Copilot was due to become available on 1 November

On 21 September, Microsoft announced that Copilot for Microsoft 365 would become generally available to enterprise customers on 1 November at USD 30 per user per month. Customers in an early access programme could use Microsoft 365 Chat in September. The other updates in Word, Excel, Outlook, and Loop were scheduled for November unless marked “coming soon”. For leaders, this was therefore a point to select workflows and prepare a pilot, rather than record broad use as an achieved benefit.

The timeline has practical consequences for project governance. Licence budgets, access groups, information sources, and training need to be ready before a team can compare the new and existing workflows. A pilot also needs a baseline. Without data on time spent producing meeting notes, collecting status, analysing data, or handing work over, it is difficult to distinguish a real improvement from the novelty effect of a new tool.

Loop would make task handovers easier

The announced Microsoft Loop features addressed teamwork directly. Participants could develop prompts together, review earlier AI interactions, and turn a table into a Loop component shared in Teams, Outlook, Whiteboard, and Word for the web. Copilot was also intended to summarise changes on a page, identify new assignments, and prepare a handover for a colleague.

This matters for project management because much delay occurs as work moves between people, meetings, and tools. A machine-generated summary can make status easier to find, but it does not replace accountability for confirming decisions, deadlines, and owners. The project manager should define which surface is the source of task status and require a named person to approve a summary before it enters governance reporting. Otherwise, a quick summary may make an ambiguous agreement more visible without making it more accurate.

New analytics features in Excel and Power BI

Microsoft also presented Copilot in Excel as assistance for analysing, formatting, and editing data, suggesting formulas, highlighting values, and producing visualisations and forecasts. This could lower the barrier to exploration, but the answer still depends on table structure, definitions, and data quality. A leader asking for “profitable customers” still needs an agreed definition of profitability, the correct period, and a policy for shared costs.

The Power BI update on 14 September showed another side of the same development. Models could be edited in the Power BI service with better preservation of layouts between the service and Desktop. Relationships could be edited in the properties pane, and testing of row-level security was improved. Git integration added paginated reports and branch switching in a workspace. Analytics development was consequently moving closer to a traceable team process in which changes can be tested and reviewed before publication. That is particularly useful when several developers maintain the same model or report portfolio.

Research showed uneven benefits from AI

A study published by the Harvard Business School AI Institute on 21 September followed 758 consultants in experiments using GPT-4. For tasks within the model’s capability frontier, the researchers reported gains in speed, quality, and completion. On a task outside that frontier, the AI group performed worse. They described the boundary as “jagged” because tasks that appear similarly difficult to people can produce very different results with AI.

This provides a useful adoption framework. A team should split the workflow into concrete tasks and check each outcome: summarising known material, drafting a project plan, analysing a defined table, or proposing a visualisation. A subject specialist should assess quality, and errors should be logged alongside time saved. The collaboration and analytics features can then be used where they demonstrably help, while decisions beyond the model’s tested capabilities retain clear human control.

Sources

Microsoft 365: Copilot general availability and Microsoft 365 Chat, 21 September 2023

Microsoft 365: Loop built for co-creation, 22 March 2023

Microsoft Power BI: September feature summary, 14 September 2023

Harvard Business School AI Institute: The jagged technological frontier, 21 September 2023

Join the discussion

Which experience from your own organisation is most relevant to this story? Share what you learned.