What is Parenting Calendar?

What is Parenting Calendar?

When parents get divorced, family lawyers need to make a calendar for child custody if the child/children is 12 years old or younger. Parenting Calendar is an embedded app within the main core software called LEAP. Its main utility is for creating calendars and comparing multiple scenarios so that both parents can settle on a final one.

When parents get divorced, family lawyers need to make a calendar for child custody if the child/children is 12 years old or younger. Parenting Calendar is an embedded app within the main core software called LEAP. Its main utility is for creating calendars and comparing multiple scenarios so that both parents can settle on a final one.

Problems

Problems

After interviewing expert stakeholders within LEAP who were practicing family lawyers I’ve learned the situation and also the struggles they need to deal with when dealing with a child custody case.

The problem is, parents will always fight and try to cause chaos to one another by taking unfair portions of the child custody. The lawyers needed to be a mediator in the middle of the storm taking place. The court’s objective for the child custody division is to act upon the child’s wellbeing so that they can grow into healthy adults.

This is where AI can come in to help conflict resolution by:

This is where AI can come in to help conflict resolution by:

1

Suggest Scenarios

Suggesting alternative scenarios as AI can use a holistic point of view to act on behalf of the child’s well-being.


2

Help defuse conflict

The family lawyer can use the AI suggestion to help their job being a mediator and also to divide the portion more equally to defuse the situation.

How the Parenting Calendar Works

1

The user will have control over the main master calendar in the main tab where they will make new parenting arrangement, add events, make edits

2

Next they publish scenarios into the scenarios tab. This is where they will compare scenarios in summary mode or view each scenario .

3

In the scenario tab we have the summary of the division for easy viewing and also a feature to compare the scenarios side by side.


4

The final output of the calendar will be a published scenario that will be shared through a web platform supported by LEAP’s other product called “LawConnect”.

Why an AI Feature?

Why an AI Feature?

This is where AI can come in to help conflict resolution by:

1

Holistic Approach

AI can take a less-biased neutral holistic approach suggesting alternative scenarios on behalf of the child’s well-being.


2

Help defuse conflict

AI suggestion to divide the portion more equally to calm the situation. Making both sides feel satisfied.

3

Help Make Better Recommendations

Suggesting alternative scenarios as AI can use a holistic point of view to act on behalf of the child’s well-being.


4

Less manual work

With AI being able to tweak the scenario as the user chat, family lawyers can focus more on being a mediator which the situation needed.

What the AI Editor can do?

What the AI Editor can do?

1

Tweak parameters on behalf of user

The AI can shift and edit the school holiday division. In addition, to also add, edit, delete events created by the parents. It can also ask follow up question if it needs more info.

2

Give reasoning to the division

The fine-tuned pre-trained AI can give specific reasoning when it shifts the existing division to give a fair rationale of why it divide it as it is.

3

Record logs on what has been changed

This is more so the system giving records, so that the user can keep track and also revert to a certain point after editing using the AI chat.

4

Respond to abstract language

It can understand abstract language if the user puts a prompt such as “my client want a strong father figure for his child” it can tweak the calendar scenario according to that prompt.

To access the AI feature the user can simply click the AI button within the selected scenario.

To access the AI feature the user can simply click the AI button within the selected scenario.

How the AI works in action

When the user pressed enter the AI will tweak the entire parameters of that scenario to match whatever prompt has been entered. It will also show the reasoning under each bar chart of why it shift it that way. We can see the loading/generate state of the AI shown above on “Term 3”.

Changelogs versioning

The user can also see the changelogs of all the changes the AI has made per prompt. It can be tracked via time stamp, and the user can also revert the state of that particular scenario into the old one.

Feasibility

The developers in my team are excited about this AI feature. I’ve worked with the developers on how it will be feasible to be built. The company has AI focused developers in another team, so I’ve invited them to a workshop session with my team’s developers to exchange knowledge on what can be done.


The AI’s ability to tweak the existing calendar system is feasible as long as the AI can detect the commands of how the code is structured so that the LLM can pickup on what to tweak according to every prompt in the chat. The pre-trained model’s ability to understand context is the magic behind it, although we need to do fine-tuning and to run some tests.


The ability of the AI to elaborate their reasoning on each bar chart within the scenario summary is also doable by tweaking the backend code to let the LLM to generate the response within the given placeholder area.


In regards with the backend prompt setup where we setup the role of the AI we set it up as: “Every time you respond I want you to take a role as an expert family lawyer who is making decisions on the child’s wellbeing so that the child can grow into a healthy adult, while making the parent’s feel that the division is fair” The AI’s role will be able to shift the direction on how the AI will respond within the context of creating a child custody calendar.”









The developers in my team are excited about this AI feature. I’ve worked with the developers on how it will be feasible to be built. The company has AI focused developers in another team, so I’ve invited them to a workshop session with my team’s developers to exchange knowledge on what can be done.


The AI’s ability to tweak the existing calendar system is feasible as long as the AI can detect the commands of how the code is structured so that the LLM can pickup on what to tweak according to every prompt in the chat. The pre-trained model’s ability to understand context is the magic behind it, although we need to do fine-tuning and to run some tests.


The ability of the AI to elaborate their reasoning on each bar chart within the scenario summary is also doable by tweaking the backend code to let the LLM to generate the response within the given placeholder area.


In regards with the backend prompt setup where we setup the role of the AI we set it up as: “Every time you respond I want you to take a role as an expert family lawyer who is making decisions on the child’s wellbeing so that the child can grow into a healthy adult, while making the parent’s feel that the division is fair” The AI’s role will be able to shift the direction on how the AI will respond within the context of creating a child custody calendar.”









Challenges with AI Editor

Challenges with AI Editor

At first I was envisioning this feature to be able to edit the main calendar editor, but due to some concerns from upper management not wanting the lawyer’s work to be “Too quick” as lawyers are paid by the hour has made the AI idea to be rejected.


However, I’ve suggested that the AI editor can be done within the scenarios to calibrate each scenario so that lawyers can tweak things easily at that point as most of the work is about listening to the client and being more of a counsellor. With this idea they are more open minded about it but they decided to put it in the backlog for V2 or V3 to be built.

At first I was envisioning this feature to be able to edit the main calendar editor, but due to some concerns from upper management not wanting the lawyer’s work to be “Too quick” as lawyers are paid by the hour has made the AI idea to be rejected.


However, I’ve suggested that the AI editor can be done within the scenarios to calibrate each scenario so that lawyers can tweak things easily at that point as most of the work is about listening to the client and being more of a counsellor. With this idea they are more open minded about it but they decided to put it in the backlog for V2 or V3 to be built.

Outcome

The Parenting Calendar will be built on the MVP version without the AI feature. The product manager and the managing director is still on the fence about doing the AI feature as it goes against how lawyers make money, and they don’t want to waste development resources to build a feature that does not benefit the customer.


My rationale is if the lawyers are there during the majority of back to back session acting like a counsellor, it is not about manual work adding to the timesheet to make money but rather the human to human experience. The AI can tweak the calendar and also give suggestions on what topics and what should they say to the client

The Parenting Calendar will be built on the MVP version without the AI feature. The product manager and the managing director is still on the fence about doing the AI feature as it goes against how lawyers make money, and they don’t want to waste development resources to build a feature that does not benefit the customer.


My rationale is if the lawyers are there during the majority of back to back session acting like a counsellor, it is not about manual work adding to the timesheet to make money but rather the human to human experience. The AI can tweak the calendar and also give suggestions on what topics and what should they say to the client

Back