human-centred ai
Case studies.
“The noblest pleasure is the joy of understanding”
H1N1 Virus.
at a Glance
CLASSIFIED CONSULTING + ANALYSIS AUGMENTED-INTELLIGENCE MODELLING DATA EXTRACTION AND REPORTING PRESCRIPTIVE ANALYTICS
Bring the experts together in one place.
- Immunologists
- Virologists
- Veterinary immunologists and virologists
- Transport & logistics
DAY 1: Each expert was given
a chance to describe, defend,
demonstrate and link their opinions and views.
DAY 2: Taking everyone’s opinions into consideration, the question of risk mitigation was modelled in our software. to 2026.
The prediction and result of our
modelling concluded the likelihood of a pandemic had increased significantly by threefold.
Our assessment took two days with actionable recommendations for Australia sent to the Government. (The US government assessment took 6 months to produce a similar result.)
In June 2009, the World Health
Organization (WHO) declared the new strain of swine-origin H1N1 as a pandemic.
Category & brand migration in the Australian motor vehicle industry.
at a Glance
Two in three customers
change buying Segments?
INTERACTIVE AUGMENTED-INTELLIGENCE ML SOLUTIONS DATA. EXTRACTION AND REPORTING CONSUMER BEHAVIOURAL ANALYTICS. CRM LOOPING (D2SD = DATA TO SALES DELIVERY).
Identify stress points. Define loyalty.
- When are customers most likely to buy again?
- What vehicle are customers likely to buy?
- How do we influence customer loyalty?
- Using data from a large scale syndicated Automotive program, we used machine learning to establish a baseline model.
- Prescriptive analytics filled in the gaps for incomplete records.
- And, we mapped previous vehicle migration including socio-economic, psycho-graphic, and behavioural data.
Customer Centric Communications.
Identified with significant certainty the vehicle type customers most likely to buy.
Potential benefit to surprise and delight the customer. With ‘They see me and my needs.’
The prediction and result of our modelling concluded a high degree of success with an ability to continually feed the Houston we Have AI model for ongoing improvement and refinement.
Predicting individual
future behaviour(s).
The Houston We Have Model increased the likelihood of customer retention by identify timing and, in turn, motivational factors of vehicle choice and purchase propensity.
Age of previous | % Correct |
|---|---|
Most likely | 73% |
Most & second likely | 90% |
Segment | % Correct |
|---|---|
Small Cars | 79% |
Small SUV | 82% |
Medium SUV | 72% |
Large SUV | 56% |
4×4 Utility | 79% |
Aggregated Total | 73% |
APPRAISE - Better Provider
profiling for health funds
at a Glance
To determine which providers warranted increased monitoring, reminders of procedures or other forms of intervention, Houston We Have Prometheus created an automated and bespoke system for profiling ancillary providers – APPRAISE.
Based on research and experience in the identification of claims errors and other forms of potentially problematic behaviour, seven indicators were identified.
APPRAISE calculates multiple probability-based measures to rank individual providers versus their peers against these indicators.
Presented as a dashboard with graphs, this information alerts health funds to the potential need to audit and or investigate further. This advanced warning systems means health funds can help drive desired claims behaviours rather than apply punitive measures once problems occur.
APPRAISE is helping funds achieve significant reductions in claims waste, improvements in claims quality and the reduction of fraud.
