Project Titanium

Project Info

Team Name

Team Visionary

Team Members

Shubham and 1 other member with an unpublished profile.

Project Description

Intro of the problem statement

In today’s dynamically evolving technology landscape, public and private sector organisations are eager to leverage AI and automation to optimise their operations, but don’t necessarily know where to start. From analysis public data to streamlining processes, there are so many real world applications of AI that government agencies are struggling to create a plan on how to enhance their operations in an ethical and transparent manner while considering ethics, data privacy and security and winning over public trust. There seems to be a real need as well as hesitance around usage of AI, specially in the public sector.

Our Solution

This is where our solution, Titanium comes into play. Titanium is a RAG Application built using Amazon Bedrock and trained on gigabytes of publicly available workforce and task based data. Titanium analyses an organisation’s operations by performing an assessment and breaking down each employee’s work into individual tasks. It assesses which tasks can be accelerated using AI, revealing insights like jobs or functions in an organisation that can benefit the most from AI. This gives executives and decision makers a practical guide on how they can harness this powerful technology, all while addressing concerns around data governance, security and privacy.

For example, Services Australia consults InfoSys to leverage AI and automation to boost efficiency and optimise operations.

InfoSys needs to then then get data from Services Australia containing
1. Job title
2. Direct Reports
3. Individual Contributor / Management / Executive
4. Job Duties

The data must be generic, containing metadata about each job role in the organisation and must be desensitised to remove any sort of PII information. If a government agency is reluctant to share this non-sensitive data, they can always provide generic information like employee count, metadata around org structure without any specific job detail. All this information can be collected and stored inside Service Australia’s AWS Account or their SharePoint library, ensuring that none of the data goes outside the organisations tenant maintaining data privacy and security.

InfoSys can then deploy Titanium’s Foundational Model straight to Service Australia’s AWS Account which comes with pre-trained workforce data from other publicly accessible places. The model can then be combined with internal data to gain contextual information.

Titanium can almost in instantly produce data driven insights for each job, detailing which components or tasks of the job can benefit from AI. The entire organisation can benefit from the set of recommendations, employees getting a recommendation engine to perform their tasks in a more efficient manner.

It also generates a heatmap, containing a macro view of each of the different type of jobs giving a score of how much of the job can be automated using AI. This provides practical insights to decision makers as well as employees so they can operate in an efficient and transparent manner.

The platform also offers the ability to maintain a publicly accessible dashboard available to citizens that demonstrates true operational efficiency, displaying the department’s commitment to transparency and optimising operations through AI.

InfoSys can also consult government agencies like Services Australia, ATO and Department of Home Affairs to set medium to long term targets around showcasing operational efficiencies gained from AI and how they could potentially lead to actual dollar cost savings. They can get dynamic and interactive roadmaps to implement AI within these organisations. Based on data from ABS, there are nearly 2.5 million public government employees in Australia across 15 departments and 13 agencies with a total of $215 billion of wages paid during the 2022-23 financial year. Even if the government can optimise 1% spend across 3-5 years, this will lead to about 21.5 billion dollars in savings.

The solution will follow a 4 step governance approach focussing on fairness and bias mitigation, data security, algorithmic accountability and data privacy.

This solution hence leverages AI to boost operational efficiency by strategically identifying gaps in the government and recommending ways to improve. This framework can also leverage InfoSys’s Scan, Shield and Steer model to help customer’s become AI first.