinkedIn.
More expert interviews to come on The CX Lead!
