Role mapping
Mapped out all the roles within Solar using AI, similar companies structure and Teams. The real research would be much more in depth and involve interviewing senior managers to comprehensively understand the company structure.
Briefing
AI efficiency potential
Solar was resistant to AI change and a colleague was pushing hard to get them to adopt it in a more open way but the communication was hitting a wall. I volunteered to produce some material and format it in a way that would make it clear to senior management how AI could benefit us and how it wouldn’t need to cost jobs but would improve output.
The plan was not to create a final researched document on where the savings could come from but to try and mock up the final result in order show the process and convince them to allocate resources to research and mapping off job roles and efficiency potential. As a result I formulated a structured plan of how we would go about a pilot scheme and then role out if it was successful
Workflow
After pitching the idea to my managment I laid out a plan to identify how we could pin down all the roles and all the efficiency savings. Given this was a quick task to win resources for further research it was a a theoretical plan rather than one that was carried out on live employees. This approach meant I was able to get relatable but general guessed data to generate examples of how we could approach the issues we were facing.
Role prioritization
Once all the tasks were mapped I proposed investigating which role were the highest priority which could mean most potential or most importance, it needed discussing but in order to do that some data on what the outcome of a priority list might be. This could easily be the AI mapping that this project outlined.
Task mapping
Once it had been decided what the prioritise would be then the goal would be to map the roles in to quarterly, monthly and weekly tasks so that we could break down what each employee did and how AI could help the do their job more efficienctly
Wireframe
With the data collected/estimated I was able to create a wireframe of the rough design I wanted to have to commuicate the data modelling so that the senior managers could accessa dn navigate around the findings in thier own time and get an understandng of the impact the work could have.
High fidelity
After finishing off the wireframing I moved to polish the design off so it would be paletable for the managment presentation and we could sell the proposal in with a professional and knowledgable feel
Strategic analysis
To identify which roles within the company could benefit most from AI, I began by mapping out the company's structure — looking closely at the departments and teams that make up the organisation and how they operate. From that assessment, I compiled a list of potential roles within Solar that stood out as strong candidates. While there were other roles across the company that could also be considered, the ones included here represented the clearest opportunities for high-value impact.
Mock interview
Whilst there were no interviews conducted in the Proposal there would be extensive interviews needed in order to get the desired results. In order to understand the roles and tasks the empoyees had to undertake it would be vital to get into their world and understand the challenges they faced so we could design ai systems to combat their struggles. I identified 18 potential key roles to investigate
Data scientist
Data from diverse suppliers (e.g., Schneider, Grundfos, or local private labels) often arrives in inconsistent formats. One supplier might label a part as "20A Circuit Breaker", while another uses "CB 20 Amp 1-pole".
Insights
- This creates duplicates in our ERP, causes search failures on our webshop, and leads to picking errors at the Vejen warehouse. Currently, Data Analysts spend countless hours in Excel performing manual "VLOOKUP surgery" to normalize these descriptions.
Financial Accountant / Controller
Currently, Financial Controllers and AP staff spend hours cross-referencing multi-page PDF or paper invoices against ERP records.
Insights
- SKU Complexity: Suppliers often use different naming conventions or internal codes than Solar’s master data.
- Volume Friction: High-frequency deliveries to Vejen mean thousands of lines must be verified weekly.
- Human Error: "Fat-finger" entries during manual matching lead to incorrect payments, missing early-pay discounts, and strained vendor relationships.
Master Data Specialist
Master Data Specialists manually re-key specs, packaging units, and barcodes from hundreds of suppliers' unstructured files into our ERP.
Insights
- Manual re-keying is error-prone and creates a massive backlog.
- It pulls specialists away from higher-value Solve work like advising on complex product solutions.
Data Scientist / Analyst
Drafting SQL Queries
Cleaning Inconsistent SKU Descriptions
Debugging Legacy PythonR Codebases
Summarizing Monthly Sales Performance for the Board
Cross-sell Opportunities for Installers
Building PowerBI Dashboards for Department Heads
Validating Data Integrity between ERP and Lakehouse
Monitoring Stock-out Risks for the Vejen Warehouse
Wireframing
We implement a Retrieval-Augmented Generation (RAG) workflow.
High fidelity
Get in touch and let's work together :)
Currently available for freelance opportunities and creative partnerships worldwide. Let's build something exceptional together.