UiPath Communications Mining brings visibility and acceleration to the Purchase to Pay mailbox
Within Espria’s Purchase to Pay (P2P) department, large numbers of emails arrive from suppliers, creditors and internal employees such as buyers. These emails are essential for a smooth purchasing and payment process, but also create a significant administrative burden because they are unstructured and end up in one central mailbox.
Together with Tacstone Technology, a project was therefore launched to organize the P2P mailbox smarter with UiPath Communications Mining (also called UiPath IXP). Instead of manually sorting and categorizing, this technology provides insight into email traffic and handles a large portion of emails automatically.
Why there was a need to use AI for email automation:
- ±150,000 emails per year with a variety of topics .
- Spent a lot of time manually sorting, prioritizing and forwarding emails
- Long response times, especially during illness/absence or vacation periods
- Work is seen as boring and repetitive
Technologies:
UiPath Communications Mining / IXP (AI classification)
Volume mailbox:
±150,000 emails per year
Result of the solution:
Insight into email traffic, less manual work, shorter response times
Reason
At Espria, all types of messages converged within the Purchase to Pay mailbox: supplier inquiries, order confirmations, invoices, specifications and notifications of delays. These messages varied widely in intent and required different actions or redirection. Because everything arrived in one place, employees had to spend a lot of time daily sorting, dragging, prioritizing and categorizing – actions that were repetitive and inefficient. And that’s even though following up on emails was often rule-based, and thus easily automated.
Approach
Implementation consisted of the following steps:
- Creating insight into email flows: Employees labeled hundreds of emails so the AI model could learn what different types of messages mean. This decoupled unstructured text from emails and made intentions and categories visible.
- Training process AI model: Training the NLP model was the biggest effort – employees labeled emails daily for several weeks to let the model learn what relevant categories are and when an email can be classified.
- Establishing mailbox logic: By establishing category structures, labels and workflows, it became clear what the main types of emails are and how to prioritize them.
- Insight leads to optimization: During the labeling process, process insight also emerged: it became clear where there was a lot of overlap in folders, which communication patterns recurred and how processes could be structurally improved.
This approach provides a foundation on which follow-up automations can be built, such as automatic assignment, notifications or integration with RPA robots for follow-up actions.
How does it work?
Functionally, the solution works as follows:
- Incoming emails from the P2P mailbox are automatically analyzed by UiPath Communications Mining.
- Classification and categorization based on intention, subject and characteristics.
- Dashboards and insights show which categories and communication streams dominate the mailbox.
- Labels and patterns are used to prioritize and reduce manual sorting work.
- Basis for automation: once classification is reliable, follow-up actions can be automated with RPA robots or other integrations
The process
Results of the solution
Although the solution initially brought mostly insight and structure, initial improvements are evident:
- Less manual sorting: employees spend significantly less time manually processing emails.
- Greater insight: automatic classification provides clear insight into communication flows and priorities.
- Foundation for automation: with clear categories and intent labels, future RPA robots and workflows can be deployed much more precisely.
- Process improvement: analyses showed that, for example, payment specifications are often unnecessarily requested as mail and that more efficient solutions are possible here.
Employees experience the technology positively and see the benefits of both the insights and the ability to automate even more steps over time.
How can you apply this in your organization?
Like Espria, do you want to gain insight and control over a busy mailbox within your organization? Contact Tacstone Technology for an introduction or demo and discover how UiPath Communications Mining can be the first step towards smart automation, less workload and more structure in your Purchase to Pay process.
