Follow us:
A customer sends a purchase order containing potentially hundreds of line items.
The workflow extracts the attached document and identifies the information required for processing.
The AI agent interprets descriptions and reconciles the customer's references against Webro's own product data.
Matched items are processed into a prepared quotation ready for review.
Webro remains in control. Once actioned, the prepared quotation can be sent to the customer.
Webro had an internal operational problem that had become difficult to solve through conventional development.
Customers around the world could send large purchase orders containing hundreds of parts. Those customers often used their own part numbers, while Webro used completely different references internally.
The descriptions might identify the same underlying product, but the reference numbers did not correspond directly. Somebody therefore had to work through the order, understand each line, determine the equivalent Webro product and prepare the quotation.
For particularly large purchase orders, that process could take up to 48 hours.
Vimana Digital is a member of the East Midlands Chamber, and the project reached us through a referral from our Chamber Business Advisor.
It was not presented as a conventional web project. Before discussing the client in detail, we were first asked to review the technical brief and establish whether it was something we believed we could solve.
After reviewing the requirements, we were confident in the approach and proposed a three-week delivery period.
We were then told that the project had previously been taken on elsewhere with a three-month delivery expectation, but had been returned after that team concluded they could not successfully deliver the requirement.
The challenge sat between documents, product data, purchasing workflows, matching logic and automated quotation preparation.
Webro later explained that both an external development team and internal technical resources had previously explored the problem without producing the required solution.
Vimana committed to a focused three-week delivery period and delivered the bespoke workflow within that timescale.
The difficulty becomes easier to understand with a simplified example.
A customer's system may call three products ABC, XYZ and PQR. Webro's own system might call those exact same products 1, 2 and 3.
There is no useful relationship between those numbers. The connection exists in the product description and surrounding information.
Humans can understand that relationship, but manually repeating that exercise across hundreds of order lines is slow and tedious.
Purchase-order information first had to be identified and extracted from incoming documents.
Customers do not necessarily structure references, descriptions and purchasing data in exactly the same way.
The system needed to establish which customer line corresponded with the appropriate Webro product.
Once the order had been processed, the relevant information needed to flow into the quotation process.
Instead of asking staff to manually repeat the same interpretation and matching process for every incoming purchase order, Vimana built a bespoke AI-powered workflow around the task.
The agent can work from the incoming email, extract the purchase-order document, process the relevant information and work through the product-matching stage.
Once the information has been reconciled, the workflow can prepare the quotation ready for the appropriate action.
The purpose of the AI is not to insert a chatbot into an existing process. It is to remove repetitive work from the process itself.
An incoming customer email contains the purchase-order document to be processed.
The relevant attachment is identified and moved into the processing workflow.
Purchase-order lines and relevant product information are extracted from the document.
Customer references are reconciled against the appropriate Webro product information.
Successfully processed information is used to prepare the quotation.
Webro can review and action the result before the quotation is sent to the customer.
A useful automated workflow should remove unnecessary human effort without removing human judgement where it is still valuable.
The system automates the repetitive stages involved in receiving, reading, processing and matching the purchase order before preparing the quotation.
Webro can then retain control at the appropriate stage, review what has been prepared and action the outgoing quotation.
Large purchase orders could previously require substantial manual processing, with particularly complex orders sometimes taking up to 48 hours.
Once the automated workflow was completed, the same type of process could be handled in around two minutes.
For a business receiving multiple purchase orders from customers around the world, the difference is not simply a faster computer task. It changes what the team can do with its time.
Following our initial meeting with Webro's director, we moved directly into understanding the actual operational process, exceptions and intended outcome.
Three weeks later, the bespoke AI workflow was ready to demonstrate.
The speed of delivery did not come from reducing the requirement. Every core requirement discussed with the client was addressed within the finished solution.
Bespoke development works best when the developer understands the underlying business problem rather than simply reproducing a specification line by line.
During the Webro project, that understanding allowed us to identify additional functionality that could make the finished workflow more useful to the business.
Those improvements were incorporated alongside the original requirements rather than treating the brief as the limit of what the project could achieve.
Unlike a website project, publishing screenshots of the finished system would reveal elements of Webro's internal operations, product data and bespoke workflow.
We therefore deliberately do not publish the live interface, source code, matching rules, internal data or customer information within this case study.
The diagrams shown here explain the type of problem solved without exposing how Webro's actual production system works behind the scenes.
Purchase orders, product matching, document processing, quoting, quality checks and repetitive operational administration.
Reconciliation, document classification, reporting, information extraction and exception handling.
Invoice processing, approvals, data capture, internal routing and recurring administrative tasks.
Document intake, information extraction, workflow routing, comparison and structured administrative processes.
Booking administration, communications, scheduling, reporting and operational coordination.
Orders, logistics, internal approvals, notifications, document handling and processes spanning multiple systems.
Webro now has a bespoke AI-powered process developed around a genuine operational requirement rather than a generic automation template.
Incoming purchase orders can be processed, their data interpreted, customer product references reconciled against Webro's own information and quotations prepared through one connected workflow.
A task that could previously absorb hours of staff time can now be completed in around two minutes for the type of complex order the solution was designed to address.
Just as importantly, the project demonstrates what becomes possible when AI is applied to a specific business bottleneck rather than introduced simply because AI is fashionable.
If your business has an internal process that is slow, repetitive, complicated or heavily dependent on manual administration, show it to us. Vimana Digital can investigate whether AI, automation or bespoke software can remove the bottleneck.
Discuss Your Workflow →