AI as the New Workplace Assistant—Promise, Limits, and Practical Realities

As organizations increasingly experiment with AI tools to answer employee questions, interpret policy documents, or guide internal procedures, it is tempting to see AI as a kind of universal workplace assistant: always available, endlessly patient, and capable of reducing administrative burden. But this week’s readings reminded me that using AI as a catch-all solution requires a far more grounded approach. Nemorin et al. (2023) highlight how AI is often surrounded by inflated promises, and this made me more cautious about positioning an AI assistant as a complete replacement for human judgment. Just the way that these AI bots take things so literally makes me think of the old Amelia Bedelia books!

If an internal AI tool provides incorrect information about procedures or compliance requirements, the consequences can be far more serious than a simple technology glitch. The authors also note that AI hype often conceals deeper issues related to privacy and surveillance, which pushed me to consider how internal search tools might inadvertently track or profile employees based on the questions they ask. This may inherently make the AI assistant bias as it collects information on the employee population and the types of questions they may be asking. Can you imagine an AI chatbot telling a high-performing employee they should just quit?

Similarly, Sofia et al. (2023) argue that AI is reshaping workforce expectations by creating constant demands for reskilling. This made me rethink the assumption that AI assistants automatically reduce workload; instead, employees need training to use these tools effectively and to understand their limitations, especially when the AI is interpreting policies or guiding procedural decisions. Their discussion on employee trust also resonated with me. Deploying AI internally is not just a technical decision, it is a cultural one. Employees are far more likely to rely on an AI assistant when the organization communicates clearly about how it works, what data it uses, and where human oversight still matters.

Touretzky et al. (2019) reinforce this human-centered approach by emphasizing the importance of AI literacy. Their argument that foundational AI understanding is essential made me realize that workplace AI assistants should not merely give answers but should support the development of employee judgment. When people understand how AI models process information, they become more discerning and less likely to accept outputs uncritically. The authors’ focus on ethical reasoning also shaped my thinking about internal AI tools. If an AI assistant is delivering guidance on workplace policies, the organization has a responsibility to ensure the system does so ethically, accurately, and in ways that support, not undermine, employee autonomy. Sometimes, this may expose initiatives in the organization such as a RIF (reduction in Force) inadvertently since AI tools don’t understand how to execute or properly incorporate the concept of timing in employee matters.

Overall, these readings helped me see AI assistants not as a replacement for employee work, but as a carefully governed support tool that requires human literacy, ethical design, and transparent communication. As I read my classmates’ reflections later this week, I’m curious how others are considering the balance between efficiency and responsibility in AI integration, and what they believe organizations owe employees when deploying such tools.

Can Feedback Elevate the Quality of Online Learning?

Reference:

Ertmer, P. A., Richardson, J. C., Belland, B., Camin, D., Connolly, P., Coulthard, G., Lei, K., & Mong, C. (2007). Using peer feedback to enhance the quality of student online postings: An exploratory study. Journal of Computer-Mediated Communication, 12(4), 412–433. https://doi.org/10.1111/j.1083-6101.2007.00331.x

Annotation:

Ertmer et al. (2007) explores whether structured peer feedback can sustain or improve the quality of graduate students’ online discussion posts in a fully online course. Using Bloom’s taxonomy as a scoring rubric, the authors examined students’ perceptions of both giving and receiving feedback and measured changes in posting quality over time. Although peer feedback did not significantly increase scores, it successfully maintained quality levels and fostered deeper reflection, metacognition, and engagement. Students valued instructor feedback more but acknowledged peer feedback as a meaningful mechanism for clarifying thinking, validating ideas, and reinforcing learning.

Ertmer et al. (2007) offer a carefully structured and methodologically transparent case study, especially notable for using a variety of tools like surveys, interviews, and rubric-based scoring. By adopting Bloom’s taxonomy as a consistent evaluation framework, the authors ensured a high degree of face validity, which is something often missing in online-learning research. A major strength lies in how they operationalized “quality” through observable cognitive indicators, rather than relying on self-reports alone. Their mixed-methods approach allowed them to capture both the stability of posting quality (quantitative) and the rich internal reasoning students engaged in while giving feedback (qualitative).

The study’s clarity in describing its procedures, anonymity protections, and reliability checks makes it replicable and trustworthy. Moreover, the article’s discussion is unusually candid about logistical constraints, like delayed feedback cycles, showing an awareness of the real-world instructional design challenges that L&D professionals regularly navigate. Overall, the study stands out for its practical applicability and its nuanced treatment of peer review as both a cognitive and social learning tool.

For Allegiant Professional Resources, where our mission is to elevate workforce learning outcomes for clients and consumers, this study reinforces a core truth: learning quality improves when learners actively evaluate and articulate understanding, not just consume content. Ertmer et al.’s insights support our belief that learning frameworks must move beyond passive LMS modules or gamified environments that prioritize activity over cognition. Giving feedback deepens learning more than receiving it and this study helps us further understand the dynamics of learning to better design effective training programs. Allegiant’s vision for a next-generation corporate learning architecture that uses reflective, socially driven, neurologically aligned pathways to strengthen memory, decision-making, and skill transfer.

As we build frameworks that tailor learning to cognitive profiles, peer-based scaffolding can become a powerful differentiator: it honors neurodiverse strengths such as pattern recognition, deep analysis, or verbal reasoning while fostering equitable, inclusive knowledge construction. This article directly informs the L&D ecosystems we design for clients, where meaningful interaction, self-assessment, and cognitive challenge become cornerstones of higher retention and real-world performance.

Applying Activity Theory to Transform Learning Impact

Reference:

Marroquín, E. M. (2025). Activity theory as framework for analysis of workplace learning in the context of technological change. Learning and Teaching: The International Journal of Higher Education in the Social Sciences, Elsevier.
https://doi.org/10.1016/j.later.2025.1000083

Annotation:

The rise of AI has happened faster than businesses and experts can adapt to the changes it has inevitably caused. Marroquín (2025) explores how Activity Theory can serve as a powerful framework for understanding how workplace learning evolves within technologically mediated environments. The author argues that as artificial intelligence and automation transform job functions, learning must be viewed not as a discrete event but as an integral part of the work activity system (comprising tools, rules, roles, community, and the object of work).

Rather than focusing on isolated training sessions, the study suggests that learning occurs through the contradictions and adaptations that arise as employees interact with new tools and changing structures. By examining these tensions, the article highlights how organizational learning can drive systemic transformation and measurable performance outcomes making this incredibly relevant to the field of organizational development.

Marroquin’s use of Activity Theory offers a rich, systems-level analysis that transcends traditional learning frameworks focused on individual cognition. The methodology draws on the framework’s core elements such as mediation, contradictions, and expansive learning which provides a structured yet flexible lens to analyze real-world complexity in workplace settings.

The strength of this article lies in its integration of theory and practice: it effectively links conceptual depth with practical implications for managing learning in AI-enabled environments. At Allegiant Professional Resources, our learning and development initiatives echo Marroquin’s perspective: learning is only valuable if it changes work outcomes. We’ve moved away from counting inputs such as “2 hours of training completed” or “5,000 skills tagged” and instead focus on impact measures, such as reduced error rates, faster cycle times, or improved decision accuracy after interventions.

Activity Theory helps us trace how those results occur by analyzing the full activity system like what tools employees use, which rules or norms guide their work, how their roles interact, and what the shared object of their activity is. When contradictions emerge (for example, when a new AI dashboard changes reporting workflows), we view them as learning opportunities rather than inefficiencies. Marroquín’s work reinforces our philosophy that training is not the outcome but instead - performance improvement is. It provides a theoretical foundation for measuring not activity, but transformation within the work system, a principle that continues to shape Allegiant’s evidence-based approach to organizational learning and impact measurement.

The Cost of Ineffective Employee Training

References:

Durgungoz, F.C., Durgungoz, A. “Interactive lessons are great, but too much is too much”: Hearing out neurodivergent students, Universal Design for Learning and the case for integrating more anonymous technology in higher education. High Educ (2025). https://doi.org/10.1007/s10734-024-01389-6

Kessler, R. C., Adler, L., Barkley, R., Biederman, J., Conners, C. K., Demler, O., … Walters, E. E. (2006). The prevalence and correlates of adult ADHD in the United States: Results from the National Comorbidity Survey Replication. The American Journal of Psychiatry, 163(4), 716-723. https://doi.org/10.1176/appi.ajp.163.4.716

Annotation:

Durgungoz’s, et al, study explores how technology-enhanced learning environments grounded in the Universal Design for Learning (UDL) framework can improve engagement for neurodivergent learners, including those with ADHD, while cautioning against overstimulation from excessive interactivity. Interestingly, the findings suggest that digital training programs are most effective when they provide flexibility, anonymity, and multiple ways to engage neurodivergent employees. The most effective programs allowed the employees to control pacing, choose preferred interaction modes, and reduce cognitive overload.

Why is this relevant to employee learning and development? According to Kessler, et al, (2006) the current adult ADHD prevalence at ~4–4.4%, and workplace studies show ADHD is associated with measurable reductions in job performance, higher absence and accident odds, and a quantifiable human-capital loss per affected worker (for example, a study of a large employer found ADHD workers averaged a 4–5% reduction in work performance and an estimated lost productivity value of roughly US$4,300 per affected worker per year).

The studies strength lies in combining both qualitative and quantitative approaches by collecting feedback from neurodivergent adults in higher education to assess emotional and cognitive engagement across different instructional formats. The researchers clearly outlines how UDL-driven technology design enhances inclusion by offering multiple means of engagement and representation, while also noting that excessive interactivity can overwhelm participants. The presentation is balanced, integrating participant voices with data analysis, and uses well-structured arguments supported by empirical findings. This approach strengthens its case for adapting UDL to corporate training by emphasizing flexibility, anonymity, and learner choice.

Allegiant Professional Resources’ mission to design corporate learning programs that genuinely enhance employee skillsets rather than simply deliver information makes the UDL approach a valuable tool in our repertoire. The study’s emphasis on UDL provides a research-based framework that supports our approach of tailoring training experiences to diverse cognitive styles and engagement preferences. Just as the article highlights the importance of balancing interaction with structure for neurodivergent adult learners, our team applies similar principles when developing corporate trainings, integrating technology that allows flexibility, pacing control, and choice in how learners engage with material.

This research reinforces the value of embedding inclusivity and intentional design into skill development programs, ensuring that each training we create is not only accessible but also effective in building lasting competencies that translate directly to workplace performance.

Additional References:

Kessler, R. C., Adler, L., Barkley, R., et al. (2006). The prevalence and correlates of adult ADHD in the United States: Results from the National Comorbidity Survey Replication. American Journal of Psychiatry, 163(4), 716-723. https://doi.org/10.1176/appi.ajp.163.4.716

Is TPACK framework effective for Executive Coaching?

Reference:

Harris, J., & Hofer, M. (2011). Technological pedagogical content knowledge (TPACK) in action: A descriptive study of secondary teachers’ curriculum-based, technology-related instructional planning. Journal of Research on Technology in Education, 43(3), 211–229. https://doi.org/10.1080/15391523.2011.10782570

Annotation:

Judith Harris and Mark Hofer (2011) examined how experienced teachers plan instruction that effectively integrates technology with content and pedagogy. Through qualitative interviews, unit analyses, and reflective journals, the researchers found that teachers’ use of technology became more conscious, strategic, and student-centered after participating in professional development focused on content-based “learning activity types.” Teachers began selecting technologies not for their novelty but for how well they served learning goals, demonstrating that effective integration requires understanding the nuanced “fit” between tools, content, and learners.

The study introduced a replicable framework for developing adaptive expertise through reflective planning and design thinking principles that extend far beyond education - hence why it is relevant to executive coaching!

For leadership and management consultants, Harris and Hofer’s study offers a powerful parallel to the challenges of coaching and organizational learning. Their methodology is deeply interpretive, reflective, and evidence-based which mirrors the way executive coaches guide leaders through awareness, experimentation, and refinement. By mapping how teachers evolve from “technocentric” to “learner-centric” thinking, the research models how leaders can move from “tool orientation” (e.g., implementing AI dashboards or engagement platforms) to behavioral fluency like integrating technology with strategy, culture, and context.

The study’s TPACK framework can be adapted for leadership enablement, where:

content = strategy

pedagogy = leadership approach

technology = tools

These facilitate decision-making and communication. HR professionals can use this lens to design more effective coaching interventions shifting from system implementation to skill integration, much like educators learned to shift focus from software features to meaningful outcomes.

Traditional Executive Coaching example:

Coach: “You’ve mentioned frustration with your team’s resistance to the analytics platform. What emotions come up for you when you see that resistance?”
COO: “It feels like they’re not moving fast enough, like they’re clinging to old ways.”
Coach: “What leadership behaviors could help model the adaptability you’d like to see?”
COO: “Maybe I could be more transparent about my learning curve too.”
Coach: “Excellent. Let’s develop a communication plan that frames your learning story and sets expectations.”

Result: The coach helps the leader become more self-aware, emotionally intelligent, and strategic in communication, but the technology integration challenge remains largely unaddressed.

TPACK Framework Executive Coaching Example:

Coach: “You’re leading a transformation that depends on your team’s ability to use data strategically. Let’s explore how your communication methods and tool use align with that goal.”
COO: “I’ve asked them to adopt the dashboard, but they still default to old reports.”
Coach: “That’s an example of a content-technology gap. What if we designed learning sessions that focus not just on using the tool but on interpreting data for strategic decisions? You could co-facilitate those sessions modeling the kind of data-driven thinking you expect.”
COO: “That makes sense. I can use our next operations meeting to walk through how I’m using the data for forecasting.”
Coach: “Exactly. That integrates the technology into your leadership pedagogy turning the tool into a platform for shared sense-making, not compliance.”

Result: The coaching moves from personal reflection to adaptive system design aligning how the leader teaches, communicates, and models behavior through the actual technology being adopted.

Coach’s Focus:

• Technology = digital tools and data systems being implemented.

• Pedagogy = the coaching approach or facilitation method (how the leader learns).

• Content = the business strategy, goals, or leadership outcomes being developed.