Lighthouse - AI Decision Automation Platform

Incedo Lighthouse harnesses the power of AI in a low code environment to deliver insights and action recommendations, every day, by leveraging the capabilities of Big Data at high speed.

Incedo Lighthouse implements the workflow consisting of: AI driven problem discovery and root cause analysis, statistical experimentation, remedial action recommendation, and business impact monitoring.

Below are the typical stages in an engagement to deploy Incedo Lighthouse:

1. Scoping and solutioning: This includes (a) customer use case detailing - understanding of business metric involved and their interrelationships, (b) data understanding - data sources, data schema, data definitions and dictionary, data gaps and mitigation plan, data characterization such as volume, and update frequency, (c) system integrations needed - on the inbound and outbound side, and (d) implementation approach finalization - extensions/changes needed to the solution, deliverables, deployment options and timelines.

2. Solution customization: Build-out of custom requirements, including minor to intermediate changes to the solution stack (front and backend, training ML models on customer data, data and model pipelines, system integrations). Incedo Lighthouse integrates with multiple sources of data through ready or open-source connectors and with systems of execution such as Salesforce for operationalizing the recommended actions. The input data that is required to run the ML models and other visualizations is ingested into a datalake, that is custom created for the use case, through the data pipelines setup to connect with raw data sources. These pipelines can be configured to run at a predetermined frequency e.g. daily, weekly.

3. User Acceptance Testing and Pilot: Post the solution is customized, it is implemented on a limited scale e.g. for a 1-2 BUs / product lines / markets, for limited customer segments, and limited datasets, to get an understanding of failure points to be able to solidify the stack.

4. Roll-out to full scale - Based on the success of pilot, plan and schedule roll-out to the full scale under scope and measure impact over a period of time.