Artificial Intelligence Tools in Office Administration: Practices And Integration in Private Institutions

1. Introduction

Artificial Intelligence (AI) has become one of the most significant technological advancements in today's world, transforming the way organizations perform administrative, operational, and decision-making tasks. Across various industries, AI is utilized to automate processes, improve communication, manage information, and enhance workplace productivity. In the field of office administration, AI tools such as virtual assistants, chatbots, automated scheduling systems, and intelligent document management applications have streamlined routine tasks and reduced human errors. As businesses continue to adopt digital solutions to remain competitive, understanding the role of AI in office administration has become increasingly important for both organizations and future office professionals.

Despite the growing adoption of AI technologies worldwide, the implementation of AI tools in office administration within private institutions in the Philippines remains relatively underexplored. Many organizations still rely on traditional administrative methods, which may result in inefficiencies, delays, and increased workloads. Furthermore, some administrative personnel face challenges in utilizing AI-based systems due to limited training, lack of technical knowledge, or resistance to technological change. These issues highlight the need to examine how AI is currently integrated into office administration practices and identify the opportunities and challenges associated with its use. Understanding these factors is essential in promoting more effective and technology-driven administrative systems.

This study aims to examine the practices and integration of Artificial Intelligence tools in office administration within private institutions. Specifically, it seeks to determine how AI enhances administrative performance, improves efficiency, and supports daily office operations while identifying the challenges encountered during its implementation. The findings of this study are expected to provide valuable insights for office professionals, educators, and Bachelor of Science in Office Administration (BSOA) students in adapting to technological advancements in the workplace. Ultimately, this research emphasizes the importance of embracing AI as a tool for innovation and productivity, helping prepare future office administrators for the demands of a modern and technologyoriented work environment.

2. Materials and Methods

This study employed a quantitative research methodology, specifically a descriptive-comparative research design, to evaluate the practices and integration of Artificial Intelligence (AI) tools in office administration across key areas such as communication, data management, scheduling and workflow, document processing, and decision support.

This design was chosen because it enabled the researchers to describe the current use of AI tools and compare the assessments of two groups of respondents, namely managers and administrative staff. The descriptive component was used to determine the extent to which AI tools were adapted and utilized in office administration, while the comparative component was used to identify whether significant differences existed between the two groups of respondents.

The respondents of the study were 87 personnel from selected private institutions, composed of 17 managers and 70 administrative staff. They were selected as participants because they were directly involved in office administration functions and had exposure to the use of AI tools in workplace settings.

A survey questionnaire was used as the main data-gathering instrument. The questionnaire was divided into four parts. The first part gathered the respondents' profile. The second part identified the AI tools used in office administration, particularly in communication, data management, scheduling and workflow, document processing, and decision support. The third part measured the respondents' assessment of AI practices and AI integration. The fourth part covered the challenges encountered in the use and integration of AI tools and served as the basis for the proposed intervention.

Before the actual data collection, the questionnaire was reviewed to ensure clarity and relevance of the items. After securing the needed permission, the researchers administered the questionnaire to the respondents and collected the accomplished forms for analysis.

Statistical Treatment of Data

To support the Descriptive-Comparative quantitative research design of this study, the following statistical tools were employed to evaluate AI practices and integration in office data administration across five functional areas: management, scheduling and communication, workflow, document processing, and decision support.

1. Percentage

This will be used to summarize the demographic profile of respondents and categorical variables.

2. Weighted Mean

This will be used to assess AI practices, problems encountered, and acceptability of the intervention plan.

Formula:

Where:

• WM = Weighted Mean

• W = Weight

• F = Frequency • N = Total Responses

Likert Scale. The Likert scales were used to determine the assessment of the three (2) groups of respondents towards the practices and integration of Artificial Intelligence (AI) tools in Office Administration.

2.1 AI Practices:

3. T-Test

This will be used to determine if there is a significant difference between the weighted means of two independent groups regarding their assessment of AI practices, problems encountered, or the acceptability of the intervention plan.

Formula: 𝑡

Where:

• 𝒕 = t-statistic

• 𝒙̅𝟏 = Mean of the first group

• 𝒙̅𝟐 = Mean of the second group

• 𝑺𝟐𝟏 = Variance of the first group

• 𝑆22 = Variance of the second group

• 𝒏𝟏 = Sample size of the first group

• 𝒏𝟐 = Sample size of the second group

Rule:

The calculated t-value will be compared to the critical t-value at a 0.05 level of significance.

• If the calculated t-value is greater than the critical t-value (or if the p-value is less than 0.05), the null hypothesis is rejected, indicating a significant difference between the groups.

• If the calculated t-value is less than the critical t-value (or if the p-value is greater than 0.05), the null hypothesis is accepted, indicating no

significant difference.

3. Results and Discussion

Table 1 presents the distribution of respondents involved in the study. Out of a total population of 100 employees from selected private institutions, a total of 87 respondents were selected using stratified random sampling. Administrative staff comprised the larger stratum with 70 respondents or 75.00% of the sample, while managers accounted for 17 respondents or 25.00%. The proportional distribution of respondents ensures that both groups are adequately represented in the study, allowing for a fair and meaningful comparative analysis of the two groups' assessments on the practices and integration of AI tools in office administration.

Table 2 shows the profile of respondents in terms of years of experience. Among managers, the majority had 1 to 3 years of experience with 8 or 47.06%, followed by those with 7 to 10 years at 5 or 29.41%, 4 to 6 years at 3 or 17.65%, and more than 10 years at 1 or 5.88%. Among administrative staff, the highest frequency was also in the 1 to 3 years bracket with 37 or 52.86%, followed by 4 to 6 years with 15 or 21.43%, 7 to 10 years with 13 or 18.57%, and more than 10 years with 5 or 7.14%. In total, the majority of respondents across both groups had 1 to 3 years of experience, accounting for 45 or 51.72% of the total sample. These results indicate that most of the respondents are relatively early in their careers, which may influence their familiarity and comfort level with AI tools in the workplace. The table shows that the majority of the respondents have 1–3 years of work experience, comprising 51.72% of the total respondents. This trend is evident among both managers (47.06%) and administrative staff (52.86%). In contrast, only 6.90% of the respondents have more than 10 years of experience. This finding implies that the workforce of the participating organizations is predominantly composed of relatively new employees who are still in the early stages of their professional careers. As a result, their perceptions, practices, and experiences may reflect current workplace trends and contemporary organizational practices. The limited number of highly experienced employees suggests that the organizations may benefit from continuous training, mentoring, and professional development programs to enhance employees' skills and competencies as they gain experience. Furthermore, the findings indicate the importance of providing adequate support and opportunities for career growth to ensure employee retention and long-term organizational effectiveness.

Table 3 presents the educational attainment of the respondents. Among managers, the majority held a Bachelor's

Degree with 10 or 58.82%, followed by those with a Master's Degree at 4 or 23.53%, and a Doctorate Degree at 3 or 17.65%. None of the managers fell under the "Others" category. Among administrative staff, the largest group also held a Bachelor's Degree with 61 or 87.14%, followed by those with a Master's

Degree at 7 or 10.00%, and 1 each or 1.43% for both Doctorate Degree and Others. In total, 71 respondents or 81.61% held a Bachelor's Degree, making it the most common educational attainment across both groups. These findings suggest that most respondents possess a degree-level education, which provides a foundational capacity for understanding and utilizing AI tools in administrative settings.

1. Artificial Intelligence Tools in Office Administration adapted.

AI Tools Adapted in Office Administration:

Research Question 1 inquired about the AI tools adapted in office administration across five categories: communication, data management, scheduling and workflow, document processing, and decision support. Table 4 presents the frequency and percentage of AI tools being used by the respondents.

Table 4 presents the frequency and percentage distribution of AI tools used by the respondents in office administration. The data reveals that Chatbots and AI Assistants were the most commonly used tools, recorded by 56 respondents or 61.54%, with 14 managers and 42 administrative staff using them. This was followed by Writing and Grammar Tools used by 53 respondents or 58.24%, with 9 managers and 44 administrative staff. Email and Communication AI ranked third with 39 respondents or 42.86%, while Document Automation Systems were used by 26 respondents or 28.57% and Scheduling and Workflow Tools by 24 respondents or 26.37%. Data Analytics and Reporting Tools recorded 16 respondents or 17.58%, and AI-based HR, Accounting, or Payroll Systems had the lowest frequency at 14 respondents or 15.38%.

Only one respondent selected "Other," and no respondent indicated non-use of AI tools. Notably, administrative staff consistently showed higher usage frequencies across almost all categories compared to managers, suggesting that frontline staff rely more extensively on AI tools to support their daily administrative tasks.

The data in Table 4 implies that while AI integration is universally present among office administration personnel in the selected private institutions, its current application remains concentrated in superficial, task-oriented domains rather than deep operational workflows. The high utilization of Chatbots (61.54%) and Writing Tools (58.24%) demonstrates that AI is predominantly treated as a digital administrative assistant to streamline textual communication and routine inquiry management. Furthermore, the substantial disparity in usage frequencies between administrative staff and managers suggests that staff members bear the brunt of operational AI adaptation, utilizing these tools to meet daily transactional demands.

Conversely, the low adoption rates for analytical and specialized functional tools—such as Data Analytics (17.58%) and AI-based HR/Payroll Systems (15.38%)—imply a critical gap in strategic AI leveraging. This indicates that administrative leadership has yet to fully exploit AI's potential for institutional foresight, complex data processing, and systemic automation. Because the baseline resistance to AI is nonexistent (0% reporting non-use), the underlying implication is not a rejection of technology, but likely a need for targeted institutional training and the formal procurement of specialized administrative enterprise AI platforms to move beyond basic productivity toward comprehensive digital transformation.

Artificial Intelligence in Office Administration

1.1 Communication

Legend:

Range Scale Verbal Interpretation Symbol

5 4.20-5.00 Highly Adapted HA

4 3.40-4.19 Adapted A

3 2.60-3.39 Moderately Adapted MA

2 1.80-2.59 Least Adapted LA

1. 1.00-1.79 Not Adapted NA

Table 5 presents the level of adaptation of AI tools in office administration as to communication. The overall composite weighted mean of 3.47 is interpreted as Adapted (A), indicating that AI tools are recognized by respondents as effectively supporting communicative functions in the workplace. Among the five indicators, AI tools improve the speed and clarity of internal communication ranked first with a composite weighted mean of 3.72, also interpreted as Adapted. This was followed by AI tools supporting real-time translation or language assistance with 3.66, also Adapted. AI tools facilitate effective communication with external stakeholders ranked third at 3.44, while AI tools improve the overall quality of organizational communication placed fourth at 3.27 and AI tools help reduce miscommunication in written correspondence ranked last at 3.26, both described as Moderately Adapted (MA). Across all indicators, managers consistently rated communication-related AI tools higher (Overall WM = 3.77, Adapted) compared to administrative staff (Overall WM = 3.40, Adapted), suggesting that managers perceive greater communicative value from AI tools, likely due to their broader exposure to communication-intensive responsibilities.

The results suggest that although integrating AI technologies greatly speeds up internal communication's execution and velocity, its ability to control semantic accuracy is still constrained. The main implication of the top-ranked item, "improving the speed and clarity of internal

communication" (WM = 3.72), is that staff members use AI as a high-speed transactional engine to get rid of backlogs in communication and large amounts of data. However, the data also suggests that AI lacks the contextual awareness necessary to decipher subtle organizational dynamics or institutional nuances, as the ability to eliminate misinterpretation in written correspondence ranked lowest (WM = 3.26). As a result, human supervision is still necessary to avoid misunderstandings and protect communication quality even though AI speeds up and increases the volume of communication.

1.2 Data Management

Legend:

Range Scale Verbal Interpretation Symbol

5 4.20-5.00 Highly Adapted HA

4 3.40-4.19 Adapted A

3 2.60-3.39 Moderately Adapted MA

2 1.80-2.59 Least Adapted LA

1 1.00-1.79 Not Adapted NA

Table 6 shows the level of adaptation of AI tools in office administration as to data management. The overall composite weighted mean of 3.39 is interpreted as Moderately Adapted (MA), indicating that while respondents recognize AI's contribution to data-related tasks, full adoption has yet to be achieved. The highest-ranked indicator was AI tools help in organizing and categorizing large volumes of data with a composite weighted mean of 3.66, interpreted as Adapted (A). This was followed by AI tools improve the accuracy of data entry and processing at 3.36 (Moderately Adapted) in second rank, and AI tools enhance the retrieval and storage of institutional data at 3.31 (Moderately Adapted) in third. Both AI tools reduce errors in data management tasks and AI tools support data-driven decisionmaking in the institution tied at 3.30, both Moderately Adapted, ranking fourth. Managers rated this dimension slightly higher (Overall WM = 3.40, Adapted) than administrative staff (Overall WM = 3.38, Moderately Adapted). The findings suggest that while AI tools show clear strength in data organization, respondents remain more reserved about their effectiveness in reducing errors and supporting strategic decision-making. Employees' trust in AI for clerical tasks versus critical data analytics appears to be sharply divided, according to the empirical data in Table 6. The high composite grade (WM = 3.66) for classifying and organizing massive amounts of data suggests that office workers are quick to adopt AI for algorithmic, mechanical sorting and file storage. On the other hand, the joint lowest rankings for data-driven decision support (WM = 3.30) and error reduction (WM = 3.30) suggest a widespread underlying skepticism about the strategic dependability and mathematical accuracy of AI systems. The implication for institutions is that although AI is implicitly trusted as an automated file clerk, it is not yet recognized as a stand-alone analytical engine that can take the place of human scrutiny in institutional governance and verification.

1.3 Scheduling and Workflow

Legend:

Range Scale Verbal Interpretation Symbol

5 4.20-5.00 Highly Adapted HA

4 3.40-4.19 Adapted A

3 2.60-3.39 Moderately Adapted MA

2 1.80-2.59 Least Adapted LA

1 1.00-1.79 Not Adapted NA

Table 7 presents the level of adaptation of AI tools in office administration as to scheduling and workflow. The overall composite weighted mean of 3.42 is interpreted as Adapted (A). The highest-ranked indicator was AI tools help in scheduling appointments and meetings efficiently at a composite weighted mean of 3.73, followed by AI tools automate repetitive workflow tasks at 3.43, both Adapted. AI tools streamline approval processes and document routing ranked third at 3.38 (Moderately Adapted), AI tools assist in monitoring and tracking task progress ranked fourth at 3.36 (Moderately Adapted), and AI tools reduce delays in workflow processes ranked last at 3.19 (Moderately Adapted). A noteworthy difference exists between the two groups: managers rated scheduling and workflow tools as Moderately Adapted (Overall WM = 3.33) while administrative staff rated the same dimension as Adapted (Overall WM = 3.44), suggesting that frontline staff derive more practical benefit from AI tools in their scheduling and workflow responsibilities compared to managers.

Table 7 reveals a critical paradox regarding the optimization of office workflows through automation. The high adaptation rate for scheduling appointments and meetings efficiently (WM = 3.73) implies that AI tools are highly effective at eliminating routine logistical friction and automating calendar management. However, because the capacity to reduce delays in workflow processes ranked the lowest (WM = 3.19), the underlying implication is that institutional bottlenecks are not caused by simple scheduling conflicts, but rather by deeper structural realities, such as human dependencies, rigid approval layers, or legacy institutional bureaucracies. Thus, the data implies that while technology can optimize individual coordination, it cannot inherently correct deep-seated systemic delays within the organization.

1.4 Document Processing

Legend:

Range Scale Verbal Interpretation Symbol

5 4.20-5.00 Highly Adapted HA

4 3.40-4.19 Adapted A

3 2.60-3.39 Moderately Adapted MA

2 1.80-2.59 Least Adapted LA

1. 1.00-1.79 Not Adapted NA

Table 8 presents the level of adaptation of AI tools in office administration as to document processing, which obtained the highest overall rating among all five AI tool dimensions. The overall composite weighted mean of 3.63 is interpreted as Adapted (A). The top-ranked indicator was AI tools help in reviewing and proofreading institutional documents with a composite weighted mean of 4.01, followed by AI tools facilitate accurate and fast document conversion at 3.62, both Adapted. AI tools assist in drafting and formatting official documents ranked third at 3.59 (Adapted), AI tools improve records management and archiving ranked fourth at 3.56 (Adapted), and AI tools reduce the time spent on document-related tasks ranked last at 3.36 (Moderately Adapted). Both managers (Overall WM = 3.63) and administrative staff (Overall WM = 3.62) gave nearly identical ratings across all indicators, reflecting strong agreement that AI tools are meaningfully and consistently adapted for document processing tasks in private institutions. The results outlined in Table 8 imply that AI tools are integrated primarily as auxiliary editorial assistants rather than end-to-end document generators. The top-ranking position of AI for reviewing and proofreading institutional documents (WM = 4.01) implies that the workforce values the technology for lexical correction, spelling verification, and formatting precision. Nevertheless, the lower rating for reducing time spent on document-related tasks (WM = 3.36) implies that the overall efficiency gains remain capped. This suggests that because administrative documentation demands absolute institutional compliance and accuracy, personnel must still invest substantial time in human auditing and fact-checking, preventing total workflow automation.

1.5 Decision Support

Legend:

Range Scale Verbal Interpretation Symbol

5 4.20-5.00 Highly Adapted HA

4 3.40-4.19 Adapted A

3 2.60-3.39 Moderately Adapted MA

2 1.80-2.59 Least Adapted LA

1 1.00-1.79 Not Adapted NA

Table 9 presents the level of adaptation of AI tools in office administration as to decision support, which obtained the lowest overall rating among the five AI tool dimensions. The overall composite weighted mean of 3.36 is interpreted as Moderately Adapted (MA). The highest-ranked indicator was AI tools assist in forecasting and strategic planning with a composite weighted mean of 3.52 (Adapted), followed by AI tools provide data analysis that supports management decisions at 3.45 (Adapted). AI tools enhance evidence-based decision-making ranked third at 3.32 (Moderately Adapted), AI tools help identify patterns or trends useful for planning ranked fourth at 3.29 (Moderately Adapted), and AI tools improve the reliability of institutional reports ranked last at 3.22 (Moderately Adapted). Both managers (Overall WM = 3.37) and staff (Overall WM = 3.36) gave similarly modest ratings, both Moderately Adapted, indicating that AI tools for higher-order decision-making functions are not yet fully leveraged or trusted in private institutions.

The data from Table 9 implies that office personnel view AI as a valuable tool for speculative, forward-looking exploration, but reject it for precise retrospective reporting. The high rating for assisting in forecasting and strategic planning (WM = 3.52) implies that administrators enjoy leveraging AI for broad trend predictions, scenario modeling, and strategic brainstorming. Yet, the lowest ranking for improving the reliability of institutional reports (WM = 3.22) implies that when formal organizational accountability is on the line, personnel do not trust AI to handle raw, definitive data. The underlying implication is that AI is welcomed for its imaginative, predictive features but heavily guarded against when it comes to static, empirical record-keeping.

Assessment of Artificial Intelligence Tools in Office

Administration by Managers and Staff in Private Institutions.

2.1 AI Practices

2.1.1 Level of Utilization

Legend:

Range Scale Verbal Interpretation Symbol

5 4.20-5.00 Highly Utilized HU

4 3.40-4.19 Utilized U

3 2.60-3.39 Moderately Utilized MU

2 1.80-2.59 Least Utilized LU

1 1.00-1.79 Not Utilized NU

Table 10 presents the assessment of AI use in terms of level of utilization. The overall composite weighted mean of 3.43 is interpreted as Utilized (U), reflecting that AI tools have been integrated into the regular operations of the respondents' institutions. The highest-ranked indicator was AI tools are regularly used in daily office operations with a composite weighted mean of 3.74 (Utilized), followed by A variety of AI tools are utilized across different tasks at 3.59 (Utilized). AI tools are used by most personnel in the institution ranked third at 3.49 (Utilized), The institution encourages the use of AI tools in administration ranked fourth at 3.42 (Utilized), and the frequency of AI tool use has increased over time ranked last at 3.40 (Utilized). Managers rated overall utilization higher at 3.67 compared to administrative staff at 3.48, though both groups fall within the Utilized range. These findings suggest that AI tools have become an established component of daily administrative operations, with management showing a stronger sense of institutional encouragement and growing utilization compared to frontline staff.

The metrics in Table 10 imply that while AI tools have successfully penetrated the daily routine of office administration, their adoption curve has reached an institutional plateau. The high score for regular use in daily office operations (WM = 3.74) implies that AI has shifted from an optional novelty to a permanent staple of the workplace. However, because the frequency of AI tool use increasing over time scored the lowest (WM = 3.40), the data implies that employee usage habits have stagnated. Without the procurement of more advanced enterprise software or the introduction of specialized training pathways, personnel will likely continue using the same basic tools for the same rudimentary tasks without naturally progressing toward higher-order integration.

2.1.2 User Competence

Legend:

Range Scale Verbal Interpretation Symbol

5 4.20-5.00 Highly Utilized HU

4 3.40-4.19 Utilized U

3 2.60-3.39 Moderately Utilized MU

2 1.80-2.59 Least Utilized LU

1 1.00-1.79 Not Utilized NU

Table 11 shows the assessment of AI use as to user competence. The overall composite weighted mean of 3.30 is interpreted as Moderately Utilized (MU), indicating that personnel possess only a moderate level of confidence and skill in using AI tools. The highest-ranked indicator was Personnel can navigate and operate AI tools with confidence at 3.49 (Utilized), followed by Personnel apply AI tools appropriately to assigned tasks at 3.35 (Moderately Utilized). Personnel continuously update their skills in using AI tools ranked third at 3.27 (Moderately Utilized), Personnel are knowledgeable about the AI tools they use ranked fourth at 3.21 (Moderately Utilized), and Personnel can troubleshoot minor issues with AI tools ranked last at 3.18 (Moderately Utilized). A notable gap was observed between managers (Overall WM = 3.62, Utilized) and administrative staff (Overall WM = 3.24, Moderately Utilized), suggesting that managers feel significantly more competent in using AI tools. This gap highlights the need for targeted and structured training programs particularly for administrative staff to bridge the competence divide.

Table 11 uncovers an operational vulnerability within the institution’s workforce regarding their actual technological literacy. The high ranking for navigating and operating AI tools with confidence (WM = 3.49) implies that personnel possess high superficial fluency—meaning they can easily operate user interfaces and generate text-based outputs. However, because the capability to troubleshoot minor issues with AI tools fell to a critically low verbal interpretation of Moderately Utilized (WM = 3.18), the underlying implication is a "black-box" dependency. The workforce knows how to prompt AI, but lacks the engineering logic to diagnose hallucinations, errors, or system failures, creating a fragile administrative infrastructure that is highly prone to stalling when technical disruptions occur.

2.1.3 Efficiency Improvement

Legend:

Range Scale Verbal Interpretation Symbol

5 4.20-5.00 Highly Utilized HU

4 3.40-4.19 Utilized U

3 2.60-3.39 Moderately Utilized MU

2 1.80-2.59 Least Utilized LU

1 1.00-1.79 Not Utilized NU

Table 12 presents the assessment of AI use as to efficiency improvement, which achieved the highest overall rating among all AI Practices sub-dimensions. The overall composite weighted mean of 3.61 is interpreted as Utilized (U). The highest-ranked indicator was AI tools have significantly reduced the time needed to complete tasks with a composite weighted mean of 3.71, followed by AI tools have increased the overall productivity of the office at 3.64, both Utilized. AI tools have improved the quality of administrative outputs ranked third at 3.62 (Utilized), AI tools have minimized manual errors in administrative processes ranked fourth at 3.57 (Utilized), and AI tools have streamlined administrative processes effectively ranked last at 3.50 (Utilized). All indicators for both managers (Overall WM = 3.78) and administrative staff (Overall WM = 3.57) consistently fell within the Utilized range, indicating strong and uniform agreement across both groups that AI tools have produced measurable gains in administrative efficiency, productivity, and output quality.

The findings in Table 12 imply that the current deployment of AI yields massive individual micro-efficiencies but fails to achieve macro-systemic transformation. The top ranking for significantly reducing the time needed to complete tasks (WM = 3.71) implies that AI is highly disruptive at the individual level, saving personnel valuable time on standalone tasks like drafting, scheduling, or summarization. Conversely, the lower ranking for streamlining administrative processes effectively (WM = 3.50) implies that these localized time-savings do not aggregate into smoother cross-departmental operations. The institutional implication is that AI tools are currently operating as isolated productivity enhancers rather than an integrated architectural solution capable of modernizing overall process workflows.

2.1.4 Data Security Practices

Legend:

Range Scale Verbal Interpretation Symbol

5 4.20-5.00 Highly Utilized HU

4 3.40-4.19 Utilized U

3 2.60-3.39 Moderately Utilized MU

2 1.80-2.59 Least Utilized LU

1 1.00-1.79 Not Utilized NU

Table 13 shows the assessment of AI use as to data security practices, which obtained the lowest overall rating among all AI Practices sub-dimensions. The overall composite weighted mean of 3.00 is interpreted as Moderately Utilized (MU). The highestranked indicator was the institution has clear guidelines on data privacy when using AI at 3.15 (Moderately Utilized), followed by Personnel are aware of data privacy risks associated with AI tools at 3.08 (Moderately Utilized). The institution regularly monitors data security in AI tool use ranked third at 2.99 (Moderately Utilized), AI tools used comply with data protection regulations ranked fourth at 2.94 (Moderately Utilized), and Personnel practice responsible handling of data through AI tools ranked last at 2.84 (Moderately Utilized). Both managers (Overall WM = 3.05) and administrative staff (Overall WM = 2.99) gave similarly modest ratings. These results reveal significant gaps in data privacy awareness, regulatory compliance, and responsible data handling practices, pointing to an urgent need for institutional policies and personnel training specifically addressing AI-related data security.

The alarming trends in Table 13 imply a profound regulatory and behavioral crisis within the target institutions. While the data implies that institutions have made baseline efforts to establish clear guidelines on data privacy (WM = 3.15), the lowest-ranked metric shows that personnel practicing responsible data handling through AI tools scored a deeply concerning 2.84. This massive statistical disconnect implies a severe policy enforcement void. In their pursuit of speed and convenience, administrative employees are routinely bypassing security protocols—likely pasting sensitive student, financial, or proprietary institutional data into open-source AI engines without any regard for legal or regulatory data compliance.

2.1.5 Ethical and Responsible Use

Legend:

Range Scale Verbal Interpretation Symbol

5 4.20-5.00 Highly Utilized HU

4 3.40-4.19 Utilized U

3 2.60-3.39 Moderately Utilized MU

2 1.80-2.59 Least Utilized LU

1 1.00-1.79 Not Utilized NU

Table 14 presents the assessment of AI use as to ethical and responsible use. The overall composite weighted mean of 3.25 is interpreted as Moderately Utilized (MU), indicating that ethical standards and institutional mechanisms for responsible AI use have not yet been fully developed or enforced. The highest-ranked indicator was the institution promotes ethical use of AI in the workplace at 3.36 (Moderately Utilized), while Personnel use AI tools responsibly and ethically and Personnel avoid misuse or over-reliance on AI tools both ranked second with a composite weighted mean of 3.26 (Moderately Utilized). Personnel respect intellectual property and confidentiality using AI ranked fourth at 3.22 (Moderately Utilized), and the institution addresses ethical issues arising from AI use ranked last at 3.16 (Moderately Utilized). Managers (Overall WM = 3.33) rated this dimension slightly higher than administrative staff (Overall WM = 3.19), though both groups fall under Moderately Utilized. These findings suggest that while basic awareness of ethical AI use exists in private institutions, concrete institutional actions to address and govern ethical concerns remain inadequate.

The data in Table 14 implies that the institutions maintain a passive, performative stance regarding technological ethics. The first-ranked metric shows that the institution actively promotes the ethical use of AI in the workplace (WM = 3.36), which implies a healthy, high-level corporate narrative. However, because the institution addressing ethical issues arising from AI use scored the lowest (WM = 3.16), the underlying implication is a complete lack of operational accountability. The institutions are content with discussing ethics in theory but remain structurally unequipped, un-policed, and legally unprepared to actively discipline, rectify, or manage real-world ethical breaches, intellectual property violations, or algorithmic biases when they manifest on the ground.

2.2 AI Integration

2.2.1 Institutional Readiness

Legend:

Range Scale Verbal Interpretation Symbol

5 4.20-5.00 Highly Integrated HI

4 3.40-4.19 Integrated I

3 2.60-3.39 Moderately Integrated MI

2 1.80-2.59 Least Integrated LI

1 1.00-1.79 Not Integrated NI

Table 15 presents the assessment of AI integration as to institutional readiness. The overall composite weighted mean of 3.12 is interpreted as Moderately Integrated (MI), indicating that private institutions are in the early stages of preparing for systematic AI adoption. The highest-ranked indicator was the institution has assessed its readiness for AI implementation at 3.32 (Moderately Integrated), followed by the institution has the infrastructure to support AI tool use at 3.30 (Moderately Integrated). The institution adapts quickly to new AI developments ranked third at 3.17 (Moderately Integrated), Leadership actively supports the adoption of AI in administration ranked fourth at 3.03 (Moderately Integrated), and the institution has a strategic plan for AI integration ranked last at 2.74 (Moderately Integrated). Managers (Overall WM = 3.37) rated institutional readiness higher than administrative staff (Overall WM = 3.14), though both remain within the Moderately Integrated range. These results suggest that while some readiness indicators are in place, formal strategic planning for AI integration is still underdeveloped, reflecting the broader challenge of AI adoption in Philippine private institutions.

The data presented in Table 15 implies that while the selected private institutions are actively conducting preliminary assessments for AI implementation, they drastically lack the concrete strategic architecture required to guide that integration. The top-ranking position of the institution assessing its readiness for AI implementation (WM = 3.32) implies that administration is highly conscious of the impending technological shift and is taking the first steps to measure its organizational posture. However, because the development of a strategic plan for AI integration ranked lowest (WM = 2.74), the underlying implication is that these institutions are operating reactively rather than proactively. Without a definitive roadmap, the current deployment remains fractured and speculative, leaving departments to adapt in isolation without a centralized institutional vision.

2.2.2 Training and Development

Legend:

Range Scale Verbal Interpretation Symbol

5 4.20-5.00 Highly Integrated HI

4 3.40-4.19 Integrated I

3 2.60-3.39 Moderately Integrated MI

2 1.80-2.59 Least Integrated LI

1 1.00-1.79 Not Integrated NI

Table 16 presents the assessment of AI integration as to training and development. The overall composite weighted mean of 2.97 is interpreted as Moderately Integrated (MI), reflecting that training programs for AI tool use in private institutions are not yet sufficiently developed or consistently implemented. The highest-ranked indicator was Training improves the competence of personnel in using AI tools at 3.06 (Moderately Integrated), followed by Training programs are regularly updated to include new AI tools at 2.99 (Moderately Integrated). The institution supports continuous learning in AI technology ranked third at 2.96 (Moderately Integrated), personnel are given sufficient time for AI training ranked fourth at 2.91 (Moderately Integrated), and the institution provides formal training on AI tool use ranked last at 2.94 (Moderately Integrated). Both managers (Overall WM = 3.11) and administrative staff (Overall WM = 2.94) rated training and development as Moderately Integrated, indicating consensus that institutions have not yet sufficiently invested in formalizing and sustaining AI training programs for their personnel.

Table 16 highlights a significant bottleneck regarding systemic capacity-building and skill acquisition within the workforce. The highest-ranked metric—training programs improving the competence of personnel in using AI tools (WM = 3.06)—implies that when training does occur, the staff find pedagogical value in it. Nevertheless, the low composite scores for providing formal training on AI tool use (WM = 2.94) and providing personnel with sufficient time for AI training (WM = 2.91) imply that these educational opportunities are severely scarce and structurally unsupported. The overarching implication is that institutions are expecting personnel to achieve technological fluency through self-directed, informal methods during regular work hours, rather than investing institutional resources into formalized, continuous learning pathways.

2.2.3 Policy and Governance

Legend:

Range Scale Verbal Interpretation Symbol

5 4.20-5.00 Highly Integrated HI

4 3.40-4.19 Integrated I

3 2.60-3.39 Moderately Integrated MI

2 1.80-2.59 Least Integrated LI

1 1.00-1.79 Not Integrated NI

Table 17 shows the assessment of AI integration as to policy and governance. The overall composite weighted mean of 2.89 is interpreted as Moderately Integrated (MI), suggesting that policies and governance frameworks for AI use in private institutions are present but insufficient and inconsistently applied. The highest-ranked indicator was Policies ensure responsible and ethical use of AI tools at 2.96 (Moderately Integrated), followed by Personnel are informed of and comply with AI use policies at 2.92 (Moderately Integrated). The institution has policies governing AI tool use and AI-related policies are regularly reviewed and updated both tied at 2.89 (Moderately Integrated), ranking third. There are clear guidelines for acceptable AI tool use ranked last at 2.81 (Moderately Integrated). Managers (Overall WM = 3.09) rated policy and governance slightly higher than staff (Overall WM = 2.85), though both remain Moderately Integrated. These results reveal that clear, comprehensive, and consistently communicated AI governance structures are still lacking in most private institutions surveyed.

The metrics in Table 17 reveal a fragile legal and regulatory environment regarding automated workflows. The firstranked item—policies ensuring the responsible and ethical use of AI tools (WM = 2.96)—indicates that compliance frameworks are slowly being initiated at a baseline conceptual level. However, because the criteria regarding clear guidelines for acceptable AI tool use scored the lowest (WM = 2.81), the underlying implication is a critical lack of operational clarity. Personnel are being told to act ethically in a broad sense, but they have not been provided with granular, practical boundaries distinguishing permissible from non-permissible AI actions. This ambiguity implies that the institution is highly exposed to compliance, data privacy, and intellectual property liabilities due to a policy framework that is too vague to enforce.

2.2.4 Resource Allocation

Legend:

Range Scale Verbal Interpretation Symbol

5 4.20-5.00 Highly Integrated HI

4 3.40-4.19 Integrated I

3 2.60-3.39 Moderately Integrated MI

2 1.80-2.59 Least Integrated LI

1 1.00-1.79 Not Integrated NI

Table 18 presents the assessment of AI integration as to resource allocation, which recorded the lowest overall rating among all Integration sub-dimensions. The overall composite weighted mean of 2.85 is interpreted as Moderately Integrated (MI), indicating that the provision of financial, technical, and human resources for AI adoption remains insufficient across the institutions surveyed. The highest-ranked indicator was the institution provides the necessary hardware and software at 2.89 (Moderately Integrated), followed by the institution prioritizes AI investment in its planning at 2.86 (Moderately Integrated). Resources for AI tool maintenance are available and personnel are given time and tools needed for AI use both tied at 2.84 (Moderately Integrated), while the institution allocates adequate budget for AI tools ranked last at 2.81 (Moderately Integrated). A notable gap was observed between managers (Overall WM = 3.22) and administrative staff (Overall WM = 2.76), with staff perceiving resource availability far more critically. This disparity suggests that the actual experience of resource constraints is more pronounced at the operational level, pointing to a significant disconnect between management's perception and staff's day-to-day reality.

The findings in Table 18 imply that administrative leadership is willing to procure physical technology but hesitant to back it up with long-term financial commitments or infrastructure support. The top ranking for providing the necessary hardware and software (WM = 2.89) implies that institutions are willing to absorb the upfront capital expenditures required to give employees basic access to AI assets. Conversely, the dead-last ranking for allocating an adequate budget for AI tools (WM = 2.81) implies a lack of fiscal sustainability. The ultimate implication is that funding for AI is being treated as a temporary, ad-hoc expense rather than a permanent, strategically integrated budgetary line item, which severely limits the institution's ability to scale up to enterprise-level administrative systems.

2.2.5 Operational Efficiency

Legend:

Range Scale Verbal Interpretation Symbol

5 4.20-5.00 Highly Integrated HI

4 3.40-4.19 Integrated I

3 2.60-3.39 Moderately Integrated MI

2 1.80-2.59 Least Integrated LI

1 1.00-1.79 Not Integrated NI

Table 19 presents the assessment of AI integration as to operational efficiency. The overall composite weighted mean of 3.27 is interpreted as Moderately Integrated (MI). The topranked indicators where AI tools have reduced operational costs in administration and overall, AI tools have improved the operational efficiency of the institution, both tied at 3.32 (Moderately Integrated). AI tools have improved service delivery to stakeholders ranked third at 3.28 (Moderately Integrated), AI tools have contributed to faster office transactions ranked fourth at 3.27 (Moderately Integrated), and AI tools have enhanced the coordination among departments ranked last at 3.16 (Moderately Integrated). The most striking finding in this table is the wide gap between the two groups: managers rated operational efficiency as Integrated (Overall WM = 3.93), while administrative staff rated it as only Moderately Integrated (Overall WM = 3.11). This significant disparity suggests that while managers perceive AI tools as already contributing meaningfully to institutional efficiency, the benefits have not yet been uniformly experienced by frontline administrative staff, pointing to an uneven distribution of AI's operational impact across organizational levels.

Table 19 exposes a profound perceptual schism between administrative leadership and operational execution staff regarding the ultimate efficacy of AI. The shared top rankings for reducing operational costs (WM = 3.32) and improving overall institutional efficiency (WM = 3.32) imply that at a macro level, the technology is successfully yielding a positive return on investment. Crucially, however, the table demonstrates a massive divergence between managers (WM = 3.93, interpreting AI as fully Integrated) and administrative staff (WM = 3.11, interpreting AI as only Moderately Integrated). This striking statistical gap implies that while managers are enjoying high-level administrative overhauls, cost reductions, and polished reports, the frontline administrative staff are encountering the friction, systemic delays, and data-entry bugs of unrefined software integration, meaning the true operational burden of AI adoption is being disproportionately felt by the lower tier of the organizational hierarchy.

3. Significant difference in the assessment of the two groups of respondents.

Research Question 3 examined whether there is a significant difference in the assessment of the two groups of respondents— managers and administrative staff—across all variables. Independent samples t-tests were conducted at the 0.05 level of significance.

Note. *Significant at p < .05. H₀: There is no significant difference in the assessment of the two groups.

The independent samples t-tests conducted to determine whether significant differences exist between managers and administrative staff in their assessments across all fifteen variables, at the 0.05 level of significance. Out of the fifteen variables tested, four yielded significant differences, leading to the rejection of the null hypothesis.

Communication registered a t-value of 2.496 with a p-value of 0.0133, where managers (WM = 3.77) rated it significantly higher than staff (WM = 3.40). User Competence showed a t-value of 2.296 with a p-value of 0.0226, with managers (WM = 3.62) perceiving their competence more favorably than staff (WM = 3.24). Resource Allocation yielded a t-value of 2.244 with a pvalue of 0.0258, where managers (WM = 3.22) assessed resource provision more positively than staff (WM = 2.76). Operational Efficiency produced the most significant difference with a tvalue of 4.169 and a p-value of 0.0000, with managers (WM = 3.93) rating it markedly higher than staff (WM = 3.11). For the remaining eleven variables — Data Management, Scheduling and Workflow, Document Processing, Decision Support, Level of Utilization, Efficiency Improvement, Data Security Practices, Ethical and Responsible Use, Institutional Readiness, Training and Development, and Policy and Governance — no statistically significant differences were found (p > .05), leading to the acceptance of the null hypothesis.

These results indicate that managers and administrative staff share broadly similar assessments across most dimensions of AI tool use, with meaningful divergence occurring primarily in areas where differences in authority, access to resources, and strategic visibility are most pronounced.

4. Conclusion

Based on the findings of this study, the following conclusions are drawn:

1. AI tools are broadly embedded in office administration in private institutions. The universal adoption across all respondent groups confirms that AI technology has become an operational standard in private institutional settings, regardless of position or years of experience. However, broad adoption has not yet translated into deep, uniform proficiency across all organizational levels.

2. Document processing is the most mature functional application of AI, while decision support remains the least developed. Private institutions have effectively leveraged AI for drafting, proofreading, document conversion, and records management, yet the wider potential of AI for strategic planning, forecasting, and evidence-based decision-making has not been fully realized.

3. Administrative staff’s user competency is moderate despite broad AI adoption. The notable gap between managers’ user competence (WM = 3.62) and administrative staff’s (WM = 3.24) indicates that AI adoption has not produced equitable skill development across organizational levels, pointing to the need for targeted, structured, and tiered AI literacy programs.

4. Data security practices and ethical use of AI represent the most critical institutional vulnerabilities. Both dimensions fall below the Utilized threshold, signaling that private institutions have not established adequate data privacy safeguards, compliance mechanisms, or ethical governance structures commensurate with their current level of AI use.

5. Resource allocation is the most significant structural barrier to equitable and sustainable AI integration. As the lowest-rated variable in the entire study (WM = 2.85), the inability to consistently fund, supply, and fairly distribute AI tools and infrastructure represents the foremost institutional constraint to advancing AI integration in private settings.

6. Institutional readiness for policy, governance, and training lags considerably behind the pace of AI adoption. The absence of formally communicated AI policies and infrequently updated training programs has created a vulnerable environment where AI is widely used without sufficient safeguards, guidance, or competency development structures.

7. Managers and administrative staff hold broadly similar assessments of AI in most dimensions, but meaningful divergence exists in four key areas. Statistically significant differences in Communication, User Competence, Resource Allocation, and Operational Efficiency reflect systematic disparities in how AI benefits are experienced across organizational levels, attributable to differences in authority, resource access, and strategic visibility.

8. The widest and most significant perception gap between managers and staff exists in Operational Efficiency. Out of fifteen variables tested, only four yielded significant differences, and therefore the null hypothesis is accepted for eleven variables. However, for Communication, User Competence, Resource Allocation, and Operational Efficiency, the null hypothesis is rejected. With a t-value of 4.169 and a p-value below .001, the operational efficiency gap signifies that the gains attributed to AI by management are not yet equitably experienced at the frontline administrative level, indicating that AI integration has not yet achieved institution-wide impact.

5. Recommendations

Based on the findings and conclusions, the following recommendations are offered:

1. The Bachelor of Science in the Office Administration Department and the Area Chair and Dean of CBPA may recommend for implementing AI Literacy training programs with particular focus on administrative staff. The training of basic tool navigation through practical application, advanced troubleshooting and responsible use and should be update regularly to keep pace with evolving AI technologies. This recommendation directly responds to the moderate user competency rating (Overall WM = 3.30) and the visible competency gap between managers (WM = 3.62) and administrative staff (WM = 3.24), which signals an urgent need for targeted, institution-level investment in AI workforce development.

2. The CBPA Colleges of EARIST together with BSOA, Professor & Students should formulate and adopt a formal AI Ethics and Responsible Use Policy that provides clear directions for the ethical, fair, and transparent use of AI tools across all administrative roles. The policy must address prevention of misuse, risks of over-reliance on AI systems, and protection of confidentiality and intellectual property. Adherence to these ethical expectations should be embedded in performance review standards, making responsible AI use a recognized and measurable professional competency rather than a vague expectation. This recommendation responds to the Ethical and Responsible Use subdimension being rated Moderately Utilized (Overall WM = 3.25), indicating that current ethical safeguards remain insufficient relative to the breadth of AI adoption.

3. The Office of the Administration, the IT Department, and Department Heads should establish a cross-functional AI Governance Committee composed of representatives from management, human resources, information technology, and administrative staff, with a formally defined Terms of Reference. This body should build and maintain a comprehensive AI Governance Framework that sets documented guidelines for acceptable AI use, policy review cycles, and systematic communication of policies to all organizational levels. This is grounded in the finding that Policy and Governance was rated Moderately Integrated (Overall WM = 2.89), the second lowest rating in the study, underscoring the need for governance structures that are operational and regularly reviewed.

4. The Budget and Finance Office and Senior Management should prioritize a dedicated AI budget line item in annual institutional planning, covering acquisition, maintenance, and upgrading of AI-related hardware, software, and training resources. Extra care must be taken to ensure administrative staff receive equitable access to AI tools, devices, and software, given that their resource ratings (WM = 2.76) were significantly lower than those of managers (WM = 3.22; t = 2.244, p = .026). Resource Allocation was the lowest-rated variable in the entire study (Overall WM = 2.85), identifying it as the most urgent structural obstacle to sustainable and equitable AI integration in private institutions.

5. The Office of the Administration and Department Heads should conduct a formal AI Readiness Assessment to systematically identify infrastructure gaps, skill deficiencies, and governance weaknesses before expanding AI usage. Assessment results should form the basis of a Strategic AI Integration Plan with clearly defined short-, medium-, and long-term benchmarks, realistic timelines, and measurable key performance indicators. Leadership commitment must be formalized through institutional directives, appointment of departmental AI Champions, and adequate resourcing for sustained implementation. This recommendation is grounded in Institutional Readiness being rated Moderately Integrated (Overall WM = 3.12), indicating that most private institutions have not yet established the foundational conditions enabling systematic and scalable AI integration.

6. The Human Resources Department and Unit Supervisors should establish regular cross-level communication channels — including team briefings, joint feedback sessions, and participatory AI decision-making forums — to bridge the significant Communication and Operational Efficiency gaps identified between managers and administrative staff. Transparent AI impact monitoring mechanisms should be instituted to ensure that operational benefits of AI tools are experienced, communicated, and acknowledged equitably at all organizational levels. This recommendation responds directly to the most significant finding of the study: the Operational Efficiency gap (t = 4.169, p < .001) between managers (WM = 3.93) and administrative staff (WM = 3.11), which indicates that the gains credited to AI by management are not yet experienced equitably at the frontline level.

7. Future researchers are encouraged to replicate and extend this study using larger and more geographically diverse samples from private institutions across the Philippines to improve the generalizability of the findings. Longitudinal, qualitative, and mixed-methods studies are especially recommended to better understand the lived experiences of administrative staff in AI-integrated environments, and to explore the organizational, cultural, and regulatory factors that enable or constrain effective and equitable AI adoption in office administration.