Note: this article was written when the most advanced OpenAI model available for Copilot was GPT4-Turbo
On November 15, 2023, Microsoft published an insightful report on Copilot's early users and its impact on workplace productivity and creativity.
Copilot allows to integrate an AI chatbot in Windows to interact with documents and applications, including word, excel, powertpoint etc. Check out this 1 minute intro video if you missed the info.
For busy project engineers, we summarize the full report found here:
Overall Productivity & Quality Increases among Copilot AI users, as reported by Microsoft:
Food for Thoughts:
"When everyone is super, no one will be" (Pixar, The Incredibles 2004)
It has been a year since we accessed chatGPT and we should recognize that AI is rapidly becoming a crucial component of our projects, seamlessly integrating with familiar software to enhance productivity across various areas. Although it is poised to solve some problems, it will inevitably introduce new, complex challenges: the surge in productivity and quality is likely to elevate average expectations for future products and services. Stakeholders may therefore expect project owners to deliver higher (added-value)/(effort) ratio.
Project owners must not only acknowledge but also proactively prepare for this intense survival competition. The forefront of this competition will be dominated either by those who own the most advanced AI and automation processes (often kept confidential) to guarantee their added value or by those who earn the highest level of trust and recognition from stakeholders.
This trend could lead to increased dependency on AI-empowered tools, potentially resulting in processes becoming locked into a single platform or service, with associated drawbacks.
Given the rapid pace of AI deployment, project owners face a dilemma regarding change implementation:
- On one hand, a time-consuming trial-and-error approach, comparing new processes in parallel, seems necessary to determine the most effective return on investment.
- On the other hand, there is a sense of urgency to remain competitive, urging the rapid rollout of new tools and putting pressure on process and data management.
Delegating the task of change is not always feasible, as it may not guarantee success and could incur substantial costs with uncertain returns. The principle of "garbage in, garbage out" applies: in many cases, the process owner is the only one who understands the required outputs and how to optimize the inputs, as data quality is crucial to successful AI integrations.
If there is no room to scale production or improve the average benefit-per-headcount, workforce reductions may become an inevitable outcome of AI automation. This could further widen the gap between large and small organizations, with very small (AI-equipped) production teams focusing on niche processes, while larger companies concentrate on volume.
Regarding regulations, the delayed availability of tools like Copilot in Europe—expected no earlier than mid-2024 due to the European Union's Digital Markets Act (DMA)—illustrates how regulatory delays can inadvertently create an uneven competitive landscape between regions.
Lastly, an essential consideration for project owners is the potential necessity of maintaining "manual" processes as a backup plan, ensuring business continuity.
