A world-leading logistics company underpinned by technology and innovation, dealing with imports and exports to/from and through the UK and across Europe. The business process up to tens of thousands of shipping invoices per day in the UK alone, through customised solutions, value-added warehousing, cold chain and other inventory solutions.
Management at the company foresaw a bottleneck in compliance operations would be caused by the impending deadline on Brexit talks between the UK and EU. They are faced with a hard deadline, related to external negotiations and with their current process would require tens of extra staff to be hired and trained before year’s end to handle the spike in workload.
This would place enormous financial strain on the organisation in an already uncertain time. Also, the extra time required to go through the hiring process and the extra workload placed on their existing staff in order to train incoming staff would likely result in high levels of stress and mistakes. Each mistake made in declaring customer information costs the company directly, through fines applied by the customs authority in the UK (HMRC). As an added challenge, the business process was being changed and optimised at the same time as the automation effort.
During initial consultation, we worked with the customer to identify a three pronged approach to the project, which could be analysed and developed in parallel. The three sub-projects we identified were a UK tariff and tax amendment bot, an EU tariff amendment bot and a GUI integration with in-house security screening software. To begin we formed a cross-organisation and cross-functional team with the customer. This ensured that all process owners were present and could lend their insights throughout.
Alphalake's RPA team began bot-building within a fortnight of initial contact, working to the agile methodology, building robust and flexible logic while the process and data analysis teams worked to develop the business logic and analyse the workloads that the bots would likely face. Our RPA platform, Automation Anywhere was deployed in a hybrid configuration for this project. This offered cloud flexibility whilst local bot agents were deployed on Virtual Machines customer side, to bring data integration benefits.
Our team shared capacity planning advice with the customer to ensure optimised bot licensing for cost-efficiency, with budget management in preparedness for production and predictable scale-up costs.
The business logic and potential limitations were solidified by analysis large client datasets of import data, to identify patterns in how the import measures are applied. The data analysis team, led by data science lead, Ravi Theja, then collected a large sample dataset from the HMRC tariff database to identify exceptions requiring human intervention. In all such exceptions the team identified unique, machine readable criteria which the bot uses to collaborate with the process experts in real time.
Given the close and rigid deadline imposed by Brexit, we focused on providing a working minimum viable product as rapidly as possible, refining this as we discovered business and technical solutions to the problems identified by our data science and process teams. Along this journey we encountered and resolved issues with:
- Decisions which the bot was technically capable of, but required human oversight.
- Regulatory and compliance issues with the customs organisations themselves.
- Dynamic business needs change, driven by similar problems from the teams with whom we collaborated.
The Value Outcome
We look forward to sharing the final rewards of this challenging project with you, pending customer permission and a full post – project analysis of bot performance. This analysis is performed as part of all of the projects that we undertake, in order to ensure that the business objectives have been met and will include:
- Analysis of FTE saved.
- Validation of calculated bot performance.
- Analysis of bot uptime and individual bot use.
Looking forward to updating you in the new year!