Introduction

A Colne Client, ITENG Translations ltd is a business specializing in medical translation and employs chartered linguists to translate various academic, medical and specialist corporate documents from source (Italian) to target (English) on behalf of clients.

Pieces of work are quite standardised in format, and generally are procured by clients through 3rd party agencies.

ITENG completes dozens of these tasks each week for different agencies and end clients; these are won and delivered to the client through a variety of interfaces, with different invoicing and payment arrangements managed through Xero (accounting software) and different currencies (GBP, EUR, USD).

Problem statement

ITENG were using an Microsoft Excel spreadsheet to manage their workload; Excel, while robust for data manipulation and analysis, falls short when compared to more flexible Customer Relationship Management (CRM) tools like Notion for task tracking and productivity. Excel's grid-based structure can become cumbersome for managing dynamic tasks and collaborations, lacking integrated features like real-time updates, customizable templates, and interactive databases.
Notion.com, on the other hand, excels in these areas, offering a highly customizable environment that adapts to various workflows, from project management to content calendaring. This adaptability makes Notion a superior choice for teams seeking a more intuitive and integrated approach to enhance productivity and streamline operations.

ITENG operate a lean business without any admin support other than outsourced accountancy services; therefore low value creating activities such as creation, tracking and invoicing of tasks is handled by translators who could be completing more valuable and billable work.
Additionally, there was no (straightforward) way to connect Excel to Xero for invoice tracking and creation.

Value Qualification

When making solution proposals for clients, Colne strives to quantify and qualify the value provided by each solution. In this case, we can consider the time taken by a translator to complete non-translation tasks as low-value activity.

Assumptions:

In the case of ITENG, admin tasks such as job creation, invoice creation and tracking take approximately 30 mins each day for each translator. We can realistically reduce this to 5 minutes/day where non-standard invoices and human intervention is required.

Translation services are priced per word, and generally vary in rates depending on the subject, urgency, client and phase (type of translation task, from full translation to proofing).
ITENG translators generally bill around £47/hour for their their time.
Each translator works 5 days a week, 50 weeks a year (accounting for holiday)

Outcome & Value statement:

An administrative burden of half an hour each day reduces potential revenue by ~£5,500 per annum for each translator.

By automating as much of these processes and small admin tasks as possible, ITENG can free up time for billable translation tasks.

Service Delivery

This follows a phased delivery, with testing and acceptance by the client (ITENG) for each phase.

Phase 1

Notion Workspace, Google form and Make Automation

Phase 1 focuses on the main skeleton of the project, delivering the upfront requirements; a space to manage tasks and client/agency details in a structured way and a form to input tasks with a focus on speed and accuracy of input; whilst being flexible if a new client or agency is onboarded.

Google forms & Make.com Automation:

A Google form was created for task input, this is a basic free option; some more feature rich forms are available such as jotform.com.

A Make scenario picks up the form response, and populates the database(s) specified by the new task.
New Agencies and Clients are also added to their respective databases, with a follow up email (with a fun AI generated italian haiku) for translators to action invoice preference setups.

Make Automation — Make.com automation to add new tasks

Notion Databases:

Colne designed and delivered a Notion workspace to the client; the normalised database schema provides a more structured way to manage client and agency preferences, which utilises Notion's flexible powerful database features without the software engineering and infrastructure costs for traditional implementations. This can be iterated upon quickly if ITENG's specific requirements change.

Tasks:
- This is the main are where work is tracked, and created; this has many calculated and 'lookup' fields from the other databases.
Agencies:
- This is primarily used for invoice information, however it is helpful to know which agency the task belongs to, as they may have specific style or format requirements. Each entry in a database has it's own 'page'; which can contain rich text and images where translators can make useful notes or guides.
Clients:
- Used to keep track of client preferences around formatting, but primarily which agencies they tend to use.
Time estimates:
- By providing a words/hour for each 'phase' of the translation process, ITENG can see at a glance on their tasks board how long each piece of work should take; and the approximate hourly rate too.
Currencies:
- Much of ITENGs work is quoted in euros and US dollars, this is a simple lookup table to convert all work into GBP for better understanding of value/rates.

Notion workspace

Notion has a fully featured project and task management suite, used by companies such as Monzo, Pixar and Headspace for their product developments. It is heavily customisable but comes with some powerful built in 'views' for databases. (Pictures are from the notion website).

Timeline:

Board:

Phase 2

Replacement with fillout form

Whilst the google form was robust and worked, it had some drawbacks, specifically that it had to be 'polled' every 15 minutes leading to a delay in the creation of tasks - and that it was unable to use conditional logic. This also used a lot of make 'operations' to keep running. It also was un-integrated with notion, so we had to manually ensure that the agencies/clients list was up to date.

After which, we discovered fillout.com which has a generous free tier and integrates directly with notion, the form also allows conditional filtering and conditional logic, so we can show fields directly related to each client/agency.

Phase 3

Xero Invoice Checking

Tracking payments can be time consuming and switching between systems and tasks can break a translator out of 'flow state' so automatically archiving delivered and paid work items can save more time. This builds upon phase 1 to automate more tasks.

This consists of a Make.com automation to run once each day;

It checks tasks to confirm each task has been invoiced
These are then checked in Xero to confirm if they have been paid.
If they have been paid, the tasks are moved to archived to mark the payment as completed.

Phase 4

Xero Invoice Creation

Building upon the invoice checking, those tasks shown as 'completed' but without an invoice reference can now have invoices created using an integration with Xero.

Using Make.com, scheduled each day.

For each completed task;

Using the agencies database, confirm whether the invoice preference is per job, or in a monthly invoice; payment terms and extract the Xero contact id.
Using the Task data, and the linked agency invoice preferences, create the invoice using a make scenario.
Using a slack.com workflow/ integration, have the user/translator approve each invoice before it is sent.

Phase 5

AI- powered data extraction & task creation from emails

The final phase of automation, adding AI to the workflow

Connect make.com to the translation requests inbox.
When certain criteria are met, use OpenAI's structured output API
Create a task in Notion with a Make scenario.

Email data extraction workflow — Phase 4 process diagram

Automations & Task Management