Manage inbound application volume with AI-Assisted Application Review

Today, we’re releasing AI-Assisted Application Review to all customers in Ashby. It’s designed to help you manage high inbound application volume and one of the biggest improvements to our customer’s productivity we’ve launched to date.

Check out a full walkthrough below:

We know a big part of adopting new AI technology into your workflow means having the proper assurances around privacy, compliance, and safeguards to protect against candidate biases.

In this post we’ll detail how the new AI-Assisted Application Review feature works, and importantly, how Ashby helps you adopt this technology in a responsible way.

The rise of inbound application volume

We’ve seen a number of trends contribute to the influx of inbound applications over the last few years:

• Macro factors leading to more applicants on the market with longer job searches
• Remote work taking hold creating bigger pools of talent for each job
• The proliferation of technology that makes it easier for candidates to apply

Talent teams have the tough job of managing high volume inbound channels while delivering on candidate experience and still carving out time for high touch recruiting activities.

Your team may be asking questions like: How do we prevent application review from being a huge time sink? How can we surface promising applicants and quickly get back to them? How can our team spend more time on high-value hiring activities?

The new AI-Assisted Application Review feature offers a step change in efficiency for reviewing applications while responsibly using the latest advancements in AI.

We receive a high volume of inbound applications, and our team was spending significant time managing the initial review process. The AI-Assisted Application Review feature from Ashby has been valuable in helping us efficiently sort through this influx. It allows us to quickly flag candidates who meet our key skill requirements while still maintaining a human-centered approach to final evaluations. - Hootan Altafi, Head of Talent Operations, Dandy

Read on to see how it works.

How AI-Assisted Application Review works in Ashby

In your Job settings you now have the option to define criteria that’s used by the integrated AI model to analyze applicant resumes.

This allows you to document criteria directly in the ATS and introduces a new level of objectivity and consistency for your reviewers.

Applicants are evaluated against each of the criteria you've defined. Under the hood, when you run these application reviews, the AI is parsing through each resume, trying to find evidence as to whether the candidate "Meets" or "Does not Meet" the criteria you’ve defined. Once complete, the AI returns the best determination, along with rationale for how the determination was made.

You can then filter and segment your application review based on any specific combination of criteria you define. This offers a marked efficiency improvement to your application review process.

By structuring and segmenting applicants based on criteria they match, you immediately start to recognize time-savings. You can:

• Review applicants that meet all your must-have criteria first and get back to them quicker
• Verify and reject candidates that meet few or none of your criteria
• Construct a dynamic set of filters in a few clicks to suit your own approach

When you’re reviewing applications you’ll always see the AI output along with citations for how the determination was made.

Then, it is up to the reviewer to advance or reject.

Using this segmented approach means you’re no longer guessing what type of resume is coming next during your review. Not knowing what type of resume you’re evaluating with each new application is often what leads to the "time sink" during manual review.

By grouping similar applications (based on criteria you set) you stay in context during the review of each batch.

This is what enables you to review efficiently across a large volume of applicants. VP of Talent at Ashby, Jim Miller, details how this approach helped him review 1500 applications in 6 hours.

How Ashby helps you responsibly adopt AI technology

It’s not enough to just build an AI-powered feature that helps with productivity. We spent a significant amount of our product development addressing the various implications of using AI in this part of your hiring process. Here are the ways we’re helping you responsibly integrate AI technology into your workflows.

Proactive Monitoring and Compliance with AI Legislation

The AI legislative landscape is rapidly evolving. Recently, we’ve seen NY Local Law 144 take effect with Colorado SB 24-205, Illinois HB3773 and the EU AI Act slated for enforcement in 2026.

We’re tracking these requirements and have already taken measures to ensure your team is set up for compliance with these regulations. This includes:

• A bias audit conducted by an independent third-party. We’ve contracted Fairnow to support both audit and AI model inventory requirements.
• DPAs and other contractual measures in place with sub-processors to ensure customer data is not used by them to train models.
• Flexible support for configuring applicant disclosures, including the ability for applicants to opt-out of AI processing.

As these regulations continue to evolve so too will AI in Ashby to stay at the forefront of compliance.

Taking Multiple Steps to Avoid Bias

We’ve built into the product several features to help protect against bias, both implicit and explicit. This includes in-app warnings around specificity and best practices for prompting,

as well as more direct warnings around potential Equal Employment Opportunity (EEO) violations.

Additionally, before resumes are sent to AI models for processing, we redact PII to avoid the possibility of bias being introduced from model training data.

We’ve also made our third-party bias audit results publicly available here: Bias Audit for Ashby’s Criteria Evaluation Model.

Removing the potential for bias to support an even playing field is a priority and we’ll continue to invest here while employing proactive product thinking.

Clarity and Transparency Around AI Model Outputs

Like other AI features in Ashby, you can view the citations of source material next to the analysis. These citations make understanding the AI evaluations transparent. No black box.

In cases where you see an AI output that you don’t agree with, you can flag, report, and override the analysis. This gives you fine-grained control over outputs and ensures you’re not compromising important parts of your review process.

You’ll also notice some criteria may be marked as skipped or unknown.

If the AI can't make a determination, the criteria evaluation will be marked as unknown. If the resume is unreadable, the criteria will be marked as skipped in the output.

Empowering Talent Teams

Our approach takes an optimistic view on AI’s ability to enhance the productivity of talent teams, while increasing fairness and transparency for candidates.

While AI surfaces key information to support decision-making, it’s ultimately the qualified humans on your team making an advance/reject decision. Fundamentally, the technology is assistive in nature and designed to expand your team’s capacity - not replace it.

Getting Started

This feature is available on all plans, with a baseline number of credits included in each plan. Incremental usage credits are available to purchase based on your company's application volume.

To get started with AI-Assisted Application Review head over to the “Opt-in Features” section in your organization’s admin settings to enable the feature. See the full setup guide over at Ashby University. If you’re interested in trying out Ashby and the new AI-Assisted Application Review you can also Book a demo to learn more.