2019Case19BAMIreland

73 L EAN C ONSTRUCTION I RELAND A NNUAL B OOK OF C ASES 2019 Through BAM’s 2020 Digi tal Vi s ion for the implementation of a digitally enabled workplace, we have been able to realise the key tenants of Lean thinking to: • increase the flow of work and information; • reduce the waste of our traditional practices; and • continuously improve our processes and the skillsets of our staff and wider supply chain. In this case study, we focus on how BAM Ireland is using cutting-edge AI technology to build on our current digital experience, to reinforce the lessons we have learnt to date, and to share these learnings across BAM Group. Construction IQ – AI Pilot Construction IQ is a web-based application with an access ible graphical inter face that di splays project information in the form of graphs and prioritised lists. Transitioning from a traditional paper-based system to a digital system reduces much of the waste of transporting, processing, and communicating the information. Where paper checklists for quality and safety were once stored locally on shelves or boxes, or construction issues were kept in a someone’s head or notebook, all this information is now instantly accessible and available to the relevant parties online and in the cloud. One of the often-neglected consequences of such improvements in the reduction of paper is the massively increased volume of digitized data and information that is now captured and available to use. Where information was previously lost through the cracks of the process, it is now captured, and this creates a new problem for the user – what information in this vast sea of information is most relevant to me right now. This is where AI is the right tool to assist in not only filtering through the data to “Sort” and “Set in Order”, but also to look for patterns to “Standardise” how the information is displayed. While the 5S work very well within manufacturing, the S of sort, shine, and standardise are the most appropriate for construction from our pilot. Figure 1. Project Address Issue Data Example Project Address Straight “out of the box” Construction IQ highlighted that we had not implemented a standard approach when setting up our projects in the B360 environment. We had not been diligent in entering the address details (they did not seem relevant at the time), the IQ map locations were incorrect, and the projects were incorrectly categorised by ‘internal BAM department’ or ‘project type’. This became our f i rst oppor tuni ty to shine our data model by backfilling all this basic information across our projects. Issues – Impediments to Project Delivery During a construction project there can be thousands of issues that can affect the delivery of the project, and on a daily basis the management team must prioritise these i ssues for resolut ion and to put in place the best resolutions. “Setting in order” this sea of data is achieved through the dark magic of the AI algorithm. Using Natural Language Processing, the algorithm looks through the text that has been entered by the project team. Where the AI sees words that relate to a high-risk issue such as “leak” or “damp”, it associates this as a water-related risk with the potential of water-related damage. It then uses its data model to automatically “weight” the severity of what it finds using the context of the surrounding issue text as a reference and the status of the issue with reference to its due date. In addition to the automated assignment of the risk, the user can fine-tune the response of the AI utilising their own experience and professional judgment, and from this the AI is trained by their input and updates the data model for future reference to a similar event. This training of the AI – a form of continual improvement – builds a risk profile for BAM within the data model. Over time, this risk profile “standardises” BAM’s response to similar issues as they arise and shares the learnings across all of the BAM users of the platform. LEAN INITIATIVE IMPROVEMENTS & IMPACT Key to the successful delivery of every project is the resolution of issues in an efficient manner. Following the initial basic housekeeping improvement within the project set-up, Construction IQ quickly identified that we, BAM, were the greatest risk to our own success. When we looked at these projects, the number of unresolved issues was uncomfortably high and represented the key to the high-risk factor to BAM. Further investigation showed that our site teams were dutifully observing and logging the issues as they occurred wi thin the B360 Field s i te management appl icat ion. Figure 2. Open Issue Data Example