Revolutionizing lending with AI
We're building a technology platform at the forefront of modernizing the mortgage lending ecosystem.
“Stare & compare” killer
Proprietary-tuned AI processes with OpenAI GPT
Custom chatbot
Custom-tuned LLM model with Google Gemini
Loan origination system
AI-underpinned
Data-based AVM
(Automated valuation model)
Partner benefits
- Lower costs
- Eliminate manual work
- Reduce bias
- Increase speed
- Improve customer satisfaction
- Enhance accuracies
By the numbers
- 95%+
Improvement in manual upfront document review labor
- $8M+
Removed in
unnecessary costs - 100%
Increase in 5-star reviews mentioning speed YoY
- 80 NPS
+2 pts from 78 (exceptional net promoter score)
Model foundations
- 6+
years of originations data
- 168k+
total documents
- 1.7M+
pages
- 2M+
attributes
Our AI team
It's more than AI data scientists who unlock the value of AI in design systems - it's a cross-functional “AI Engineering" team of data scientists, data engineers, and application, product, and platform engineers.
While model training is important, by orienting ourselves cross-functionally around AI we are moving faster and more productively.
AI Eng works closely with Product and “Curation” teams of subject matter experts.
Figure's CTO, Chief Data Officer and Legal/Compliance teams serve as cohesive leadership.
Our AI driven lifecycle
- 1Extract
AI Eng team extracts 12 areas of data points that were traditionally key to the “stare and compare” process
- 2Evaluate
AI Eng team evaluates data - including 168K+ total documents, 1.7M+ pages and 2M+ attributes
- 3Provide feedback
“Curation team” of Customer Service agents and managers provide feedback based on hours of daily interactions with customers and the technology
- 4Fine-tune
AI Eng team fine-tunes the model in ways that improve accuracy and speed
Our AI principles
At Figure, we're confident in the power of AI to improve how we work and how we positively impact customers, partners and the industry at large. We're applying AI in several aspects of our business today and expect to accelerate its usage for years to come. We'll revisit and update these AI principles as part of our commitment to being ethical, accountable and values-driven.
- We ensure models can be explained
- We gather feedback across Figure to improve the models and AI projects
- We establish clear guidance on AI models' strong points and limitations
- We monitor and regularly disseminate key performance metrics and outcomes with our teams
- We continuously improve performance and usefulness to Figure and our stakeholders
- We monitor and evaluate the latest AI innovations with an eye to testing and deployment within Figure
- We monitor deployed models for unexpected behavior, ensuring timely detection and mitigation of issues
- We strive to ensure AI systems are designed to be fair, unbiased, and equitable for our stakeholders
- We curate our AI outputs to confirm our solutions are working as intended
- We protect user privacy and data security in the models
- We work cross-functionally, fostering collaboration to align goals and allowing for different perspectives to achieve the best outcome