Inspiration

Different decision owners felt the need for documenting a decision outcome, along with the process which was executed to come out with the final decision. Additionally, the owners wish to review the past executed/performed decisions in future along with the comments coming from different involved collaborators, ultimately improving the overall decision process & adding high business value in their respective work areas.

What it does

Allows individual/ group of individuals configure their decision models (consisting of a combination of the decision matrix, decision tree, voting/ polling), & executing these with various facts, assumptions, outputs from previous decisions/ or decision steps to come with a decision based on the configured logic on the corresponding decision model. Not just that - the decision model owner can choose to enable “collaboration” with multiple people in the department/ team or organization to get their inputs across multiple stages of decisions during its execution that can further act as inputs to subsequent decisions steps, overall contributing to the accuracy of the outcome of the decision.

Additional business value propositions this application brings for the users of the application and thereby for the entire organization: 1) The application also offers appraising the decision process - by letting the decision-makers to step back and review the process by which decision was made coming through the feedback’s, lessons learnt & collaborators comments from previous decisions. 2) With this entire decision-making ecosystem - the application automatically puts the core decisions taken and related comments to be documented for future reference that can further act as key inputs for future decisions. 3) Enables continuous improvement & learning when any decision turns/proves out incorrect as the reasoning of every decision stages are documented for required retrospective phase for future decisions, thereby helping an organization to make better decisions in future.

How we built it

The application has the following core phases: 1) Configuration: This is where a decision owner can come and configure their respective decision model(s) using the different decision matrix, decision trees, questions & answers. They can choose to specify the individual or group of individuals they wish to collaborate with on the given decision model (this can also be changed during the team when the same decision model is executed). This entire configuration of each decision model is stored in the database. Due to the complexity of the nature of various components each decision model can have (i.e., sets of the decision matrix, trees, etc), the configuration for them was maintained in JSON and stored it against a table. Also, the different mathematical formulae that can be applied across columns/ rows of the decision matrix (editable grid) was stored along with what different formulas can be chosen depending on the type of that corresponding row/columns.

2) Actual Execution: This is where a decision owner executes their configured decision model with our powerful decision execution engine outputting the decision outcomes at each stage of decision model - with each instance of a decision model tracked on Appian records to have full visibility of the snapshot of the entire decision cycle.

3) On top of this, we have used Appian tasks to collaborate with collaborators to capture their inputs, feedback & comments, and then giving the decision owners to reconcile the comments together.

Challenges we ran into

The time we started thinking of this idea - we all knew this will help us and others to improve the decision making, but had high uncertainty in terms of its design & implementation. We had to spend a couple of weeks concluding & coming with concrete & scalable design. The aspect that adds high complexity to the design is because a decision model can consist of following multiple components: 1) Decision matrix - an excel-like structure with different formulae configured on different cells of rows/columns of the editable grid. Giving an excel-like user experience on Appian, modelling decision matrix on Appian 2) Decision Tree - Giving a decision tree-like view/ structure on Appian, modelling decision tree on Appian.

Few of the other initial challenges were: 1) Modelling the entire decision model (consisting of multiple decision matrix, decision tree) 2) How to give task collaboration capabilities to multiple people on each stage/step of a decision

Accomplishments that we're proud of

We were able to accomplish the solution conquering the above challenges, given a time we had for hackathon submission. Apart from the design & implementation - this subject being quite powerful for business yet very difficult to convey through a presentation and video - we need to think from different angles (data setup, video presentation) to come out with something that allows them to understand what this is all about. The team performed outstanding teamwork collaborating with both engineering & marketing team to prepare the required deliverables before the submission deadline.

What we learned

Following are the key learnings from this application: 1) How important are decisions to be documented every time you decide as it acts as a single source of information to come back in future as required 2) We learnt the need for involving the right people/ team at the right stages of any decision making to come out with as accurate & correct decisions. 3) The idea was a bit focussed when this was started, but as we went along in design & implementation - we as the team realized the immense potential this has in future & hence we are planned to take this to the next level as described below in “what next for decisions” section.

What's next for Decisions - A process based approach for better results

The application was built too fast & still has scope to improve the design and the performance. The collaboration with annotation, sentimental analysis and analytics are the next steps.

Additionally, the plan is to have an element of AI/ self-learning model incorporated in decisions, leveraging the value of the documented data gathered on each decision, so that solution recommends the decision owners and the collaborators of the relevant suggestions, feedbacks automatically whenever they are in process of taking any decisions.

Built With

  • appian
  • mysql
  • network-graph-plugin-component
Share this project:

Updates