Building Smarter: Unlocking the Power of Knowledge Management in Construction
Leveraging Emerging Technologies to Streamline Information, Enhance Decision-Making, and Drive Success
The age of Artificial Intelligence (AI) is upon us, with Large Language Models (LLMs) like OpenAI's GPT-4 and its interface, ChatGPT, revolutionising the technological landscape. These groundbreaking advancements have spurred the development of spin-off solutions and platforms, such as AutoGPT, that are capable of answering questions and performing tasks via Application Programming Interfaces (APIs).
In this rapidly evolving environment, enterprise knowledge management has emerged as a critical aspect for organisations seeking to harness the power of AI, knowledge graphs, and data pipelines. As the pace of change accelerates, it becomes increasingly challenging for individuals and organisations to determine the most effective solutions for their needs.
Gone are the days when businesses could rely on a single technology for years; adaptability has become essential for organisations aiming to leverage technology effectively. Industries like construction, often seen as laggards in technology adoption, must now be prepared to embrace continuous evolution and adapt to new solutions rapidly. As we move on we’ll explore the various technologies shaping the future of enterprise knowledge management and the strategies businesses can employ to stay ahead of the curve.
Image credit - Shelby Klingerman - Enabling Supply Chain Innovation with Emerging Technology
Laying the Foundation for Success
In this article, we will focus on helping construction enterprises prepare themselves to fully harness the potential of emerging technologies by leveraging existing resources and tools.
While AI is an enticing and rapidly evolving field, there are several challenges that must be considered before construction enterprises dive in:
Solutions are evolving so quickly that they may become obsolete almost as soon as they are adopted.
Information security, resilience, and intellectual property rights remain largely untested and murky in most cases.
Most enterprises are not yet prepared to maximise the benefits offered by AI technologies, as the answers provided by tools like ChatGPT may be generic rather than tailored to an organisation’s specific knowledge base.
Now is the time to focus on tasks that have long been overdue, such as organising internal knowledge, adopting open standards, and implementing digital mechanisms to structure and store valuable information.
Brain Power
Consider the knowledge within an organisation as its brain. This may include:
Policies, procedures, and standards
Best practices and lessons learned
Meeting minutes and transcriptions
Records of people, performance, projects, estimates, and financial/commercial data
Non-conformities
Models, drawings, and specifications
Method statements and risk assessments
In many organisations, this knowledge is dispersed, poorly structured, and difficult to share or extract insights from. By addressing these issues, construction enterprises can better position themselves to make the most of emerging technologies like AI, knowledge graphs, and data pipelines.
Embracing Openness
One effective strategy to mitigate the challenges posed by constant technological change is to base company and project standards and processes on open standards, such as those developed by buildingSMART and international standards. Open standards are publicly accessible guidelines that promote compatibility and interoperability between different systems, platforms, and software. They are a focus of a previous article, linked below:
Adopting open standards brings several key benefits:
It enhances interoperability, allowing different systems to work together more seamlessly.
It reduces vendor lock-in, giving organisations the flexibility to switch platforms with minimal effort.
It fosters collaboration and knowledge sharing within the organisation and across the industry.
A practical example of the advantages of open standards is the use of UniClass 2015 for classifying documentation. By organising information according to this standardised classification system, organisations can more easily locate and connect documented knowledge both internally and with other industry players. This streamlined approach facilitates knowledge management and collaboration, ultimately helping construction enterprises make better use of emerging technologies
Knowledge Management
Redefining the Basics
The Oxford Dictionary defines knowledge as "the information, understanding, and skills that you gain through education or experience." In construction, we rely on our senses, experience, and the collective expertise of our teams to navigate the complex landscape of risks and opportunities in our field.
We draw upon our experiences and learnings to make informed decisions based on the information we receive. But can we improve this process? Can we accumulate and share knowledge in such a way that both organisations and the entire industry can benefit from the lessons learned? How can we prevent the repeated occurrence of costly, dangerous, or damaging mistakes?
Harnessing the Power of Emerging Technologies
Technology holds the key to enhancing our knowledge management capabilities. The question is, how can we effectively structure and store our acquired knowledge to make it easily accessible and useful for machines? By doing so, we can enable AI to work alongside our workforce, generating value and driving innovation in the construction industry.
Building a Digital 'Brain': A Step-by-Step Guide to Managing Knowledge in Construction Enterprises
I’ve broken this task down into steps. These are:
Identify the Knowledge Needs
Gather Knowledge, including existing/historical information
Process and Store Knowledge
Use and Share Knowledge
Identify Knowledge Needs
We pick up knowledge all the time, sometimes useful, sometimes not. Unless you’re in the business of turning a profit from pub quizzes you will want to try to make the knowledge you gain as an organisation is targeted. There is little point, not to mention costs and increased risk, of storing data and information that is not required in the future.
Therefore building in the main should be planned. Whether at personal, project, organisation, or industry level there are sources of information available on which to base this plan. For example:
Project Execution Plans
Project risk and opportunity register
Contracts and agreements
Business plans
Business objectives and goals
Key Performance Indicators (KPI), and Objectives and Key Results (OKR)
Roadmaps and Development Plans
By understanding the requirements of these sources a plan can be formed to identify knowledge gaps and the information and data which will fill them.
Gather Knowledge - new and existing
Once a specific plan is in place then data and information generating can begin, generating knowledge in our target areas. This is current and future data, but also the mountains of data an organisation already owns. There are a plethora of sources for information and data, some examples are:
Sensors and data capture
Performance measurements and metrics
Reports and publications
Meeting minutes
Research papers
Correspondence, including emails, letters, and messaging
Technical data sheets
If you look at this list you will notice that some of these are structured (spreadsheets, SQL databases) and unstructured sources (text, photo’s, video). To generate knowledge from each of these takes specific approaches to organising and storing, which we’ll cover next.
Processing and Storing Knowledge
There are two main types of knowledge that an enterprise needs to store: structured data and unstructured data.
1. Structured Data:
Structured data is organised into a standardised format or schema, making it easy to input, store, query, and manipulate. The challenge for enterprises lies in ensuring that data is structured correctly and that related datasets are properly connected for comprehensive querying. Some common storage methods for structured data include spreadsheets, SQL on-premise databases, and cloud databases.
Examples of structured construction data sources:
Customer Relationship Management (CRM) system data
Programme/Schedule data
Personnel records
Project Management System data (e.g., Autodesk BIM360 or Bentley Synchro Site)
Common Data Environment metadata
2. Unstructured Data:
Unstructured data encompasses everything else that an enterprise produces and consumes. Traditional data tools and methods struggle to make sense of unstructured data without significant preparation. The goal is to process the information so it can be structured and analysed at scale.
“The variety, velocity and volume of today’s unstructured data can overwhelm traditional data platforms built for structured or semi-structured data.” Snowflake - Best Practices for Managing Unstructured Data
Examples of unstructured construction data sources:
Contract documents (Works and Site Information)
Method statements
Meeting minutes and transcripts
Video recordings (CCTV, timelapse, meetings)
Photographs
Email
Social media posts and comments
To store unstructured data effectively, three main storage methods can be employed: NoSQL databases, Data Lakes & Warehouses, and Knowledge Graphs.
Data Processing: Converting Unstructured Data to Structured Data
To maximise the value of unstructured data at scale, it must be processed. This involves running data sources through algorithms to remove noise and apply a structure/schema before re-storing the data.
Examples of data processing techniques are Natural Language Processing (NLP) for text processing, and Computer Vision for video and photo analysis
Data Pipelines
A data pipeline is a method for taking raw data, either structured or unstructured, from its source and pushing it through a process to an endpoint, such as storage and/or analysis. Data typically must undergo some form of processing, such as cleaning, filtering, transformation, or standardisation, before being stored.
There are three main components to a data pipeline:
Data Ingestion:
Data is extracted from the source and ideally stored in a data warehouse or data lake before processing.
Data Transformation:
Processing jobs are performed on the data to fit the format of the endpoint storage or process. This process is typically automated to handle large volumes of incoming data with minimal effort.
Data Storage:
Transformed data is stored in a repository (e.g., SQL or NoSQL database, Knowledge Graph) where it can be accessed and used by stakeholders through business intelligence tools, automated processes, or machine learning processes.
NoSQL databases and Knowledge Graphs
NoSQL databases store data in formats other than relational tables, making them more suitable for semi-structured data. They feature a flexible schema, allowing the database to scale and adapt while data is being collected and stored. This agility makes NoSQL databases suitable for dynamic scenarios.
A subset of NoSQL databases is Graph Store, which houses data with relationships that can be represented in a Knowledge Graph, creating a network of connections. This enables a deeper understanding of objects and relationships across a subject or an entire organisation, providing insights and knowledge.
Knowledge Graphs are data structures that help understand relationships between items. They are highly scalable and flexible, consisting of nodes (objects) and edges (relationships). Data can be stored within both the items and the relationships, creating a 'map' of information that improves understanding and insight. Knowledge Graphs can be integrated with existing data storage within an organisation, as they can quickly ingest and process information.
Using and Sharing Knowledge
We've identified our need for knowledge, gathered and sourced our data, processed and stored it. Now, it's time to put it to good use.
In the past, an organisation's knowledge was primarily stored in documentation or, more often, in employees' heads. To learn how to operate the photocopier or build a balanced cantilever bridge, you needed to consult Bob down the hall, offer him a cup of tea, and ask questions or have him on the project. This approach posed challenges—if Bob was on leave, tasks could be delayed, or simple errors might occur in the absence of his expertise.
Documented knowledge offers some advantages—unlike employees, A4 binders or digital files don't take leave or move to another job. However, they can become outdated and require maintenance to stay current, with the risk of people using outdated information.
In the sections above, we've developed a digital 'brain' for the organisation. This 'brain' gathers information from various sources, processes and stores it, and continuously updates to reflect the current state of the business. This digital knowledge repository is more complex than relying on 'a Bob' or a library of A4 binders, but it offers numerous benefits.
Continuously Updated:
The digital 'brain' is constantly updated, ensuring that the organisation's knowledge remains relevant in a rapidly changing environment. This dynamic approach reduces the risk of outdated information and ensures that best practices are always accessible.
Leveraged:
Unlike Bob, who may not be known to everyone, the digital 'brain' can be accessed by all who need it. This way, everyone can access the latest knowledge to support their work, increasing efficiency and collaboration across the organisation.
Verified Against:
Traditional audits involve manual checks of past performance, standards, or procedures, which can be time-consuming and reactive. With a digital 'brain,' real-time monitoring and guidance help ensure that employees adhere to expected practices. Discrepancies can be identified and addressed promptly, and the knowledge base can be updated accordingly. This approach results in fewer errors and more effective resource utilisation.
By adopting a digital 'brain' for managing knowledge, organisations can stay up-to-date, leverage collective wisdom, and verify processes in real-time, ultimately fostering a more efficient and innovative work environment.
Accessing and Sharing Knowledge
With the digital ‘brain’ in place it can now be used. How this is done is dependant on the use case, available information, and outcomes required - this includes the use of emerging technologies such as AI/Machine Learning.
Below I’ve listed some of the ways of accessing and sharing:
Automation:
Building knowledge into digital processes to automate and reduce human interventions. This may be done through building APIs to be able to call the knowledge into algorithms or computer code.
Search and Recommendations:
Search functionality can help people find what they need. We can use the relationships between pieces of information (including documents) to provide recommendations of further knowledge which may be useful
Prompt and response support:
ChatGPT has ushered us towards a fundamentally evolved future with tech. Using similar technology our knowledge can become part of the training material for a Large Language Model (LLM) such as GPT-4 for fine tuning, allowing users to query the knowledge to get a broad response much wider than a single document
Analytics and Data Visualisation:
PowerBI and other Business Intelligence tooling has taken off in the industry. A well built graphical interface giving people useful insight into the organisations knowledge (performance, prediction, analysis) and allow them to make a decision by finding trends and patterns.
Wiki’s and Forums:
Collaboration tools like wikis and forums can be used to encourage employees to share their knowledge and expertise with others in the organisation.
Final Thoughts
Adapting to the rapidly changing landscape of the construction industry requires organisations to embrace digital transformation, adopt open standards, and create a digital brain to manage their knowledge effectively.
1. Embracing digital transformation involves understanding the potential of technology, being agile in adopting new tools, and nurturing a culture that values innovation. In this context, it is crucial for organisations to identify their digital maturity level, assess their readiness for change, and develop a digital transformation roadmap that aligns with their goals.
2. Adopting open standards, such as buildingSMART and International Standards, allows organisations to ensure interoperability and seamless data exchange across platforms. By classifying documentation in accordance with standards like UniClass 2015, companies can more easily find, connect, and share knowledge across the organisation and the industry.
3. Creating a digital brain for storing and managing knowledge involves a multi-step process that includes identifying knowledge needs, gathering existing and new knowledge, processing and storing structured and unstructured data, and using and sharing that knowledge effectively. This digital brain allows organisations to access up-to-date information, leverage insights for decision-making, and verify processes in real-time.
By implementing these strategies, construction companies can not only adapt to the evolving technological landscape but also stay ahead of the curve, maximising their potential for innovation, efficiency, and growth in the industry.