Loka’s Custom-Built LLM Improves GenAI Talent-Assessment Platform
Industry
Analytics, SaaS,
Human Resources,
Corporate Training
Tech & TOOLS
Amazon SageMaker, Amazon Bedrock, AWS Lambda,
AWS API Gateway
Teams & Services
GenAI, Data and
ML Engineering
milestones
84% team improvement,
100-200 hrs saved/mo
Industry
Analytics, SaaS,
Human Resources,
Corporate Training
Tech & TOOLS
Amazon SageMaker, Amazon Bedrock, AWS Lambda, AWS API Gateway
Teams & Services
GenAI, Data and
ML Engineering
milestones
84% team improvement 100-200 hrs saved/mo
Underprepared or inexperienced managers diminish employee well-being and performance.
Deeper Signals--and GenAI-- can help.
The situation
Most new managers receive no formal management training, an oversight that takes a toll on teams and companies. One survey found negative experiences with a poor manager resulted in 41% of employees feeling stressed or anxious about reporting to work and 31% losing confidence in their company overall.
The goal
Deeper Signals provides personality assessments that help managers leverage greater understanding of their teams’ unique traits and skills to drive more positive trends. Deeper Signals saw an opportunity to enhance its offerings with GenAI and turned to Loka and Amazon Web Services to bring the vision to life.
The solution
Loka built a foundational architecture for Deeper Signals on Amazon Bedrock with Anthropic’s Claude, Langchain and ChromasDB. Loka used retrieval augmented generation (RAG) to help the large language model (LLM) understand Deeper Signals’ methodology of assessing personality characteristics.
The AI takes into account Deeper Signals’ Core Drivers personality model, which measures 12 behavioral characteristics that shape the way people work, collaborate and lead. Then it automatically generate insights about employees’ talents based on their voluntary responses to surveys.
The result
Using the Deeper Signals platform, every manager can now query their team reports and receive AI-assisted coaching playbooks personalized to the unique personality profile of their team. Managers can ask AI to help them solve a variety of common leadership challenges and instantly receive a step-by-step guide to help coach their team to success.
By leveraging GenAI and Amazon Bedrock, Loka is helping eliminate conservatively 100-200 people hours per month building reports. Information and guidance that previously could cost thousands of dollars to acquire from executive coaches is now available to every team and leader.
Project highlights
This new service offering is currently being used by an exclusive set of Deeper Signals customers to support leadership and team effectiveness at scale, with wider availability expected in 2024.
84% team improvement
Using GenAI and AWS services, Loka enabled Deeper Signals’ personality assessments to deliver more accurate insights and offer suggestions to improve manager coaching. The platform has helped users achieve 84% team improvement with Deeper Signals feedback.
Efficiencies through expertise
Loka’s RAG-based solution outperformed foundational models, even though the information used to augment the model is publicly available.
Better quality, fewer hallucinations
To create and augment the LLM, Loka utilized Amazon Bedrock because it enables teams to easily switch between models for different types of queries. This modularity was essential for enhancing the product quality and accuracy.
Bias for action
Using AI to generate on-demand coaching playbooks, Loka and AWS helped Deeper Signals deliver personalized coaching feedback at a scale and speed. At 1-2 hours per report and 100 reports per month, customers save 1,200-2,400 hours on manual reporting every year.
"Assisted by AI, every manager can deeply understand their teams’ unique dynamics and receive on-demand coaching support to unlock performance and teamwork.”
