Underperforming agents meant engineers rewriting prompts by hand and guessing. Now a self-correcting system reads scores and rewrites them.
An enterprise software company building platforms for health plans and providers for 12 years, in a regulated industry where new tools typically take weeks to get off the ground.
When AI agents underperformed in production, engineers had to rewrite prompts manually and guess at what would improve scores, a feedback loop that...
“What Anthropic has built covers the whole arc of the work in healthcare tech: research, brainstorming, execution, validation, design, security, and more, all in one place, end to end. And it’s connected to the dozens of other tools we already use every day using MCP. Claude also keeps shipping features before we even know we want them. Our engineers, and everyone else, spend their time on harder problems now because we trust Claude with the execution. That has taken us further and faster than we expected.”
Healthcare consumer experience platform for payers, providers, and employers.
Anthropic is a technology company specializing in artificial intelligence and machine learning solutions.
League's Automated prompt optimization is part of this use case:
Related implementations across industries and use cases
Compliance kept AI in pockets; security reviews backed up three weeks. After a weekend rollout, two-week sprints became three-hour sessions.
Scattered data and basic coding tools bottlenecked engineers. A 9-agent AI workflow shifts them from writing code to directing AI teams.
Prompt iterations rubber-banded: each engineer's fix overcorrected the last. Evals are now a merge requirement—no eval, no commit.
Prompt iterations rubber-banded: each engineer's fix overcorrected the last. Evals are now a merge requirement—no eval, no commit.
Novus's agent drew on analytics or code—never both—invisible to the team. Traces caught it; 60% of problems now caught before customers.
Surging calls caused long holds and overtime. A 24/7 AI voice agent handles routine payroll, freeing 700 HR partners for advisory work.
Keyword bots bottlenecked 100 agents supporting millions. Now, AI resolves FAQs, freeing staff to mine chat logs for product feedback.
Software updates were tied to rigid vehicle production cycles. A GenAI platform now frees 5,000 developers to release code independently.
Scattered AI tools and manual document searches slowed engineers. Now, a unified AI rapidly retrieves specialized technical answers.
Underperforming agents meant engineers rewriting prompts by hand and guessing. Now a self-correcting system reads scores and rewrites them.
An enterprise software company building platforms for health plans and providers for 12 years, in a regulated industry where new tools typically take weeks to get off the ground.
When AI agents underperformed in production, engineers had to rewrite prompts manually and guess at what would improve scores, a feedback loop that...
“What Anthropic has built covers the whole arc of the work in healthcare tech: research, brainstorming, execution, validation, design, security, and more, all in one place, end to end. And it’s connected to the dozens of other tools we already use every day using MCP. Claude also keeps shipping features before we even know we want them. Our engineers, and everyone else, spend their time on harder problems now because we trust Claude with the execution. That has taken us further and faster than we expected.”
Healthcare consumer experience platform for payers, providers, and employers.
Anthropic is a technology company specializing in artificial intelligence and machine learning solutions.
League's Automated prompt optimization is part of this use case:
Related implementations across industries and use cases
Compliance kept AI in pockets; security reviews backed up three weeks. After a weekend rollout, two-week sprints became three-hour sessions.
Scattered data and basic coding tools bottlenecked engineers. A 9-agent AI workflow shifts them from writing code to directing AI teams.
Prompt iterations rubber-banded: each engineer's fix overcorrected the last. Evals are now a merge requirement—no eval, no commit.
Prompt iterations rubber-banded: each engineer's fix overcorrected the last. Evals are now a merge requirement—no eval, no commit.
Novus's agent drew on analytics or code—never both—invisible to the team. Traces caught it; 60% of problems now caught before customers.
Surging calls caused long holds and overtime. A 24/7 AI voice agent handles routine payroll, freeing 700 HR partners for advisory work.
Keyword bots bottlenecked 100 agents supporting millions. Now, AI resolves FAQs, freeing staff to mine chat logs for product feedback.
Software updates were tied to rigid vehicle production cycles. A GenAI platform now frees 5,000 developers to release code independently.
Scattered AI tools and manual document searches slowed engineers. Now, a unified AI rapidly retrieves specialized technical answers.