At Gulf Consulting, we help organisations improve operational performance by designing and implementing practical improvement strategies. We do this by supporting clients to define clear and measurable operating strategies, stabilise and standardise core processes, and deliver achievable performance outcomes.
Gulf Consulting was engaged to evaluate whether a pre-trained, cloud-hosted AI image classification model could be integrated into mine site operations to support workplace safety monitoring — without the cost and complexity of building a custom model. The project reflects a core capability of our practice: identifying where off-the-shelf pre-trained AI models can deliver measurable process improvements quickly, and where deeper investment is warranted.
The engagement focused on a question our clients frequently raise: can a pre-trained model, activated as-is, deliver enough value to justify integration into an existing process? To test this, we selected the AWS Personal Protective Equipment (PPE) detection model and paired it with a lightweight web application. No fine-tuning was performed, and no retrieval-augmented generation (RAG) was applied. The intent was to isolate the usability and performance of the base model as an integration candidate.
We designed and built a web-based image interface allowing operational staff to upload photographs of workers on site. The AWS PPE detection model analysed each image, identifying head, face, and hand PPE per person, drawing bounding boxes over the detected items, and returning confidence scores for each detection. Users could then export results directly to a formatted PDF report suitable for compliance review and record-keeping. The end-to-end workflow — from image capture to reportable output — was operational within a short timeframe and required no model training.
Testing revealed a clear performance profile for the base model:
The project reinforced several principles that shape how Gulf Consulting approaches AI integration for our clients:
An infrastructure construction client within an ongoing alliance established to deliver major rail infrastructure projects in the state of Victoria, Australia
As part of its statement of strategic intent to continue leading the alliance partnership into the next 5-year phase of major rail infrastructure project upgrades, the client decided to implement its critical strategic objectives by using the Hoshin (Policy) Deployment methodology.
Policy Deployment is a strategy development and deployment methodology that translates the voice of the customer into new products and business operating systems to drive reliable profit growth. At its heart, it is an organisational learning method and a competitive resource development system.
A series of ongoing workshops involving the senior leadership team and key managers analysed the current business and external environments to identify 3 key strategic objectives to focus over 12 months. Hypotheses were developed on what and how to achieve each objective together with a set of implementation projects, goals and measurable targets.
Deployment teams were created to work with internal and external stakeholders to plan and execute the activities and tasks needed. Monthly reporting tracking strategic execution was created using visual management techniques. Given the complexity and effort required across all 3 objectives and being the first time the leadership team was exposed to this methodology, 2 of the 3 objectives' activities and tasks were 50% completed within the first 12 months.
A freight logistics client with significant numbers of active mobile assets servicing over 400 customers and responsible for the movement of millions of tonnes of raw resources, retail goods, manufacturing products, and construction materials across Australia.
Increases in journey cycle times and challenges in increasing the number of services available to individual customers drove the need to better understand:
The objective was to make additional services available to move more products to grow revenues and broaden its growth ambitions.
The client was assisted in discovering, adopting and implementing a time usage framework based on the Overall Equipment Effectiveness (OEE) methodology and adapted for its industry operating model.
The framework was digitised using Internet of Things (IoT) technologies and equipment to identify, measure, report and calculate major locomotive and train delays experienced in the field in real time. Delays were also tied to key events and proximity relationships to identify likely root causes for management action.
An innovative user interface was developed to give an intense and rich visual representation of delay data and information to users and management.
The framework was tested and deployed in a key market over several key assets. The data and information collected drove the instigation of several improvement projects aimed at reducing delays due to external network providers preferencing their assets' use of the rail network and review of maintenance and provisioning processes to identify opportunities to reduce associated maintenance time delays.
A global mining client with significant underground operations in North America supplies base metals essential to the green energy transition and AI infrastructure development.
Advances in battery-electric mining equipment, combined with rising commodity prices for base metals used in green energy infrastructure and AI-related development, created the financial and operational case to expand development drilling and blasting at an existing nickel mine.
To help the client access deeper sections of the ore body within a strict timeframe, the Business Improvement team received project management and execution support across targeted improvement initiatives designed to accelerate development rates and meet planned production ore and stope blasting milestones.
The work focused on practical operational improvements, including streamlining underground material logistics, reducing ore and waste haulage cycle times to surface stockpiles, shortening bolting cycle times, and improving the availability of battery-powered mining equipment, including jumbos and bolters.
Nine improvement projects directly and indirectly supporting a target development rate of 24 metres per day were initiated, managed, and completed.
The engagement also strengthened the Business Improvement team’s diagnostic and problem-solving capability while helping establish internal governance structures for future business improvement project management.
An Australian Government Business Enterprise is delivering one of the nation’s largest pumped hydro projects, creating critical clean-energy infrastructure and a major source of renewable generation.
The project required a complex network of service and emergency access tunnels and a new underground transformer hall, delivered using drill-and-blast methodology. However, the scale and complexity of the works created significant schedule pressure, contributing to construction delays and cost escalation.
To support the client in replanning and controlling the revised drill-and-blast schedule, the Last Planner Methodology (a form of Lean Pull Production Planning) was introduced as a production control mechanism. This enabled crews to align daily execution with the revised blast plan while reducing the risk of further cost escalation. The approach was reinforced through integrated visual daily management boards that tracked drill-and-blast performance, highlighted deviations from plan, and identified containment actions and countermeasures for construction and related services teams to address in order to continue to meet planned blasts.
Daily shift-to-shift visual management gave teams clear visibility of completed work, incoming priorities, upcoming blasts, and constraints affecting current and future cycles. Defined actions and owners reduced the risk of unresolved issues carrying across shifts. A two-week look-ahead planning routine brought together all trades and services influencing the drill-and-blast cycle, improving coordination, schedule reliability, and confidence in achieving the replanned blast targets.