One-Line Summary
Rebuilt the sales quotation process — from paper and Excel — into an online tool that supports showroom use, real-time pricing, digital signature, PDF storage, and downstream project handoff. A first step toward automating deal reporting and finance workflows.
After the tool went live and entered real showroom use, I continued iterating based on field feedback — adjusting device options, refining discount and payment terms, and improving the print layout — gradually turning a first-version tool into a sales workflow system that fits how the team actually works.
Background
One day, a sales colleague handed me two printed quote sheets.
Each listed estimated prices for two different scenarios: partial installation versus whole-home installation. The intention was to give sales reps something to reach for during showroom visits — a quick way to show customers the price range and help them understand their options.
At first glance, nothing looked obviously broken. The sheets were solid and functional, consistent with how the company had always worked.
But the first question that came to mind was:
If I were the one doing the sales pitch, would I actually want to use this?
Pricing changes depending on device count, installation scope, application context, and discount terms. When all of those variables still depend on manual entry, paper confirmation, Excel edits, and follow-up handoffs, the process accumulates a lot of room for error, repeated work, and information gaps.
And realistically — sales reps in the showroom aren’t sitting at a desk calmly opening Excel. They’re holding a tablet, demoing products, and having a live conversation about needs and budget.
So I started thinking: if this quote sheet were a tool instead of a document, what would it look like?
My Role
This project didn’t start as a formal assignment. It started from a daily observation.
My role included:
- Understanding how sales actually use quotes in the field
- Translating the Excel pricing logic into a web-based tool
- Collaborating with AI to build a working prototype
- Filling in process details based on needs from sales, clients, engineers, and finance
- Redesigning the visual layout and information structure using the brand guidelines
- Helping the team think through downstream workflow automation possibilities
What I Observed
The original quotation process had a few obvious friction points.
One: the process was still mostly manual. Prices, device counts, discounts, and notes all required manual input and adjustment — time-consuming and prone to version mismatches or calculation errors.
Two: the format didn’t fit how showrooms actually work. Sales reps in the showroom are typically using a tablet to demo product scenarios. A tool that doesn’t fit that context forces a constant context switch between “showing the product” and “managing the quote.”
Three: the quote affects downstream handoffs, not just the customer. After a deal closes, there’s still site survey, equipment planning, installation, invoicing, and payment confirmation. Without enough structured information in the quote, everything after that depends on verbal handoffs or the sales rep filling in gaps manually.
Four: the visual design was outdated. A quote is a brand touchpoint. What the customer sees isn’t just a price list — it’s a first impression of whether the company is professional and trustworthy.
What I Built
I didn’t want to just make a “web-based Excel.”
The original formulas and pricing logic were already there — that part could carry over. But once the tool became online, it should actually fit the context of a live sales conversation.
Earlier this year, I took a Python course offered by NTU’s computer science department. I can’t write a full system from scratch yet, but I now understand programming logic, how data gets processed, and how to break down requirements into tasks I can work on with AI.
So I used Claude to translate the Excel spreadsheet into a web-based quote tool. The original formulas and pricing logic already existed in Excel, so that part wasn’t the hardest. The harder part was: now that it’s online, what should it do that Excel couldn’t?
Key Design Decisions
1. Tablet-first showroom experience
I assumed sales reps would be using a tablet in the showroom. Everything — button size, field layout, information density, pricing display — had to account for that context. The customer might be watching the screen at the same time, and the sales rep needs to quickly switch between device counts, packages, and discounts.
The goal wasn’t to make quoting more complex. It was to make it less friction in front of a customer.
2. Real-time pricing and discounts
The tool is a calculator. Sales reps can adjust device counts and apply discount ranges, and the total at the bottom updates immediately. This lets them explore different combinations on the spot, rather than going back to the office to revise a quote later.
In later iterations, I also added named discount presets, integer-rounded discount settings, and floor-rounding on the final amount — making the quote output feel more natural in a live conversation and more appropriate as a formal customer document.
3. Digital signature and PDF export
Once the quote is digital, the signature shouldn’t go back to paper. The tool includes an inline signature field — the customer signs directly, and the signed document saves as a PDF. Sign, confirm, and archive on-site.
4. Sales rep profile and draft management
Multiple reps use the same tool, so it includes a rep name selector that pulls in the corresponding contact info and signature. Drafts auto-save, so a quote doesn’t have to be completed in one sitting.
5. Date picker for quote validity
Instead of manually entering a number of days, sales reps now select an expiry date directly and the system calculates the validity period automatically. This is more intuitive than counting days on the spot, and reduces the risk of mental-math errors and inconsistent dates on final documents.
6. Field notes for project handoff
The quote isn’t just for the customer — it also goes to the engineers handling site survey and installation. I added optional notes fields where sales reps can fill in room locations, specific devices, and context. Those notes appear directly on the quote and reduce the amount of verbal or manual handoff afterward.
This turns the quote from a pre-close document into something that carries information forward into the next stage.
7. Visual redesign
Rebuilt the layout using the brand guidelines, and reorganized the information hierarchy. The goal wasn’t to strip it down — it was to keep all the necessary information while making it shorter, clearer, and more like what a professional brand would hand a customer.
How AI Fit Into This
A large part of this project involved AI collaboration, but I wouldn’t describe it as “AI built me a tool.”
The biggest thing AI helped with was compressing the time between an idea and a testable prototype. Whether the tool would actually be useful still depended on everything before that: understanding who uses it, where, what’s getting stuck, what information can’t be missed, and how it needs to fit into how the company already works.
For me this is a pretty typical field problem approach — understand the flow first, then use the right tool to put it into a more workable shape.
An Unexpected Extension
After finishing the first version, I walked the sales team through the tool. Not long after, finance joined the conversation.
It turned out finance had already been planning to rework part of their process — right now, after a deal closes, someone still has to manually write an email to notify the relevant people and confirm invoicing and payment.
The online quote I built could become the starting point for that automation. Once the quote is structured data, there’s a path toward connecting deal reporting, finance notifications, and invoice confirmation.
Sometimes digitization doesn’t start from a large system plan. It starts from a small tool that happens to sit at the right point in the workflow.
Status
The tool is in active showroom use and has accumulated feedback from real quoting situations. Based on that operational experience, I recently completed a second round of improvements: filling in missing device options, adjusting discount and payment term settings, refining the print preview layout, and updating the quote header to better match the tone of a formal customer document.
This also made something clearer to me: this project was never just about moving Excel to a webpage. It’s about gradually shifting the quoting process from “knowledge held by people” to “logic that a tool can carry.”
What I Learned
At the start, this project looked like a quote sheet. Once I got into it, I saw it was connected to a lot of other things — how sales reps handle customer conversations, how customers understand pricing, how engineers receive handoff information, how finance confirms payment.
This project also made clear to me that for AI collaboration to actually work, someone has to first think through how the data is structured, what the inputs and outputs are, and what the real operating context looks like.
Having taken a programming course, I understood that once logic can be cleanly broken down, it can be implemented in code. For example, when a sales rep selects a certain device, the system can automatically include required accessories in the calculation — turning what used to be a rule held in someone’s memory into logic the tool can handle.
Once the tool entered real use, sales feedback surfaced details I never would have anticipated. For instance, some customers haven’t received their unit yet when they’re getting a quote, so they can’t decide on switch panel colors right away. Based on that, I added a “color TBD” option and a way to display the available color choices on the quote — so customers can refer back to it later.
The whole project worked in a cycle: digitize the existing quote logic, gather real feedback from the field, identify what design changes would make the tool more rigorous and easier to use, and iterate. That loop is what turned it from a first draft into something sales reps actually reach for.
Recently, sales reps who weren’t part of the original development have started asking how the tool was built and sharing positive feedback. That meant a lot — it’s a sign that it’s been recognized by people using it in the real world, not just the ones who were involved from the start.