From messages to
operational workflows
Connecting email, quotes, orders and operational states in one clear workflow.
- Design lead
- Lorenzo Rogante
- Design scope
- Head of Design / Design Lead
- Collaboration
- Management and development team
- Designed for
- Teams handling customer requests and operations
The initial concept centred on an intelligent email client. The design direction shifted towards the work created by each message: its requirements, progress, exceptions and handover between people.
or explore the workspace
Key decisions
Work beyond the inbox
Requests become connected workflows, preserving the source message as the work progresses.
Phase and status
A quote can be ready, blocked or awaiting a reply. Its stage and condition need separate signals.
Specific human input
Missing information opens a focused intervention rather than making the team repeat the entire task.
Column = phase · Badge = status · Explanation = next action
Keep the source beside the extracted data
The source message sits beside the information extracted from it. This arrangement supports checking an interpretation and supplying a missing value without reconstructing the conversation elsewhere.
From message to work
Known: quantity, format, delivery · Missing: paper weight
Separate progress from operational status
Request, Quote, Order and Production describe the stage of the work. Status explains its condition within that stage. Keeping these signals separate distinguishes a quote awaiting input from one awaiting a customer reply.
One connected workflow
Known: quantity, format, delivery · Missing: paper weight
Explain why a workflow is blocked
A missing paper specification keeps the example request in Quote while complete work progresses. The board makes the exception visible without suggesting that all activity has stopped.
Paper weight is missing
Known: quantity, format, delivery · Missing: paper weight
Ask for the missing value in context
The detail combines the message, extracted fields and missing paper weight. The intervention is limited to the information needed to continue, with the original request available for reference.
The right input, in context
Known: quantity, format, delivery · Missing: paper weight
Define the boundaries of automation
Confirmed steps and more automated processing share explicit intervention points. Missing information and consequential actions remain visible; in this scenario, a customer reply is required before the quote becomes an order.
Review before advancing
300 gsm confirmed · Review and customer confirmation remain explicit
Carry the request history into production
The message, specifications and decisions remain connected as the confirmed order moves to production. The handover preserves the context needed by the next person, rather than creating an isolated task.
Ready for production
300 gsm confirmed · Review and customer confirmation remain explicit
Phase and status answer different questions
This case study presents professional product design work. Interfaces have been recreated for the portfolio using synthetic records and generic names. Interactive sequences illustrate design decisions rather than record the original application. The figures are examples, not measured project outcomes.
From a smart inbox to a workflow platform
A customer message can set an entire business process in motion. Understanding the request is only the beginning: a team also needs to know what should happen next, what information is missing, and whether the work can move forward.
My contribution to MailSense centred on that transition: helping define a product experience that turns incoming communication into structured, actionable workflows.
The challenge
MailSense supports teams managing customer requests, including requests that lead to quotes, orders and production. Handling large volumes of communication manually means repeatedly finding information, interpreting requirements and keeping track of progress.
The design challenge was to connect the original request with the work it generates. The interface needed to make progress, blockers and next actions understandable across a process that combines people and AI.
The AI uses company information, product catalogues and price lists to interpret requests and prepare the information needed to handle them. My responsibility was to shape how that capability became a usable product experience.
My role
As Head of Design / Design Lead at BackwellTech, I helped shape the product direction and took responsibility for the user experience and visual language. My work covered information architecture, interaction flows, interface design, reusable components, prototyping and collaboration with developers.
Product requirements were informed by management input and discussions with engineering. I translated those inputs into an experience centred on the work that customer communication creates. A central proposal was to evolve MailSense from a smart email client into a platform for managing workflows and production.
Three design decisions
1. Make the workflow the main unit of work
I shifted the organising principle from individual emails to the request and its progression. Communication provides the context; the workflow makes the operational work visible.
A team should be able to follow a request through the relevant steps, understand where it stands and recognise what needs attention. For production-related work, the model includes Request, Quote, Order and Production. MailSense also handles broader customer requests, so the experience needs to support work beyond production alone.
2. Separate the phase from the status
A workflow’s position in the process and its condition are different pieces of information. A quote may be ready for review, waiting for a client or blocked by missing specifications.
I proposed separating these concepts: the phase describes where the work is, while the status describes what is happening within that phase. The card then connects that status to a reason and a next action. For example: Quote → Blocked → Missing technical details → Add missing information.
This gives the interface a consistent way to explain progress and exceptions without requiring the user to reconstruct the situation from the email thread.
3. Make human involvement explicit
MailSense has an assisted mode, in which users confirm workflow steps, and an automated mode, in which the software handles the progression more autonomously.
I treated the points where a person needs to intervene as part of the workflow experience. Missing information, an unclear request or an exception should appear in context, with a clear indication of what is needed to proceed.
The system can normalise data and resolve some issues autonomously. When additional business information or human interpretation is needed, the interface has to make that handoff understandable. The two modes therefore need clear treatment of confirmations and exceptions.
Outcome
MailSense is used by customers. My contribution defined its UX and visual direction and proposed a workflow-centred model for managing requests, progress and exceptions.
The phase-and-status separation presented here is a recent design proposal to make progress, blockers and next actions easier to understand.
The work produced a UX and visual foundation for the product, a workflow-centred organising model, and a proposal for clearer handling of progress and exceptions.
What I would evaluate next
I would evaluate how quickly users can identify requests needing attention, understand a blocker and determine the next action. I would also check whether they understand when the system will proceed automatically and when their input or confirmation is required.