Deploying The AI Operating Model – 3 Key Strategies

by | Jun 20, 2024 | Insights

The evolution of operational excellence has entered a new phase as AI and digital technologies have matured swiftly and are enabling businesses to increase agility and quality, while reducing costs.  Large companies and early adopters are already learning and applying AI to build competitive advantages in their operating models which drive how value is delivered to customers. The early movers will expand their reach into previously uncompetitive markets and products with the scale and cost advantages of AI and digital technology eroding the profits and market share of competitors.  AI and digital capabilities are especially critical to small growth and mid-sized companies to defend and increase market share.   The central question for many businesses is what and how do you apply the right set of technology and practices to resolve business challenges while yielding the highest return on investment.  There are three key operational strategies that enable businesses to fully capture the power of AI and digital technologies.

  • Pilot and Deploy AI Models – machine learning (ML) and deep learning (DL) AI models are now outperforming traditional statistical methods for most business and technical analysis.  These performance improvements lead directly to lower operational costs, improved quality, and higher customer satisfaction.  The performance gap also increases as more data is collected and processed.  Deploy AI Pilots using agile methods, rapidly and iteratively, to foster learning, build capabilities, and manage technical risks.  To clarify terminology, Artificial Intelligence (AI) is the overarching field of building machines or algorithms to imitate human intelligence.  ML is a subset of AI and are algorithms that learn to make predictions based on data.  DL is a subset of ML and are algorithms based on multi- layered neural networks that can learn more complex patterns in data.  GenAI assistants are also providing valuable design insights and code snippets when developing specific AI models, increasing developer productivity and speeding deployment.  AI operating model use cases include but are not limited to demand and financial forecasting, product and process quality optimization, predictive maintenance, and sales, inventory, and operations planning (SIOP).  
  • Deploy Automated Flow Operations (AFO) – when optimizing and streamlining operational processes leveraging lean, flow principles and low-code technologies such as Robotic Process Automation (RPA), improvements are rapidly implemented.  Typical operational cost savings run 30-50%, productivity increases of 3-4x, and error reduction over 90%.  Similar to AI models, GenAI assistants provide valuable code or flow snippets when developing RPA work flows, again increasing developer productivity and speeding deployment.   When properly designed, AFO not only automates tasks across end-to-end processes such as Sales Forecast-To-Cash but also monitor and report Key Performance Indicators (KPI).
  • Implement an Innovation Management System (IMS)  –  continuous innovation of products, services, and operational processes has become a strategic necessity for many businesses.  Management systems are defined as a set of policies, procedures and processes used by an organization to meet business objectives.  Traditional hierarchical management systems are designed more for efficiency and perform best on managing established products and services.  The IMS is more of an entrepreneurial network that integrates and operates in parallel with the existing management system and has the necessary autonomy to meet key transformation objectives.  Enterprise or Scaled Agile frameworks are common frameworks utilized in an IMS and offer an excellent starting point and an essential tool in deploying AI and digital technologies.

Pilot and Deploy AI Models

An enterprise agile method works best when developing AI models, allowing the teams to learn how to best apply the technology, manage risk, and implement changes for new or existing products and processes. Establish internal or external benchmarks to compare existing methods and processes with AI models.  For example, in sales demand forecasting, measure the accuracy of traditional methods like Croston’s versus AI ML and DL models.  The benchmarks will illustrate the advantages of AI and its business case, for example, how even small gains in accuracy lead to optimizing inventory across SKUs, lowering holding costs, while increasing service levels and revenue capture.  Better demand forecasts also lead to more accurate financial forecasts, improving cash flow and resource management.  For manufacturing companies, unscheduled downtime can severely impact the efficiency of operations and customer satisfaction.  AI predictive maintenance models analyze motor current and vibration data to better forecast equipment failure allowing planning of preventative maintenance during scheduled downtime.   AI models also detect patterns in data not recognized before leading to further business insights for teams to capitalize on.  As AI initiatives expand, plan for data management requirements to increase in order support the additional data collection, cleansing and enhancements needed.

Implement Automated Flow Operations

Automated Flow Operations is an optimal blend of best practices and technology to increase productivity and throughput, lower operational costs, and reduce defects.  Continuous flow operations are designed to move work products seamlessly through each stage of the end-to-end process with minimal interruptions.  Robotic Process Automation plays a key role in automating any manual repetitive tasks, tracking and monitoring flow, and reporting performance metrics.  RPA technology (e.g. MS Power Automate) integrates across virtually all software systems (ERP, email, spreadsheets, web) making it an excellent solution to implement flow principles.  For example, RPA flows can monitor Forecast To Cash process bottlenecks and send reminders or emails to responsible individuals or teams and escalate to management when thresholds are exceeded.  Also, RPA flows can scan data for missing or incomplete fields and either populate the correct values or notify the responsible team.

Implement an Innovation Management System     

The IMS is used to manage the design, development, and implementation of innovative product and process improvements.  It is a separate organization that is created from existing resources and given the autonomy to meet innovation objectives.  IMS is optimized to innovate rapidly, fail fast, constantly evolve and iterate until solutions are ready and tested.  AI and digital technology introduction into products, services, and processes is an excellent example use case for IMS.  A leading enterprise agile framework is SAFe from Scaled Agile Inc. and has a comprehensive library of methods and tools that are well maintained and continuously improved.

Digital Operational Excellence

The three operational strategies discussed above are critical to realizing the next level of operational excellence being enabled by AI and digital technologies. They complement and integrate together for a more rapid transformation and realization of benefits.  Implementing an IMS will ensure the organization can design and operate new AI models and automation while also sustaining the improvements.  Automated Flow Operations will streamline and optimize processes, reducing waste and errors even as new products and processes are rolled out.  AI models will generate better forecasts, improve quality control, optimize inventory, and provide additional business insights not available with traditional methods.