User data analysis

Project: AI tools in design

Channel: design process

Timeline: 2025 - 2026

Role: Product designer

Tool: MS Copilot

Team: Savings & Investment SLAM, personal project

Problem: In today's fast-paced environment, there is an increasing demand to deliver valuable and user-centric solutions based on the qualitative research faster & in more effective manner. This situation made me to reevaluate my research routine and assess its efficiency to maximize the positive impact on my solutions.

Solution: Application of own analytical prompts to get the detailed customer data analysis and synthetizing in-dept insights to inform related design decisions.

Situation

I personally felt that as new AI technologies emerged and rapidly improved, the surrounding time and conditions shifted. All design fellows noticed these changes and had to adjust accordingly. Eventually, I decided to examine opportunities that might be benefical for strenghtening my own argumentation regarding customers' voice towards my business partners. I recognized the potential of AI prompting to support my design work through the obvious synergistic effect. So, I seized this momentum as an opportunity for my growth.

Personal objective

Reevaluate my UX research process to identify the potential impacts of an AI application on this process.
Determine supportive areas that could be streamlined or explored in greater depth.
Identify any weak points that could be eliminated or improved to enhance time management.
Keep my focus and time spent on core assessment and judgmental activities related to critical thinking and finding better solutions.

Context

Challenges

Constraints

The main limits for me in this challenge were the corporate licence setup, which later on worked out for me.

Parallel, I needed to figure out the optimal combinations of intended prompting for the narrowed purpose of feedback & user data in the regulated propositions

All that data collection happened in the form of a standardized reporting summary, blind to some inner connections.

Thus, I managed to have the Excel raw data database for better targeted analysis & data synthesis.

Due to an extensive data set & a limited time frame to go through each month for each product, I looked for AI assistance.

UX approarch

For this workflow level-up I took the case of our regular reporting about after-sales clients' feedback & satisfaction withe products and digital processes. I decided to incorporate AI advanced promting in the actual process.

I prepared & fine-tuned my own set of analytical prompts based on my experience in conducting business and customer analysis.
Knowing the weak spots and gaps that were challenging to address—and that required an excessive amount of time for deep dives—I focused my attention accordingly.
After executing the usual steps, I used my prepared prompts to conduct a thorough analysis with Copilot. This approach allowed me to maximize results while minimizing effort.

Key impact

Interestingly, some carefully crafted questions, supported by detailed follow-up inquiries, led to significant perspective shifts for my business stakeholders.

In a few rounds of this prompting, I was able to uncover some hidden patterns & clients' shifts for our key digital processes.
This gave me additional puzzle pieces and allowed me to craft my particular UX suggestions & recommendations with additional strong arguments.
With this deeper analysis outputs produced for my business partners, I managed to convince them to act now in particular areas, rather than postpone them for later priority changes.
It was also evident that some colleagues were waiting for decisions until spotted potential stagnation in the presented projections & future trends, which could directly impact our planned KPIs.

My analytical prompting kit

I decided to integrate this advancement into my existing process because I believe that AI inputs can significantly enhance my observations and help correct any biases that may limit my perspective. Ultimately, I am confident that this support may enrich the final insights derived from the reporting and lead to improved qualitative interpretations for future decision-making.

I ended up with a set of eleven analytical prompts to cover all aspects of this analysis in certain layers so that I could achieve a repetitive effect for each analyzed product and process and get strong arguments for my UX recommendations & design decisions in front of my business partners.

Through this analytical prompting exercise, I aimed to achieve the following goals:

1. Evaluate actual product processes from the clients' perspective, based on their own user experiences.

2. Connect various influences, including demographics, seasonal shifts, and motivators, to the types of service processes, while considering scores from NPS, CSAT, and CES, along with personal comments and responses.

3. Identify the next short-term steps to achieve quick wins for immediate improvements.

4. Anticipate potential long-term patterns and trends in clients' perceptions of digital processes.

5. Explore areas of interest for enhancing user experience and addressing process challenges on a long-term scale.

With the intention of a deep understanding of the reporting data as well as to distill the strongest insights, my key criteria for this prompting were:

Feedback differences in different seasons of the year - inspect any effects of digital campaigns, special limited offers with benefits & seasonal specifics.

The most frequently cited negative areas of the process - focused on comparing the perceived value & usability of the current processes with the planned feature releases.

Shifts in comments & scores comparing pure digital & digital-assisted services - aimed at enhancing understanding of the differences in customer needs and challenges addressed by different channels and service types.

Connection of the comments with the granted NPS / CSAT score - supposed to link the received feedback comments with granted scores in the watched metrics data.

Identification of the repeated patterns & areas over the past year - analyze all trends in the feedback data collected in the last 12 months. This analysis will help to form UX assumptions and hypotheses that inform future corrective actions.

User data analysis with Copilot

How to pursue a proper quarterly analysis without losing yourself in details

Exactly, this was the moment I decided to involve active AI prompting into the process of data mining to squeeze maximum insights from the reporting with a smart amount of effort spent on this exercise. I set the following steps in motion.

Every month, I received regular reports on all our digital savings & investment products from our insight partner.

I update the ever-lasting pivot tables with the newest inputs from clients' raw data for each of these products.

After quarter closure and a finalized update for that period, I initiated a prompting conversation in these Excel files and initiated the Copilot analysis with my predesigned set of prompts.

This served two purposes in parallel. Firsthand, the main data search & mining was running as required.

On the other side, I was able to jump into the conversation for minor adaptations where the digging could go even deeper or customize the prompts for that particular case.

As a result, I had detailed data analysis at hand in a few minutes (or a lunch break) together with an overview of possible actions to take and a backlog plan.

This material helped me prepare a presentation filled with arguments based on the clients' perspectives. It will be useful for my regular meetings with business partners to develop an effective action plan for maintaining optimal processes.

Part of my job involved regularly reviewing after-sales feedback from our clients to identify potential gaps, issues, and complaints regarding the digital processes for our savings and investment retail portfolio. Every quarter period, I was in charge of producing the data analysis together with improvement suggestions. Since I was responsible for four main product streams, this task was quite challenging and time-consuming. It required me to manage my workload effectively, as my capacity was spread across many other tasks on my to-do list.

Final reflextions

Reflecting on this AI-related initiative, I am truly grateful that I chose to listen to my analytical side and allowed new technology to enhance my research capabilities. The entire process proceeded smoothly and iteratively, maximizing the benefits for my work.

The effort, time, and creativity I invested paid off, as I found joy in this exercise each time. I shared my experiences and knowledge with my team to help establish a standardized best practice aimed at improving our daily design routines during feedback and research activities, ultimately strengthening our UX argumentation.

After leaving ČSOB, I am confident that I will carry this expertise with me into all my future projects. While this has been just one of my AI-related initiatives, I plan to continue expanding my AI toolkit further.