EXCLUSIVE OFFER: Save 12% on all AI Certifications. Exclusive for Job Seekers. Offer Ends on May 31, 2024!
Use Voucher Code:  UNEM12AIC .
The Evolution of ERP from Servitude to Liberation via AI and the No-Code Revolution/ai-insights/the-evolution-of-erp-from-servitude-to-liberation-via-ai-and-the-no-code-revolution

The Evolution of ERP from Servitude to Liberation via AI and the No-Code Revolution

April 10, 2024

The Evolution of ERP from Servitude to Liberation via AI and the No-Code Revolution

ERPs started from a necessity to create one source of truth, especially in the financial realm in which compliance is imperative to stay in business. They grew into systems that provided functional processes as executives experienced FOMO but could not afford to develop non-core functionality on their own. Businesses saw what innovators could do. For example, UPS figured out decades ago that right turns are more efficient than left turns and that therefore routing with mostly right turns led to more efficiency and a stronger bottom line. Technology manufacturers used robotics in warehouses with temperature control and lighting decades ago.

Technology businesses were happy to oblige the FOMO and create economies of scale for businesses by licensing non-core functionality; ERP, TMS, WMS, YMS, S&OP, HRMS, CRM, MMS, etc. You name the function and there is a system acronym. The problem is that these systems continue to exacerbate a siloed organizational mentality as processes are linear and in which users access processes and data only for the silo in which they reside. It is likely that each functional silo also creates its own one source of truth. 

As we think of ERP in the future, What If, we progress from Artificial Neural Networks (ANN), through Convolutional Neural Networks (CNN) and Recurrent Neural Networks (RNN) to Liquid Neural Networks (LNN)...

  • What If the ERP of the future is not functionally oriented, but rather data-oriented?
  • What If the purpose of the ERP of the future is orienting data structure and databases such that...
    • Businesses can choose from a library of automation based on business goals with AI-empowered statistical methods that learn how to improve the outcomes.  All this by utilizing Natural Language Processing (NLP) through Large Language Models (LLMs).
  • Allowing business to focus on
    • Validating rather than doing
    • Relationships rather than tracking
    • Products and sourcing rather than purchasing
    • Supply, demand, and price rather than forecasting
    • And the list goes on

Then: Analysts Galore

While Amazon and UPS, for example, mastered eCommerce, warehouse automation, and logistics ahead of other businesses, others struggled to catch up. Less than a decade ago, the industrial landscape was undergoing a significant transformation as companies moved away from traditional paper-based systems or static PDF marketing collateral to embrace the digital realm and eCommerce. Separate efforts to forecast, simulate, and model business financials and operational flows, revenue, cost, etc. were performed by business analysts using proven age-old statistical techniques such as classification and regression, clustering, association, search algorithms, and Monte Carlo analysis to bring accuracy and foresight to decision-making processes. These were cyclical, likely monthly, or project-based efforts. Automation was viewed as macros created within an application, or scripts, created, and run by IT programmers for interaction between programs. These methods, coupled with the digital transformation to eCommerce, are the underpinning of what started the AI evolution and revolution soon after.  Until then, AI, to the business masses, was still theoretical and left to innovators and early adopters.

Fast forward three to five years, and the industry was experiencing the initial stages of incorporating machine learning into crucial areas such as pricing strategies and customer churn prediction, utilizing the RFM (Recency, Frequency, Monetary) model.  SEO was a real and important function. The quest for defining product attributes through AI became front of mind as the labor-intensive Master Data funnel constrained revenue growth or decreased margin as the precision of aligning meaningful attributes with images took precious time. Optimizing SEO was born and almost immediately arcane.

Companies established dedicated teams comprising Master Data experts, Data Scientists, and Report Writers to tackle these problems and others. The tools of the trade included programming languages like Python and R, along with visualization tools like Power BI or Tableau, all as tagalongs to ERP systems and other operating systems.

This period posed additional challenges as common libraries and standardized code for AI/ML models we just in development. Robotic Process Automations (RPAs) were in their infancy, struggling with fundamental commands like right-click functionalities.

Natural Language Processing was the bane of large companies as comedy skits were based on their customers yelling into their phone, “Agent, person!”  And chatbots were just people halfway around the world responding in real-time to questions.

Now: The Era of Low Code and AI Libraries

In the present, the landscape has evolved significantly. The introduction of low-code platforms has democratized the development process, enabling a broader spectrum of professionals to actively participate in building applications and workflows. Fear of Missing Out (FOMO) is palpable as companies must strive to stay at the forefront of technological advancements or be left behind. Yet, ERP and all the bolt-on modules still exist and continue to silo organizations and data. Most bolt-on modules now use Machine Learning (ML) within their function incorporating AI and ML directly into the workstreams and analytics outputs. Demand forecasting uses ML within S&OP modules. WMS uses ML for pick routing and stocking ABC analysis. TMS uses ML for fleet routing. This provides gains in operational efficiency, forecasting customer behaviors, financial outcomes, product demand, and much more but does not democratize data.

Libraries for AI/ML models, particularly those coded in Python, have become widely accessible, streamlining the development and deployment processes. RPAs have matured, overcoming initial process barriers, and demonstrating their effectiveness in automating routine tasks. Generative AI has made Natural Language Processing and Image Recognition commonly available. Chatbots are now AI-driven through Deep Learning, requiring validation checks as opposed to an army of customer service reps.

The processing speed of all these automation and machine learning through neural networks is far quicker than human beings.

Coming Soon: The No-Code Revolution and Evolution of ERP from Servitude to Liberation 

Looking ahead, the industrial sector is on the brink of a no-code revolution. The concept of writing extensive lines of code is gradually fading away as more intuitive, user-friendly interfaces, and copilots, emerge. Today’s FOMO remains a driving force and is a well-reasoned fear. Every day a business hesitates to embrace these emerging technologies is a day in which a competitor may gain a significant edge. There will come a point when, rather than buying siloed process-driven technology, creating your own from well-staged data and a library of automation will be cheaper, faster, and drive a more productive business and workforce.

The impending era of no-code solutions promises to empower even non-technical professionals to actively contribute to the development and implementation of AI solutions and automation, democratizing and liberating businesses from age-old systems and process constraints. The imperative for businesses is clear - adapt and embrace these innovations swiftly, for tomorrow's success depends on the decisions made today.