Chapter II

REVIEW OF RELATED LITERATURE

AI Assistants in Asset Management and Maintenance

As technology-enabled asset management solutions are increasingly

adopted by organizations, artificial intelligence (AI) has emerged as a

precious resource to improve asset tracking, maintenance work, and

predictive maintenance. AI-based assistants are transforming equipment

maintenance in institutions by facilitating efficient tracking, on-time repair,

and less downtime in operations. One way AI can be applied in asset

management is through the utilization of AI-driven diagnostics in facilities

maintenance and repair services. Maintenance management and tracking of

legacy assets typically depend on manual data entry, paper records, and

human verification, causing inefficiency and inaccuracy. AI-based systems

are able to offer automated data collection, identify patterns of asset

performance, and forecast failures prior to their occurrence. A top innovation

highly applicable is the application of AI-based QR code-based asset tracking

platforms that enable customers to scan a QR code to access immediate

information on asset details, review maintenance history, and request a

repair. With AI integration, the platforms can handle repair requests,

recommend solution protocols, and supply technician’s tutorial resources like

video tutorials.

Additionally, AI asset management assistants schedule preventive

maintenance to ensure that all past repair history, usage patterns at the

time, and weather are considered in trying to recommend the most optimal

maintenance cycles. This minimizes downtime, maximizes asset life, and

maximizes utilization of institutional assets, and optimizes utilization of

institutional resources. In spite of these advantages, AI asset management

tools have limitations. Internet reliance has potential to influence real-time

data access and diagnostics, and repair recommendations from AI await

validation by humans for accuracy and reliability. Cybersecurity solutions

also need to be covered for guarding sensitive institutional asset information

from potential threats. Overall, AI assistants are now well involved in

transforming asset management through computerized monitoring, repair

process optimization, and maintenance decision-making based on facts. With

further improvement in AI technology, its use in asset tracking and

maintenance systems will always improve efficiency, lower costs, and

enhance the reliability of assets as a whole (Anderson, P.,2023).

AI-Driven Troubleshooting in Software Operations

 As software systems become increasingly sophisticated, the need for

automated fault correction and problem-solving has grown exponentially.

Fault correcting solutions based on AI are transforming the manner in which

IT teams handle system crashes, reducing manual intervention and

improving operational efficiency. A wonderful example is Resolve AI, with the

goal of automating software with AI-driven solutions that can only

automatically detect, diagnose, and repair manufacturing defects. These are

created to eliminate downtime, stabilize the system, and have engineers

spend less time firefighting and more time innovating and designing.

Resolve AI relies on advanced machine learning algorithms, anomaly

detection, and predictive analytics to automatically detect problems before

they happen. Alongside cloud platforms such as AWS and coding repositories

such as GitHub, the system is exploring the capability to rapidly scan logs,

determine root causes, and repair issues automatically to reduce the amount

of time needed to fix problems in the interim. These solutions with artificial

intelligence can also logically triage alarms, separating high-priority failures

from low-priority anomalies in order not to cause alert storms for engineers.

One of the main issues is the capability of AI to correctly diagnose

sophisticated system failures in certain situations; human intuition is still

required to make sure that automated fixes do not wrongly diagnose

entrenched problems. In addition to troubleshooting, Resolve AI also intends

to utilize its abilities in other central areas of software behavior like incident
prevention, cloud cost optimization, and auto-scaling of IT resources. With

forecasts of future system crashes, AI can dynamically patch, configure, or

rebalance resources to yield optimal performance. Moreover, AI-based cost

optimization can also assist companies in keeping a check on cloud spend

better via usage behavior analysis and suggestion of reduction in wasteful

spending. Challenges apart, however, the various advantages of AI-based

troubleshooting persist.

Additionally, as minute-by-minute automation gets more intelligent

through the use of AI, IT support teams have to keep re-training models on

the basis of data from installed systems time and again to be effective. AI-

based troubleshooting in most cases is revolutionizing software operations

with fewer touches by humans, system stability, and pre-emptive cure.

Additionally, the functionality of an AI system depends on the quantity and

quality of the input employed. Garbage input data or faulty input data may

result in inaccurate diagnosis or inappropriate maintenance advice. With

companies such as Resolve AI at the forefront, the future of software

maintenance and IT operations is tilting towards an autonomous, intelligent,

and streamlined one (Lee, C. (2019).

AI Integration in Military Asset Management

 The use of artificial intelligence (AI) in military asset management has

transformed the way defense institutions maintain and optimize equipment,

vehicles, and infrastructure. The USS Fitzgerald is perhaps one of the most

extreme cases of AI deployment in military operation, as it now becomes the

world's first warship with AI. However, this technology represents a giant

leap towards the US Navy's vision of enhancing fleet readiness and

operational effectiveness. At the center of this technological revolution is

Enterprise Remote Monitoring Version 4 (ERM v4), an artificial intelligence

based software that applies machine learning algorithms to forecast,

diagnose, and mitigate maintenance failures before they manifest into full-

scale problems.

The predictive maintenance feature helps ensure that extremely vital

naval assets remain in peak working condition, limiting the occurrence of

unplanned system failures on missions. Through ongoing monitoring and

cross-checking against live performance data from the different components

of a vessel, ERM v4 is able to spot anomalies, compute wear and tear, and

suggest pro-active maintenance measures. Use of AI in military asset

management has a variety of strategic benefits. To start with, it reduces

downtime by quite a large margin through pro-active repair, so naval ships

and other military assets are ever mission-ready. Second, AI-driven

maintenance maximizes the use of resources so that time is concentrated

among military engineers and technicians for more urgent repairs instead of

spent on gratuitous routine inspections. Third, AI-driven diagnostics

maximize cost-effectiveness since early identification of impending issues

reduces the need for costly emergency repairs or replacement of parts.

Apart from naval use, AI is increasingly being used in other military branches

to maximize asset management and operational readiness. For instance,

predictive maintenance using artificial intelligence is being applied in Air

Force combat fighter aircraft, ground vehicles, and missiles to keep them in

the best operating state.

In addition to this, AI algorithms assist military logistics personnel in

enabling supply chain management so that essential supplies such as spares

and fuel reach frontline forces efficiently.

While such benefits are important, they are at the expense of the

challenges of implementing AI in military asset management. One such

major concern is AI reliability and security. Military assets need to function

under austere environments, and the use of AI has to be supplemented with

strong cybersecurity to avoid hacking, data intrusion, or manipulation by

outside parties. Additionally, the quantity and caliber of input used

determines how well an AI system performs. Inadequate or faulty input

information may result in incorrect diagnosis or inaccurate maintenance

suggestions.

Additionally, as minute-by-minute automation gets more intelligent

through the use of AI, IT support teams have to keep re-training models on

the basis of data from installed systems time and again to be effective. AI-

based troubleshooting in most cases is revolutionizing software operations

with fewer touches by humans, system stability, and pre-emptive cure.

Additionally, the functionality of an AI system depends on the quantity and

quality of the input employed. Inaccurate diagnosis or inappropriate

maintenance advice may arise from garbage or flawed input data.

Application of ERM v4 in USS Fitzgerald is a ground-breaking case of AI-

powered defense innovation, opening up for its wider adoption among global

military forces. As technology for AI continues to advance, its application in

military logistics, predictive maintenance, and combat performance will

become even more critical (Johnson & Smith, 2022).

RepairIt Application and AI-Powered Repair Diagnostics

The use of artificial intelligence (AI) in repair troubleshooting has, in

the past few years, been hugely popular owing to its ability to increase

efficiency, lower cost, and increase accessibility to repair solutions. One such

innovation includes the RepairIt Application, a smart AI app that was created

during the Gemini API Developer Competition. This tool leverages the vast

capability of Gemini AI to evaluate the reparability of devices and inform

users about how and when to repair their devices. Simply by entering certain

data on the device, users are presented with in-depth reparability scores that

dictate the ease or complexity of repairing a specific device. Additionally, the

app provides customized repair suggestions, which include DIY repair

manuals in simple steps, local repair shops access, and information

regarding replacement parts availability. This functionality enables

customers to make an informed choice on whether they should try a repair

by hand or hire the services of an expert. The RepairIt App supports

worldwide initiatives to advance the Right-to-Repair movement, which

campaigns for consumers to have increased access to repair information and

services. By providing users with AI-based repair quotes, the app not only

optimizes device lifespan but also helps in environmental sustainability

through minimizing electronic waste (Google AI Developer, 2024).

Artificial Intelligence in IT Operations (AIOps)

Where IT infrastructure is business-centric, such as in banking and

finance, healthcare, retail and e-commerce, and telecommunications, AIOps

is especially important. For example, in the banking sector, AIOps can detect

security breaches and fraudulent transactions in real-time, while in the

healthcare sector, it can monitor patient data systems so that it can obtain

access to life-critical medical records in real-time. AIOps fixes this by

providing informed, real-time suggestions based on insights that help

organizations detect anomalies, predict system failure, and analysis from

sources such as network traffic, user activity, system performance, and

application logs. AIOps solutions leverages machine learning algorithms to

detect patterns, correlation, and possible threats that may evade human

eyes. The moment an unusual surge in resource consumption occurs

suddenly, a probable security attack, or probable system crash in

anticipation of AIOps is detected, it can initiate auto-mitigation measures to

quarantine the issue without disrupting business processes. One of the

significant advantages of AIOps is that it assists in improving the efficiency of

IT by simplifying the workload of human IT staff. AIOps is also an important

tool in predictive maintenance where probable IT infrastructure crashes can

be predicted in advance.

Through constant analysis of history and real-time data, the system

can predict when hardware bits of equipment, software programs, or

network appliances are going to crash. It provides an ability for organizations

to pre-emptively address issues, preventing costly outages and realizing the

highest uptime for services. Cloud providers and large organizations also

employ AIOps in order to reduce wastage of resources, prevent server

crashes, and automate IT operations. While businesses are racing towards

digitalization, AIOps will even more be essential in providing strong,

effective, and scalable IT infrastructures (Smith, J., & Brown, T., 2021).

Predictive Maintenance in Asset Management

Is it an upscale asset management method that employs machine

learning, data analysis, and constant monitoring to forecast equipment

failure in advance? Predictive maintenance looks at the present status of

infrastructure and assets and makes a decision on replacing or servicing it,

as opposed to conventional preventive maintenance that includes periodic

servicing conducted at scheduled time intervals. Through this method,

companies can optimize asset life and reduce wasteful service costs just by
doing the maintenance where and when it is required. Predictive

maintenance plans track for patterns and trends in historic and real-time

data in the hopes of being able to forecast impending failures. Such basic

performance metrics like vibration, temperature, energy consumption, and

other performance characteristics are monitored by predictive maintenance

systems through sensors, Internet-of-things sensors, and machine learning

software. When faulty conditions are sensed, alert messages, and

maintenance personnel can respond prior to a break-down. Predictive

maintenance is not only more cost-effective to operate; it is safer in the

workplace as much as avoiding surprise failure, reduces catastrophic failure,

and no longer generates downtime that is expensive for manufacturing.

Additionally, predictive maintenance saves sustainability efforts

through maintaining maximum resource utilization with minimum wastage of

resources in terms of parts replacement that is unnecessary. Predictive

maintenance has been growing more and more noticeable in the transport,

manufacturing, health care, and energy industries in their attempts to

optimize asset reliability. For example, electric power generating power

plants utilize predictive analysis to guarantee that power is being produced

around the clock, and air transport employs predictive maintenance to

monitor aircraft engines for the purpose of avoiding engine breakdown while

in flight. Even though it has its advantages, predictive maintenance setup

does come with one-time cost in terms of sensor technology, data analytics

equipment, and staff that is trained to analyze information correctly.

Predictive maintenance is a cost-effective solution for companies and

organizations willing to spend on effective asset management procedures as

AI-based solutions keep getting cheaper and the technological environment

keeps evolving.

AI-Powered Predictive Maintenance in the Mining Industry

The mining sector has been a front-runner in adopting AI-driven asset

management technologies, using predictive maintenance technology to

improve performance and lower equipment downtime. A case in point is

Modular Mining Systems' MineCare Maintenance Management System,

computer software that uses AI to monitor and schedule equipment

maintenance to maximize performance. MineCare has the specific focus on

reducing response time for maintenance, with real-time information on

machinery performance.

In contrast to the traditional reactive maintenance, where repairs occur

following a breakdown, MineCare utilizes predictive and reliability-based

maintenance to prevent breakdowns from occurring. In addition to cost

savings, AI maintenance enhances safety within mining. Through early

warning of mechanical failure, organizations avoid dangerous accidents due

to unexpected failure of equipment. The capacity to foresee failures also

improves supply chain effectiveness, with maintenance materials and spares

in place when required. As industries adopt AI into asset management,

technologies such as MineCare demonstrate how predictive maintenance can

change equipment reliability, operational efficiency, and industrial

sustainability overall, (Johnson, R., & Patel, K., 2021).

QR Code-Based Asset Tracking in Educational Institutions

Educational institutions are increasingly integrating QR code-based

asset tracking systems to enhance resource management and improve

operational efficiency. These systems involve assigning unique QR codes to

institutional assets, such as laboratory equipment, computers, furniture, and

books, allowing for seamless tracking, monitoring, and retrieval of asset

information. Traditional asset tracking methods, such as manual logbooks or

barcode scanning, often lead to inefficiencies, errors, and difficulties in

updating records. By contrast, QR code-based tracking provides a faster,

more accurate, and automated solution, ensuring that asset records remain

up to date with minimal human intervention. When scanned using a

smartphone or a dedicated scanner, the QR code instantly retrieves essential

details about the asset, including purchase date, maintenance history, and

status.

One of the major advantages of QR code-based asset tracking is its

ability to reduce asset misplacement and losses by promoting better

accountability. Faculty members, administrative staff, and even students can

scan and verify asset locations, minimizing unauthorized transfers and theft.

Additionally, automated tracking simplifies inventory audits, allowing

institutions to generate reports on asset conditions, depreciation, and

maintenance schedules with ease. Beyond asset tracking, some educational

institutions have started integrating QR codes with maintenance and repair

management systems. By scanning an asset’s QR code, users can quickly

report issues, request repairs, and access troubleshooting guides, reducing

downtime and ensuring that equipment remains functional. As digital

transformation continues to shape educational management, QR code-based

asset tracking systems offer a cost-effective and scalable solution for

institutions looking to enhance asset visibility, accountability, and long-term

sustainability. (Williams, M., & Brown, T., 2020).

Web-Based Maintenance Management Systems

Web-based maintenance management systems (WBMS) are valuable

programs for organizations which wish to automate maintenance, better

track assets, and generally increase efficiency. The program provides an

internet-based system from which maintenance operations can be planned,

tracked, and documented more centrally, decreasing the need for manual

spreadsheet-based or paper-based methods. Through the use of cloud

computing and web access to real-time information, WBMS facilitates better

co-ordination among maintenance teams, managers, and decision-makers

irrespective of the location. Possibly the most significant benefit of web-

based maintenance systems lies in their capacity to automate maintenance

operation scheduling as well as work order administration. Inspect,

preventive maintenance, and repair may be scheduled systematically,

followed, and re-scheduled by one system. This helps with the achievement

of no key maintenance task going unattended, minimizing downtime and

equipment breakdowns. WBMS also supports data-driven decision-making

through storing and learning from maintenance history. By tracking

maintenance and replacement parts, and trended performance,

organizations can determine trends in equipment failures, predict

maintenance requirements, and practice predictive maintenance techniques

for asset life optimization.

The data also enables compliance with regulations by providing true

records of maintenance activities to facilitate auditing and inspection. Yet

another important advantage of WBMS is that it is able to coordinate

communication between maintenance groups. Technicians, managers, and

stakeholders may interact in real time, be provided with real-time

notification of work progress on maintenance, and have access to key asset

data through web-based dashboards. This saves response times, enhances

coordination of the workflow, and increases accountability for maintenance

operations. Web-based maintenance management systems represent an

extension of a larger trend towards digitalization in asset and facility

management. Firms across industries like manufacturing, medical care,

education, and transport are all implementing WBMS to simplify

maintenance practices, reduce operating costs, and offer asset reliability. Aid

is provided despite advantages in the form of proper training to implement

it, secure data handling, and compatibility with current IT infrastructure.

(Lee, J., & Smith, R., 2021).

AI Assistants Join the Factory Floor

Artificial intelligence agents like Microsoft's Factory Operations Agent

bridge the loopholes of conventional processes by identifying failed

conditions automatically, detecting patterns of machine behavior, and

forecasting failure prior to its occurrence. Employing machine learning codes

and big data analytics, these computer programs continuously scan factory

machines, identify deviations, and alert plant managers and technicians

about the same. One of the greatest benefits of AI assistants for

manufacturing is that they can automate operations and cut delays. Instead

of allowing a breakdown to stop production dead in heavy before anything

can be accomplished about it, AI-based monitoring systems can detect small

deviations in equipment operation—temperature fluctuations, unusual

vibration, changes in energy absorption—so maintenance personnel can be

one step [Link] predictive maintenance approach not only reduces

unexpected breakdowns but also extends factory machine lifespan, and it

makes astronomical profits. Additionally, quality control procedures are

made simple by AI-based assistants. AI assistants in manufacturing are not

without disadvantages, though. Cyber-attacks, data privacy concerns, and

upskilling the labor force are among some of the prime drivers for

implementing these technologies. Despite all these issues, AI-powered

assistants are revolutionizing the factory shop floor, ushering in the era of

intelligent, half-autonomous factories. Despite continued development of AI

and machine learning, factory AI assistants will become the norm in the

future factory plan, enhancing working efficiency, reducing downtime, and

facilitating decision-making based on data in industrial configurations

(Johnson, R., & Patel, K., 2021).

Groundbreaking’ Free AI App Can Save Drivers £100s on Repair

Bills

With vehicle maintenance costs always on the increase, artificial

intelligence (AI) is increasingly being used to enable drivers to diagnose

problems and estimate repair costs more efficiently. A new AI-based app has

been created that can scour a massive database of car fault records, repair

records, and manufacturer guides in an effort to counsel car owners in depth

about possible faults and estimated repair fees. The app functions through

enabling users to enter symptoms or error codes for problems in the

performance of their car. Through the use of machine learning algorithms

and predictive analytics, the AI crosschecks this information with repair

histories for millions of cars and identifies typical faults, recommending

probable fixes. Furthermore, the app can offer quotes for repair costs

utilizing historical data, enabling drivers to plan more effectively for repairs.

One of the best features of this AI system is that it can save drivers from

unnecessary repair and overcharge by car mechanics. Most car owners are
not technically advanced and go to repair shops only for diagnosis, leaving

them at the mercy of false estimates or too costly services. With an AI-based

app, drivers can compare repair estimates, check out mechanic suggestions,

and decide intelligently before going for costly repairs. Beyond cost saving,

the app improves car maintenance by enabling the diagnosis of faults at an

early stage. By examining common failure patterns, the AI system can notify

owners of potential faults before they are in advanced states, consequently

precluding the occurrence of breakage and costly breakdown repair. In

others, the AI is even capable of triggering elementary DIY repair for simple

errors, preventing professional intervention in the first instance. But as with

all good things, there are a couple of caveats to AI car diagnosis. Even when

the app can accurately estimate and recommend fixes, it can't substitute the

sweat of skilled technicians in the event of complicated repairs involving

hands-on inspection or expert tools. And since the AI is drawing on historical

data and statistical extrapolation, its suggestions may not always be

precisely the right size for any particular car. However, AI-powered repair

estimate software is a huge leap for the automotive sector, offering

motorists more freedom, greater control over maintenance and care, and

even possible cost savings. As advancements in AI technology continue to

unfold, the future promises even more precise diagnostic capacity and

broader applicability in predictive maintenance systems for personal and

commercial vehicles (Jacob Jaffa, 2024).

 AI-Powered Diagnostic Systems in Maintenance

Artificial intelligence-based diagnostic solutions have come forth as a

game-changing solution to anticipate and forestall machinery breakdown in

advance because the sophistication of machines and industrial equipment

evolves day by day. The systems can explore equipment condition and

enable companies to conduct proactive maintenance processes instead of

reactive repair through implementing machine learning algorithms, real-time

analytics, and predictive modeling into practice. Conventional maintenance

practices like the preventative maintenance (maintenance at a set time) and

repair maintenance (equipment maintenance after an occurrence) will be

prone to leading to unnecessary downtime, expensive repair, and inefficient

consumption of resources.

AI-diagnostic systems constantly check in real time for sensor data,

detect abnormal patterns, and learn precursors to future equipment failure

so that interventions can be made before such failure can occur

catastrophically. One of the largest benefits of AI diagnostics is they can

process enormous amounts of data from IoT-connected sensors monitoring

temperature, vibration, pressure, and energy consumption. AI, using real-

time data, can forecast where and when a failure would happen and reduce

unplanned downtime by half and maximize asset dependability and

operating efficiency. The achievement of AI diagnostics would rely on

employee training, quality data, and integration with existing systems.

Businesses would need to spend on sensor technology, data cloud storage,

and AI model training. (Williams, K., & Brown, J., 2020).

Challenges in Implementing AI in Asset Management

Artificial Intelligence (AI) implementation in asset management holds

the key to transforming maintenance, monitoring, and decision-making

functions through predictive analytics, automation, and smart monitoring.

But as a result of its sheer benefits, applications of AI in asset management

are plagued with critical challenges that need to be faced by organizations in

order to make the process smooth and reap its entire benefits. Availability

and quality of data is perhaps the greatest problem. Asset management

depends on immense amounts of real-time, quality data to feed machine

learning models and make smart predictions.

Most companies possess incomplete, inconsistent, or outdated asset

data, resulting in poor-quality insights and false predictions. To cater to this,

organizations have to spend on data standardization, improved data

gathering mechanisms, and high-quality IoT sensors providing accurate and

formatted datasets for AI analysis. System integration and tech compatibility

are another primary issue. Most industries, especially government

organizations, manufacturing, and infrastructure, continue to have legacy

asset management systems that are not on AI platforms. To shift to AI would

necessitate a high degree of system upgradation, custom API building, and

cloud-based infrastructure, which is expensive and time-consuming.

Organizations need to evaluate their current IT infrastructure and plug in

scalable options that enable seamless integration with AI. Resistance to

change and labor accommodation are also key hindrances to the application

of AI in asset management.

Workers accustomed to operating in the context of legacy asset

management practices can be resistant to AI-driven automation because

they do not want to lose their jobs, experience more complexity, or ineptness

with new technology. Organizations can avoid this resistance by highest

priority giving employee training, providing upskilling AI courses, and putting

emphasis on AI as an add-on and never a substitute for human judgment.

Involvement of workers in the AI adoption process can enhance acceptance

and bring overall implementation success. Secondly, cybersecurity and data

privacy threats are another limitation with asset management using AI

(Patel, S., 2022).

Smart Asset Tracking with IoT and AI

The study explores the convergence space of artificial intelligence (AI)

and Internet of Things (IoT) in asset monitoring platforms and details the
extent to which these technologies are involved in enhancing real-time data

capturing, tracking, and decision-making. Organizations can end the risk of

theft, inefficiency, and misplaced assets with the assistance of IoT-based

sensors, GPS tracking, and AI-driven analysis to offer improved visibility into

the location of the resources, usage trend, and state of condition.

Huge amounts of asset data are scanned by AI software to detect

anomalies, predict the requirement for maintenance, and execute automated

tracking operations that enhance the utilization of resources, minimize cost,

and boost the process efficiency. Apart from that, research leans towards

greater adoption of cloud-based AI platforms with remote access and plug-

and-play compatibility with current enterprise platforms to enhance

scalability and real-time data for high-level asset management. (Williams, K.,

& Brown, J., 2020).

Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks

Two RAG model formulations, RAG-Sequence and RAG-Token, were

suggested by the researchers. RAG-Sequence uses the same set of retrieved

documents for the entire output sequence, while RAG-Token allows different

documents to influence individual tokens. Their findings showed that RAG

significantly outperformed traditional seq2seq models and extractive

systems across various open-domain question answering (QA) tasks. This

foundational work directly supports the theoretical basis for the proposed

repair AI assistant by demonstrating that RAG models can enhance both

factual accuracy and specificity in responses (NeurIPS 2020).

Knowledge Management for Maintenance, Repair and Service of

Manufacturing System

Effective asset management and maintenance in institutional and

industrial settings rely heavily on the proper documentation, analysis, and

utilization of repair data. A study conducted by Wan, Gao, Li, and Evans

(2014) addressed this concern within the context of manufacturing systems.

Their paper, Knowledge Management for Maintenance, Repair and Service of

Manufacturing System, presented at the International Conference on

Manufacturing Research, explored how knowledge derived from maintenance

processes can be better organized to enhance decision-making and

efficiency.

The objective of the study was to evaluate the current state of

maintenance operations in a power generator manufacturing company and

identify inefficiencies in knowledge handling. The researchers conducted a

case study to assess how maintenance records and fault data were stored
and used. Their findings revealed that while vast amounts of maintenance

data existed, the lack of consistency in documentation and absence of an

integrated knowledge management system made it difficult for personnel to

share, reuse, or learn from past experiences. To address this, the study

proposed the development of a knowledge-based system that integrates

equipment condition monitoring, fault prediction, and maintenance histories

to assist technicians and decision-makers. This research is particularly

relevant to the present study, which focuses on developing the I’SMART: ICT

Smart Asset Repair & Tracking with AI Assistant System for ASSCAT. Both

studies highlight the importance of centralized and intelligent systems to

manage repair data, streamline maintenance workflows, and reduce asset

downtime. While Wan et al. (2014) concentrated on manufacturing

environments, the I’SMART system applies these principles within an

educational institution's ICT infrastructure. The integration of QR code

tracking and AI-powered repair assistance in I’SMART builds upon the

concept of digital knowledge management by enhancing data retrieval and

automated diagnostics, even in offline settings. A key similarity lies in the

mutual goal of improving repair efficiency through better data handling.

However, the I’SMART system expands the application by introducing AI

language models and video-based instructional support, thereby offering

real-time guidance for technicians — a component not addressed in the

manufacturing study. Furthermore, while Wan et al. emphasized

organizational-level decision-making, the I’SMART system is designed to

empower both administrative users and on-ground technicians, creating a

more inclusive maintenance ecosystem. The study by Wan et al. (2014)

underscores the critical need for structured maintenance knowledge

management — a foundation upon which the I’SMART system builds and

innovates. By addressing limitations in manual diagnostics and isolated data

systems, the current research contributes to the evolving field of smart

maintenance platforms, particularly in contexts with constrained technical

support and infrastructure (Wan et al. 2014).

Evaluating Retrieval Quality in Retrieval-Augmented Generation

Recent advancements in Retrieval-Augmented Generation (RAG) have

significantly enhanced AI systems that depend on external knowledge bases,

such as intelligent repair systems. A core challenge with RAG is evaluating

the retrieval component, which directly influences the accuracy of the

generated responses. In a 2024 study, Salemi and Zamani introduced eRAG,

a method that evaluates each retrieved document based on its contribution

to a language model’s final output. This method is more in line with real-

world results, raising Kendall's tau correlation to 0.494 and making it much

more efficient by using up to 50 times less GPU memory. This is especially

relevant for systems like I’SMART: ICT Smart Asset Repair & Tracking with AI

Assistant, which plans to use RAG to improve diagnostic suggestions for

technicians. eRAG’s insights support the need for accurate retrieval in

ASSCAT’s LAN-hosted AI system, ensuring it performs well on limited

hardware. However, a gap remains in adapting such techniques for

multilingual environments, such as those requiring Tagalog and Bisaya

languages central to ASSCAT’s context. This study provides a strong

foundation for optimizing retrieval logic in the proposed system (Salemi, A.,

& Zamani, H. 2024).

Informal Learning on YouTube

Patricia G. Lange (2018) highlights YouTube’s role as a dynamic

platform for informal, peer-driven learning through tutorials, walkthroughs,

and interactive feedback. This decentralized model empowers users to

engage in self-paced, interest-based learning that supplements formal

education—particularly valuable in fields like ICT repair. Lange’s findings are

highly relevant to the I’SMART: ICT Smart Asset Repair & Tracking with AI

Assistant System, which integrates YouTube video references to guide

technicians through diagnostics and troubleshooting. By drawing from

YouTube's vast content, the AI Assistant extends support beyond traditional

manuals and procedures. However, Lange also identifies challenges such as

exposure to unverified content, lack of assessment, and potential for

negative public interaction, which suggest the need for safeguards like AI-

generated keyword prompts in systems like I’SMART. Her research

underscores both the potential and limitations of video-based learning,

validating its use as an adaptive training tool in institutional maintenance

environments such as ASSCAT (Lange, P. G. 2018).

Application of KKS Coding and QR Code Technology in

Transmission Asset Management

Zhenhua Wang et al. (2021) present an innovative framework that

applies QR code technology and a structured KKS-based coding system to

enhance transmission asset management. The study highlights the

limitations of traditional asset tracking systems, such as data fragmentation

and inefficiency in lifecycle monitoring. By implementing QR codes in

combination with lifecycle information codes—such as Interactive Location

Codes and Equipment Information Codes—the system enables seamless

tracking, recording, and digitization across multiple platforms. This

integration eliminates data silos and improves real-time access to critical

asset data. The approach aligns with current trends in smart infrastructure

and digital transformation, reinforcing the relevance of QR-based systems in

institutional environments like ASSCAT. For the I’SMART system, this

research supports the technical foundation for applying QR codes in ICT

asset tracking, ensuring greater accuracy, speed, and transparency.

However, while the transmission asset model focuses on electrical systems,

it opens opportunities for further adaptation to ICT contexts, particularly in

educational institutions managing multiple interconnected devices and

systems. The study also identifies gaps in cross-platform compatibility and

user training—areas that I’SMART addresses through LAN-based AI

integration and role-specific functionality. Thus, this work provides both a

technological precedent and a justification for adopting QR-coded lifecycle

tracking in institutional maintenance systems.