OpenAI icon

We're excited to announce major updates for Endorsed's API Platform Partners.

Integrating with Endorsed is now 40%+ easier than in 2023. We've also made substantial AI accuracy improvements, in addition to standardizing and breaking down pricing to help reduce and create predictable costs.

For platforms using Endorsed for internal applicant review and testing, you can also now fully self-service using the Endorsed UI.

Integration Difficulty Reduced by 40%+

As we've been successfully integrating with new Platform API Partners, we've learned how to reduce the complexity of an integration considerably. This reduces both the time and internal cost to integrate with Endorsed.

A few major improvements:

Simplified API architecture
remote_id support on endpoints, reducing your calls in half and eliminating the need to store Endorsed IDs
Uploading attachments using public URLs
Dedicated endpoints for retrieving evaluations and scores

Some improvements on our roadmap:

New simple API for lower-scale implementations, integrable in roughly ~2 weeks
SDK
GraphQL endpoints
Paginated retrieval

Please reach out if you'd like to request any of these updates sooner.

Cost Standardization

Historically, we quoted per AI review for our platform customers. We realized this was overly restrictive because some customers have simpler needs, and there are other unique use cases we initially didn't predict.

Now pricing is done per "evaluation," which effectively is a question you'd ask the AI. For example, "Did this candidate work at a Y Combinator startup?" would be one evaluation.

We've also created a standardized pricing model for volumes up to 50,000 evaluations/mo, making the costs of using Endorsed more predictable upfront.

Increased AI Review Accuracy

For any criteria, from years of experience to industry experience, Endorsed's AI is now far more accurate in its evaluations.

Increased Human-Quality AI Review Speed

With our most powerful models, AI review speed can now reach roughly 40 candidates per minute, representing a 20-40x improvement over human review.

We're also working on new models and architectures to power general candidate search features, with industry-leading match quality at ~5 second response speeds.

Maybe Statuses

Instead of Endorsed guessing whether a candidate meets various criteria, now it will say when it's not sure. This typically happens when the dataset doesn't have enough information listed, and a human couldn't make an accurate determination either.

Full Self-Service Review & Testing

The Endorsed UI is now available for marketplace platforms to review talent internally and for other platform customers to use for testing and debugging.

Features include the ability to run AI reviews, edit and update criteria, assign labels to criteria, perform smaller sample runs, and more.