Intelligent assistants for all!
23 Jun 2024 - Frans Vanhaelewijck
From Mission down to your action list
Let’s share a pioneering concept that finds its roots in an old yet ingenious idea - building the object model for management. A proposition that might seem revolutionary at first glance but is deeply rooted in the traditional structure of organizational hierarchy.
An organization is akin to a pyramid, with the apex being the overall mission which cascades down into strategic statements. These strategies then translate into objectives and goals that shape programs, projects, sub-projects, and tasks. This shapes an intricate, almost infinite hierarchy from the mission at the top, down to the most detailed task in a subordinate’s to-do list.
Once a generic model encapsulating this hierarchy is established, it calls for a simple array of tools - a parent-child relationship framework and tagging mechanism - to connect all the elements.
However, this isn’t a standalone solution. To make this work effectively, two essential mechanisms need to be introduced.
- Firstly, an alerting framework - a ‘check engine light’ that signals when things are not going as planned.
- Secondly, an indicator framework or a dashboard where one can monitor the progress of tasks.
In a previous life, we’ve put this framework to the test and it garnered traction in specific markets.
“History doesn’t repeat itself, but it often rhymes.” - Mark Twain
Today, we have seemingly intelligent artificial intelligence systems and large language models. This brings us to an intriguing proposition: Can we flip the pyramid and work bottom-up, maintaining the same essence of automating management?
To tackle this, we need four primary components, and this post aims to describe these and seek your feedback and ideas to refine them further.
1. Daily Management
We propose confining ourselves to the daily management of small teams, frontline teams, etc. Lets consider the many disruptions and interactions that managers have to handle. We propose equipping everyone with an AI assistant, not just an account for everyone on an AI model like ChatGPT, but a proactive assistant.
2. Values and Beliefs Database
This wouldn’t encompass the formal aspects like mission, strategy, and objectives, but rather the ethos of the organization. The beliefs-base would be an embodiment of how we handle customers, interact with suppliers, and what kind of company or department we identify as.
3. Connections to Business Entities
The AI assistant must have access to essential databases such as customer, product, order, or complaint databases, as well as custom data objects unique to your company or department.
4. Iterative Input for AI Assistant
Lastly, the AI assistant must be designed to accept iterative input. It should yield three potential outcomes
- acceptance of the AI assistant’s suggestions,
- an iterative dialogue to gain more confidence, or
- an option to escalate the task to a human colleague for a second opinion.
In essence, we must always remember that while AI assists us, it’s the human judgment that must ultimately prevail.
Your insights, thoughts, and suggestions will help us refine and perfect this idea. We invite you to envision how this concept might come to life within your organization. Consider the potential impact, the transformations, the enhancements, the potential pitfalls.
I am looking forward for any feedback.