Discovering Actionable Insights From Customer Feedback With Azure OpenAI

Today, businesses often collect vast amounts of customer feedback through various channels such as websites, social media, and direct surveys. And for any business, understanding customer sentiment and preferences is crucial for developing products that resonate with the market.  However, this feedback is typically unstructured, making it challenging to analyse systematically. As a result, companies struggle to sift through the noise to identify key themes and sentiments that could inform product development, quality assurance, and set up vendor management strategies.

Collecting the data

This client had an abundance of data at its fingertips but needed to determine how best to capture the right information to create actionable insights. Delays of two to three months in processing customer feedback were hampering the company’s ability to nimbly respond to individual feedback and market trends. The heart of the challenge was in collecting feedback across various touchpoints and interpreting it in a way that could drive meaningful change within the organisation.

The Protiviti team, which included experts in data management and analytics, began by partnering with product and quality control leaders and their teams to identify how the current data collection and interpretation process was working.  We reviewed the client’s situation and discovered that its existing processes for gathering and analysing customer commentary were time-consuming, leading to inconsistencies and missed insights that slowed progress and decision-making.

Armed with that information, we developed an AI solution using Azure OpenAI services capable of summarising 133,000 customer comments into approximately 25 classified themes, ranging from pet wellness, product love, and food safety to packaging, price, and sustainability. This intelligent summarisation process also included sentiment analysis to gauge customer emotions associated with each theme. These features were instrumental in getting the customer data needed:

• Automated data collection: The automated processes we established collect customer comments from the website via web scraping and upload to the centralised database. 
  • The development of a dynamic dashboard provides real-time customer feedback insights, enabling quick responses to emerging trends and issues.
• Theme classification: Using advanced natural language processing (NLP) algorithms within Azure OpenAI, feedback is classified into relevant themes such as product design, usability, pricing, logistics, safety, and more. 
• Sentiment analysis: AI models assess the sentiment behind each comment—positive, negative or neutral—to provide deeper insights into customer opinions.
• Integration with Power BI: The summarised data, along with sentiment scores, are integrated into Power BI dashboards where they can be visualised in real-time and explored by the product team and quality assurance team.

Doggone good results

By implementing this AI-powered solution using Azure OpenAI services coupled with Power BI integration, this client saw immediate and sustainable results, including:

• Enhanced product development: The product team gained valuable insights into what customers like or dislike about current offerings, enabling them to prioritise features accordingly. 
• Improved quality assurance: QA teams now monitor recurring issues highlighted by customers, ensuring timely resolution before escalation. 
• Strategic vendor management: Identifying patterns related to vendor-supplied components helps the client make better-informed decisions about continuing partnerships or seeking alternatives. 
• Reduced manual review: Customer comment review time has decreased tenfold. Instead of taking a month to review comments, it can now be done in two to three days, allowing the product team to focus on more strategic tasks. This, in turn, leads to enhanced productivity and faster decision-making, ultimately accelerating the product development cycles and improving overall operational efficiency. 
• Higher accuracy rate and consistency in theme classification: Achieving a higher accuracy   rate and maintaining consistency in theme classification ensures that insights derived from customer feedback are reliable and actionable. This precision enables the product and quality assurance teams to make informed decisions based on accurate and consistent data, leading to more targeted improvements, increased customer satisfaction, and a stronger competitive edge in the market. 

Innovation for transformation

At Protiviti, we are  dedicated to using innovation to drive meaningful business change. This project, leveraging Azure OpenAI to convert customer feedback into actionable insights, reflects our commitment to excellence and customisation. We know that each organisation has unique needs and objectives and tailor our AI-driven strategies to align perfectly with those specific goals so that data is used to its full potential. Imagine a scenario where manual review times are slashed, freeing up resources that can be redirected towards strategic initiatives, with reliable insights that drive informed decision-making. We create this reality for our clients every day and we continuously refine our solutions based on real-world feedback and evolving needs. By integrating advanced technologies like Azure OpenAI with intuitive platforms such as Power BI, we empower teams to make data-driven decisions effortlessly.