Issue #8: The Lost Enterprise Intelligence: Rediscovering Tribal Knowledge and Hidden Efficiencies
Robotic Process Automation, AI, and LLMs: Harmonising Automation, Intelligence, and Adaptability in Modern Enterprises
Welcome to Issue #8 of 'The Uncharted Algorithm,' where we delve into the evolving landscape of AI and Large Language Models (LLMs) and their impact on modern business operations, especially the aspect of workflows and knowledge sharing.
These technologies are revolutionising how we process and interpret vast amounts of data, offering unprecedented insights into patterns and trends. This capability is being used for enhancing efficiency and productivity.
Yet, the real game-changer is the proficiency of AI/LLMs in deciphering and emulating human-like nuances and the subtle, often overlooked tribal knowledge within organisations.
This advancement enables the automation of complex tasks, including decision-making and problem-solving, which were traditionally thought to be exclusive to human intelligence.
However, integrating these sophisticated technologies comes with its challenges. Robotic Process Automation (RPA), known for automating repetitive tasks, often falls short when faced with dynamic changes and exceptions, crucial elements in business operations.
This edition focuses on how the synergy of AI, LLMs, and RPA can overcome these hurdles. We're tapping into the hidden data and workflows, a treasure trove within organisations, waiting to be discovered and utilised.
Hidden Workflow Optimisation Opportunities
In the current business landscape, in order to unlock hidden workflows one should firstly think about modernising legacy systems and secondly focus on capturing tribal knowledge.
Each of these areas, though distinct, interconnects to create a holistic approach to enhancing business efficiency and innovation.
Modernising Legacy Systems
Legacy systems in businesses are often like old machinery – functional but falling behind in today's digital acceleration. The true potential locked within these systems is not just in their basic operations but in the vast reservoirs of data and un-utilised workflows they contain. Modernising these systems can be thought of as embarking on a journey of discovery, uncovering insights hidden beneath layers of outdated technology.
One study from 2017 highlighted that enterprises in the US and UK lose a staggering $140 billion annually due to disconnected data, with 76% of businesses reporting critical data trapped in legacy systems, unable to integrate with modern cloud services. This revelation underscores the immense hidden potential in data that remains inaccessible and underutilised within these archaic systems.
When businesses undertake the modernisation of these legacy systems, the initial costs are soon outweighed by the benefits. On average, the investment in modernising legacy data systems is recouped within a few months or a couple of years. This return includes improved developer productivity, enhanced operational efficiency, and optimisation of infrastructure and licensing.
Furthermore, data modernisation significantly boosts productivity and employee motivation. These improvements are crucial in today's competitive and rapidly changing global marketplace, where agility and efficiency are key to success.
In conclusion, the modernisation of legacy systems offers far more than a technological update; it provides a pathway to unlocking the hidden potential of data and workflows.
This process not only enhances operational efficiency and productivity but also strategically positions businesses for future growth and success in an increasingly digital world.
Capturing Tribal Knowledge
In the intricate ecosystem of modern businesses, there lies a largely untapped resource – the hidden workflows and unstructured data, reminiscent of the rich ‘oral traditions’ of past civilisations. This unseen realm, filled with tribal knowledge and subtle processes, holds immense potential for optimisation and efficiency gains.
Tribal knowledge refers to the unwritten, unspoken, and often implicit understanding, skills, and practices that exist within a group or organisation.
It is the collective wisdom of the organisation, encompassing valuable information that has accumulated through informal channels but remains undocumented and isolated from the rest of the organisation. This knowledge can include best practices, processes, procedures, and even unwritten rules and norm. Tribal knowledge is typically shared by word-of-mouth and can be a valuable resource for organisations if managed well. However, if it remains locked behind gatekeepers within an organisation, it can become a force that cripples the growth of a company
Internal knowledge-sharing sessions can offer a platform for employees to share their undocumented tips, tricks, and insights. These sessions, when recorded and transcribed, can turn ephemeral knowledge into a tangible asset. Incentivising documentation also plays a crucial role. By rewarding employees for codifying their expertise, businesses can foster a culture of open information exchange.
Moreover, the power of AI can't be overstated in this context. These technologies dive into the sea of unstructured communications, analysing emails, messages, and informal interactions. They not only structure and codify this data but also reveal patterns and insights that can revolutionise how a company operates.
Through these approaches, businesses can unlock the hidden potential within their operations, turning overlooked workflows and tribal knowledge into a driving force for innovation and growth.
This process not only preserves the invaluable wisdom accumulated over time but also aligns it with the company's strategic goals, ensuring that these hidden treasures contribute significantly to the overall success of the organisation.
Current Automation Hits Limits
So how do we effectively harness this newfound wealth of data and insights?
This is where the limitations of RPA become apparent. While RPA excels in automating routine tasks, it falls short in extracting and utilising the intricate, often unstructured knowledge and data hidden within organisations.
It lacks the cognitive ability to interpret, analyse, and act upon the nuanced information that resides in these uncharted territories.
Incomplete Process Automation
RPA is designed to automate specific tasks, not entire complex processes. It works best with structured data and may struggle with unstructured data, which is often found in hidden workflows and data sources. This limitation means that RPA may not be able to fully unlock the potential of hidden workflows and data.
Scalability and Maintenance
As RPA systems become more complex and handle a greater volume of processes, they can become more difficult to maintain and scale. This can be particularly challenging when dealing with hidden workflows and data, as these often involve complex and changing processes.
Data Integration Challenges
Organisations often have diverse and fragmented data infrastructures, making it difficult for RPA bots to interact with and extract data seamlessly. Legacy systems may use outdated technology or non-standard data formats, further complicating the integration process. When RPA bots cannot easily access, process, and update data, it hinders the automation of critical business processes, limiting the technology's potential to streamline operations.
Change Management
Implementing RPA can require changes to existing processes, which can be difficult for employees to adapt to. This can be particularly challenging when dealing with hidden workflows and data, as these often involve human nuances and undocumented knowledge.
So, while RPA has brought significant benefits to many organisations, it faces limitations when dealing with hidden workflows and data. To unlock the full potential of these hidden assets, organisations may need to consider more adaptable solutions, such as blending RPA with AI and LLMs, which can better handle complex processes, unstructured data, and human nuances.
Practical Applications of RPA and AI/LLMs
The combination of RPA and AI/LLMs can unlock significant opportunities in the enterprise, particularly in the context of tribal knowledge, hidden data, and workflows.
Below are some specific examples of how this synergy can be leveraged:
Customer Support
Many organisations receive a constant influx of customer support messages that need responses. An LLM can assist customer support representatives in generating replies to these messages. When used together, RPA can automate the process of sorting and categorising incoming customer queries, while an LLM can generate appropriate responses based on the nature of the query. This can significantly speed up response times and improve customer satisfaction.
Knowledge Management
Organisations with large datasets can benefit from LLMs for knowledge management. Employees can type in natural language queries, and the LLM, trained on the organisation's dataset, can provide an answer. This can be particularly useful for leveraging tribal knowledge, as the LLM can be trained on internal documents, emails, and other sources of information that might not be widely known or accessible.
Compliance Report Generation
Enterprises that are in the business of generating compliance reports can greatly benefit from the integration of RPA and AI/LLMs. RPA can automate the process of gathering and organising data required for the reports, such as financial data, operational data, and regulatory information. Meanwhile, LLMs can analyse this data, identify trends, and generate insights that can be included in the reports.
For instance, a company might need to generate a compliance report to demonstrate adherence to the Sarbanes-Oxley Act (SOX). RPA could automate the process of extracting relevant financial data, while an LLM could analyse this data to identify any potential compliance issues. The LLM could then generate a section of the report that explains these issues and suggests potential solutions.
Compliance Audit Support
Compliance audits are a critical part of many businesses, and they often involve a significant amount of data analysis and report generation. RPA can automate the process of gathering and organising the data required for these audits, while LLMs can analyse the data and generate insights.
For example, during a compliance audit, an organisation might need to demonstrate that its internal processes are in line with GDPR or HIPAA regulations. RPA could automate the process of extracting data from various systems to demonstrate compliance, while an LLM could analyse this data to identify any potential areas of non-compliance. The LLM could then generate a section of the audit report that explains these issues and suggests potential solutions.
Regulatory Change Management
Regulatory environments are often dynamic, with new regulations being introduced and existing ones being updated regularly. RPA can automate the process of monitoring regulatory changes and updating internal systems and processes accordingly. Meanwhile, LLMs can analyse the new regulations, understand their implications, and generate insights that can help the organisation adapt to the changes.
For example, when a new financial regulation is introduced, RPA could automate the process of updating the organisation's financial systems to comply with the new regulation. An LLM could analyse the new regulation, understand its implications, and generate a report that explains these implications and suggests how the organisation should adapt.
Key Takeaways
If you are an enterprise decision maker that is looking to extend your organisations capabilities beyond automation, or even if you are a beginner in the automation journey, here are some takeaways that might help.
If you are yet to start your journey it might be a blessing in disguise as you have an opportunity to leapfrog the competition, but there might be other cultural challenges that could be top of mind.
Rethink Automation Beyond Routine Tasks: Move beyond using RPA for just repetitive tasks. Leverage AI's cognitive abilities to handle complex, unstructured data and workflows.
Embrace Change Management: Understand that implementing these technologies requires a shift in organisational culture and processes. Be proactive in managing this change.
Focus on Scalability and Integration: Ensure that your automation strategy is scalable and integrates seamlessly across different business functions for maximum impact.
Strategically Leverage Data: Encourage a shift in perspective to see data not just as a by-product of business operations but as a strategic asset to be harnessed and optimised.
Prioritise Data Accessibility: Make data accessibility a cornerstone of your strategy. Hidden data locked in legacy systems is a lost opportunity; liberating this data can drive significant business growth.
Final Thoughts
There is a separate debate about whether AI/LLMs will eventually supersede RPA systems, including those integrated with AI features. I've been deeply involved in a project investigating this very topic. Our preliminary findings with cutting-edge LLMs suggest that it's too early to draw definitive conclusions.
For now, the synergy of RPA and AI/LLMs is where the action is.
However, as the capabilities of AI/LLMs continue to grow and stabilise at an enterprise level, this area is poised for significant evolution. I'm excited to delve into this subject in future editions of my newsletter.
Stay tuned for updates on this evolving topic!