Treasury’s Evaluation of M365 Copilot: Employees Require a 13-Minute Daily Increase to Offset Licence Expenses
We independently review everything we recommend. When you buy through our links, we may earn a commission which is paid directly to our Australia-based writers, editors, and support staff. Thank you for your support!
Brief Overview
- A trial of Microsoft 365 Copilot by the Australian Treasury indicates that a mid-tier government employee requires only 13 minutes of weekly productivity savings to warrant the licensing expense.
- The Australian Centre for Evaluation’s (ACE) internal review contrasts with the broader appraisal of Copilot’s benefits from the Digital Transformation Agency (DTA).
- Even though the trial pointed to likely time savings, it did not provide specific measurements, resulting in ambiguity surrounding Copilot’s overall impact.
- Less experienced employees recognized the potential for gaining more time for skill enhancement and higher-level tasks.
- Obstacles included system limitations, security issues, and insufficient training resources, all of which impeded both adoption and performance.
- The DTA opted against further extensive government trials of generative AI, allowing separate agencies to perform their own assessments.
Assessing Microsoft 365 Copilot within the Australian Public Sector
Microsoft 365 Copilot, the AI-driven tool aimed at boosting productivity, has faced scrutiny after a 14-week examination by the Australian Treasury. The findings illuminated some possible advantages but also brought forth concerns related to its efficacy, cost justification, and wider adoption throughout government entities.
Government Employees Only Require 13 Minutes of Efficiency Gains
The Treasury’s internal analysis by the Australian Centre for Evaluation (ACE) determined that an APS6-level public servant—whose salary ranges from $94,300 to $114,243 per year—only needs to save 13 minutes weekly on administrative duties to make the Copilot licence financially justifiable. While this establishes a seemingly modest target for efficiency improvements, the analysis did not provide concrete time savings data.
Are These Savings Achievable?
ACE considered the 13-minute productivity gain “likely,” but the lack of solid evidence makes it challenging to firmly back this assertion. In a separate review, the Digital Transformation Agency (DTA) proposed a significantly more positive average time saving of one hour daily. If this claim holds true, it would imply that Copilot could easily justify its cost.
Advantages of Copilot for Junior Employees
A notable insight from the trial was the potential for junior employees to gain from AI-assisted support. The analysis indicated that Copilot facilitated the reduction of time spent on routine administrative tasks, enabling junior personnel to concentrate on skill development and more high-priority work such as creating policy briefs or conducting data analysis.
Long-Term Consequences
The results imply that AI tools like Copilot might significantly enhance professional growth for junior employees. By managing repetitive administrative work, AI assistants could enable newcomers to engage in more significant tasks sooner in their careers.
Challenges in Quantifying Productivity Improvements
A major hurdle in the evaluation was the difficulty of assessing Copilot’s effects. Although many participants reported experiencing efficiency gains, their managers often did not perceive a marked difference. This indicates that while AI may assist individuals, it doesn’t necessarily result in overarching organisational enhancements.
Incomplete AI Work Outputs
Another issue was that Copilot frequently did not produce fully finished work, necessitating human involvement for completion. Consequently, it became challenging to ascertain if the AI actually conserved time or merely shifted workload from one phase of a task to another.
Limitations in Security and Training
The Treasury’s strict security and privacy protocols curtailed Copilot’s capabilities, rendering it less productive than unrestricted AI tools available outside the governmental framework. Furthermore, employees received minimal training on how to effectively utilize Copilot, further diminishing its potential effectiveness.
Lost Opportunities Due to Inadequate Resources
The absence of dedicated onboarding and educational resources meant that numerous participants were not adequately prepared to fully exploit Copilot’s potential. A more organized rollout coupled with comprehensive training could have yielded clearer perspectives on its real value.
The Future of AI in Public Sector
As of early 2025, the Digital Transformation Agency has no intention of initiating any further comprehensive government-wide AI trials. Instead, individual agencies will be tasked with conducting their own evaluations to ascertain whether AI tools like Copilot are worthwhile investments.
Is Custom AI Development an Option?
The government is not currently investigating a specialized, all-encompassing AI model. However, agencies may decide to procure or create tailored AI solutions to satisfy their unique requirements.
Conclusion
The Australian Treasury’s investigation into Microsoft 365 Copilot demonstrated potential but also underscored various challenges related to implementation, training, and assessing efficiency gains. While the AI tool could enable junior employees to prioritize skill growth, the broader organizational advantages remain uncertain. With the DTA stepping back from comprehensive government trials, the decision to adopt Copilot and similar tools now rests with individual agencies.
Q&A Session
Q: What exactly is Microsoft 365 Copilot?
A:
Microsoft 365 Copilot is an AI-supported assistant integrated within Microsoft’s productivity tool suite, crafted to automate and optimize administrative functions.
Q: How much productivity would an APS6 employee need to save for Copilot to be financially viable?
A:
The Treasury’s review states that an APS6 employee needs to save around 13 minutes weekly on lower-value tasks to make the cost of a Copilot licence justifiable.
Q: Did the trial demonstrate that Copilot enhances productivity?
A:
The trial suggested that productivity enhancements were “likely” but lacked definitive data to measure time savings accurately.
Q: Which group benefits the most from Copilot’s use?
A:
Junior staff seemed to gain the most, as Copilot allowed them to redirect time towards higher-value projects and professional growth.
Q: What primary challenges did implementing Copilot entail?
A:
Significant challenges encompassed security limitations, insufficient training, and obstacles in quantifying its effect on productivity.
Q: Will the Australian government proceed with AI trials in the future?
A:
As of early 2025, the Digital Transformation Agency does not plan to continue its extensive government AI trials, permitting individual agencies to conduct their evaluations.
Q: Is there a possibility for the government to develop its AI model?
A:
Currently, there is no exploration of a specialized generative AI model by the government, although individual agencies might consider creating or acquiring specific AI solutions.
