Applying AI to drive business outcomes
REIMAGINING EMPLOYEE ENGAGEMENT
Reduce redundant internal user support request driven tickets by 50% creating opportunities for improved employee engagement
Whenever a new support request (installation problem, configuration problem, usage issue, performance degradation, etc.) or defect report is received from an user for internal apps, Facility, Infra, Network, it needs to be analyzed to see if any existing ticket can be utilized to automatically convert this incoming request to a self-service experience (supervised learning on 10 years of tickets of various types).
This include providing auto response with instructions, attachments, links, etc. related to the resolution for the equivalent incident in the past. Initial pilot run performance was lower than PoC leading to identification of new data source and refinement of algorithm
Toolset: GCP/Tensorflow, other Leading / Open Source ML/AI Platforms AI / ML → NLP → Intent Analysis (LDA) followed by Random Forest/ RNN. Visualization using Google Data Studio.
Impact: 30% cost savings overall but the biggest gain is the efficient way of getting things done while boosting employee satisfaction. Reduction in mean time to resolve issues with automated routing by 50%. This shaved off precious triage time (L1) on the customer queries and made resolution process faster. Additionally, for commonly asked questions, a cognitive chatbot is interacting with customers and automatically converting “simple” incoming requests to a self-service experience.
Ability to convert free wifi endpoints into on-premise sales channels and drive customer engagement.
Startup sold the solution to multiple restaurants, hotel chains for in-room services, and currently targeting stadiums. Our team looked at their transaction data and able to use predictive intelligence
- Market-Basket analysis/association rules and see which products can be offered together
- Create an affinity score between the customers and the products so as to do better recommendation
- Advanced segmentation based on product combo selection, weather, time of the day, the day of the week and more
Impact: 20% uplift in combo ordering and recommended orders. Higher revenue and traction for the startup.
MONETIZING DATA FOR A HOSPITALITY NEW VENTURE
IMPROVING CUSTOMER ENGAGEMENT FOR LUXURY RETAIL BOUTIQUE USING AI-POWERED SHOPPABLE VIDEOS
Improve customer engagement and increase brand recognition.
Worked with a new venture called RealValue. ai to build a product -service that monetizes any media asset, image or video. It does so by using our proprietary Computer Vision algorithms trained on several different data sets.
Impact: More than 30% uplift in customer engagement.
IMPROVING FORECASTING ACCURACY
Create a simpler, more accurate matching of thousands of client healthcare purchases to negotiated business contracts. Standard forecasting models are limited in several ways, including limited ability to deal with trends, seasonality, and momentum.
Reviewed machine learning literature to find the best solution for this customer. Extended an existing time-series algorithm (L1TF) to give an improved forecast that beat human forecasting for the contracts, reduced manual effort and gave the team time to focus on higher-level activity, such as interpreting results and responding to them. This resulted in a follow up to rewrite the contract matching engine to handle more complex scenarios and larger datasets.
Toolset: Hadoop, D3.js
Impact: Rewrite of the contract matching engine enabled client to handle more complex scenarios and larger datasets with
Increased record matches from 80% to 95%
Reduced time to results from hours to minutes
Data visualization replaced Excel
Streamlined conversion process with opportunity to lower cost of healthcare purchases in hospitals
Automate an error-ridden manual process to boost efficiency and improve employee engagement.
We automated a multitude of authentication and CRM entry tasks including but not limited to
Create a new employee account in the Active Directory and assign appropriate roles and permissions.
Create an account in different healthcare websites and configure permissions
Create an account in the internal time-tracking system
Impact: Achieved a high level of automation and data accuracy for onboarding 100s of employees every month in an otherwise manual process for a mid-sized healthcare cost management provider.
AUTOMATING ONBOARDING OF TEMPORARY EMPLOYEES FOR A HEALTHCARE CLINIC
Do a technology due diligence to assess the defensibility of technology architecture (including data pipeline and machine learning algorithms) implemented by a rising prop-tech startup in order to support multi million dollars of VC investments.
Assessed the startup on multiple dimensions like Value, Algorithms, Platform, Infrastructure, Culture and Defensibility through a series of interviews, document reviews and architecture reviews.
Impact: Overview and final reports help VC firm make inform decisions regarding their investments. The new venture got funded in less than 2 weeks.
TECHNOLOGY DUE DILIGENCE FOR A PROPTECH VENTURE FUND
Our core team was involved in implementing a new way to creating, distributing, and measuring work by rejiggering the annual budget process (~$350 mm) for a name brand financial services firm and mutual funds leader. The engagement started by reviewing the entire portfolio of work and then using data collected from various sources to identify the top 7 new initiatives for the calendar year.
Customer outcomes, regulatory hurdles, and many other factors were used in a workshop like setting to select these new initiatives and then they were funded only for a quarter - each subsequent quarters funding was provided based on outcomes achieved.
DATA DRIVEN PORTFOLIO BUDGETING FOR A LARGE FINANCIAL SERVICES FIRM
WORKING WITH A COHORT OF STARTUPS FOR BETTER USE OF DATA
Working with a set of funded and ambitious startups helping them define, refine and implement advanced data analytics strategies that allow for the quicker and ethical collection of data and use of the data for generating insights for the next level of product growth.