Automating Your Business

 View Only

  • 1.  Business automation and AI

    Posted Mon June 26, 2023 11:02 AM

    What are some successful real-world examples of integrating business automation and AI technologies to streamline operations and improve efficiency?



    ------------------------------
    Selin Moran
    ------------------------------


  • 2.  RE: Business automation and AI

    IBM Champion
    Posted Tue June 27, 2023 08:52 AM

    Here's a few that we've built over the years:

    1. Looking at a mortgage packet that was scanned in as a single giant document and using a classifier to break it into its components, then using NLP to extract the relevant metadata. You orchestrate this within the workflow and then use it to automate a number of processing steps.
    2. Prioritizing sales opportunities (i.e. work items in a pre-sales workflow) by propensity to buy and weighted value.
    3. Using propensity models to automatically build audiences for marketing campaigns (paired with marketing automation, fully automating campaigns).
    4. Forecasting demand for the thousands of products in a supermarket across all stores in a chain to automate reorder processes and balance stock levels.


    ------------------------------
    Eric Walk
    Director

    O: 617-453-9983 | NASDAQ: PRFT | Perficient.com
    ------------------------------

    Original Message


  • 3.  RE: Business automation and AI

    IBM Champion
    Posted Wed June 28, 2023 11:14 AM

    Selin

    We focus entirely on decision automation in our practice - using ODM, ADS and DMOE to automate high volume, transactional decisions. We find that almost all of these can be enhanced using machine learning and AI - there are some that are 100% compliance oriented (all rules) but that's actually pretty rare. We tackle these projects "DecisionsFirst" to make sure we understand the structure of the decision (not just the list of rules) and this reliably reveals the need for advanced analytics. For instance

    • A claims decision that uses a predictive analytic ML model to predict the likelihood of non-disclosure and integrates that with other explicit red-flags to recommend a fraud review
    • An origination decision that uses predictions of credit risk, life time value, likelihood of use of banking products at other banks and several other predictions as part of deciding if the loan is acceptable or not (more valuable customers require a less profitable loan but all loans must meet credit risk guidelines)
    • Next best offer/next best action decisions that use predictions of life time value and propensity to accept to prioritize offers while also using rules to make sure that the offers make sense given the customer's current products and are compliant.
    • Scheduling and assignment decisions that use predictions of arrival time and likelihood to get additional stops added to refine the assignment and scheduling rules.
    • Preventive maintenance systems that use both predicted problems and warranty rules to identify high value servicing tasks

    And so on. If there are rules, there are almost always also ML/AI needs. If there's an ML/AI need, there's always a rules need.

    We have some very effective workshops to help clients think through this too and I've written a book about it called Digital Decisioning:Using Decision Management to Deliver Business Impact from AI



    ------------------------------
    James Taylor
    CEO
    Decision Management Solutions
    Palo Alto CA
    16504003029
    james@decisionmanagementsolutions.com
    ------------------------------

    Original Message


  • 4.  RE: Business automation and AI

    Posted Fri December 15, 2023 07:06 AM

    I think the most interesting and commonly known example nowadays is "Tesla's use of AI in its Autopilot system is a prime example of automation in the automotive industry. This advanced driver-assistance system incorporates AI to enable vehicles to navigate with minimal human intervention, improving safety and efficiency."



    ------------------------------
    Kashif Yaqoob
    ------------------------------

    Original Message