Recruiting does not have a data problem. It has a usability problem. Most reporting tools recruiters use today were borrowed from finance or sales, built around tagged fields, static dashboards, and quarterly views. The interview layer, where the highest-signal data in hiring lives, sits outside all of it.
AI Reports 2.0 is what changes when the reporting layer is built directly on the interview corpus, queryable in natural language, and refreshed in real time. It is not a faster dashboard. It is a different unit of analysis: every conversation, every signal, every comp number, every dealbreaker, structured and explorable from the moment the call ends.
This post walks through what changed in 2.0, the four new capabilities that recruiters and hiring leaders are running on it, the workflow shift from "export a static dashboard" to "ask a question, get an answer," and how AI Reports fits with the rest of the agentic recruiting platform.
Reporting that thinks like a recruiter
The recruiting industry has spent a decade trying to make hiring decisions from the wrong data layer. ATS fields, calendar timestamps, recruiter-tagged stages: necessary, but never the whole picture. The conversation itself, what the candidate actually said about comp, about timeline, about competing offers, about the role, is the highest-signal layer in hiring, and it has historically been the hardest to query.
AI Reports started in 2024 as Metaview's attempt to put a queryable surface on top of that conversation layer. AI Reports 2.0 is the version where it stops feeling like a dashboard and starts feeling like a conversation with your hiring data.
The shift matters because the rest of the hiring stack has changed underneath it. According to Metaview's 2026 AI & Hiring Alignment Report - surveying 505 recruiting leaders and hiring managers across North America and EMEA, fielded with Cint - 85% of companies exceeding their hiring goals use AI in hiring, and teams where AI is core to the process are 3.8x more likely to rate their cross-functional relationship as excellent. Reporting is no longer a lagging quarterly review. It is operational infrastructure, and the teams winning are the ones treating it that way.
What is new in AI Reports 2.0
Four upgrades ship together, each addressing a specific complaint we heard from customers running the original AI Reports daily.
Ask questions of the interview corpus in plain English instead of building a report. Results are returned with the source transcripts and scorecards already linked.
Filter by AI columns, ATS custom fields, scorecard submission time, or any combination. Set alerts on sensitive signals, comp, timeline, dealbreakers, so you hear about them before the next round.
Grouped reports now include AI columns as aggregation attributes. Numerical columns return averages across every conversation in the group, not a sample. Yes or no columns return percentages.
Start with the highest-signal reports we have seen customers build: candidate rediscovery, interview quality, comp benchmarks, market intel. Configure them in minutes instead of from scratch.
The underlying point is structural. Each of these upgrades treats the interview corpus as the primary data layer and the ATS as a passthrough, not the other way around. That inversion is what makes the rest of the platform line up.
Metaview's Views functionality, if you know, you know. I thought the interview transcribing was a game changer, ‘views’ is next-level. Interview insights that can be drawn down and compared against a cohort, or across all interviews is amazing!”
What customers are running on it
We worked with customers running the original AI Reports daily to figure out what 2.0 should optimize for. Four workflow stories surfaced as the highest-leverage. Each is now a templated report you can configure in minutes; each is also fully explorable from the natural-language query interface.
Candidate rediscovery as a CRM
Search the interview archive like a CRM, with no tagging required. Find candidates by skills, seniority, location, comp expectation, anything they mentioned in a conversation. The candidates a recruiter spoke to six months ago for a different role are often the strongest pipeline for a new one, and until now they have lived as untagged audio.
Real-time alerts on dealbreakers
Set alerts for sensitive signals so you hear about them in the next round, not at the offer stage. Comp expectation outside band, timeline mismatch, relocation friction, competing offers in late stages. Alerts are scoped to the same permissions Reports already inherits, so a sourcer does not get paged on a VP search.
Interview quality and panel coaching
Spot gaps in feedback or misalignment across interviewers. Pull every panel for a role and ask which competencies actually got assessed, which interviewers default to the same three questions, which scorecards landed with the most detail. That is the input to coaching conversations that are specific instead of generic.
Market intel from inside the funnel
Track themes across conversations to see where talent is moving, what comp ranges candidates are actually citing, what competitors are paying. Most market reports lag the market by a quarter; the conversations in your own funnel lag by hours.
Without and with: the workflow shift
The clearest way to understand what 2.0 changes is to look at the daily workflow of a recruiter trying to answer a single question: which candidates in last quarter's pipeline mentioned a comp expectation above band, and how did that correlate with offer acceptance.
- Export the ATS funnel report, hope someone tagged comp in a structured field.
- Open recruiter screen transcripts one by one in a separate tool.
- Build a spreadsheet that goes stale the moment any new offer closes.
- Ask an analyst for the report and wait two weeks for a quarterly review.
- Ship a presentation that everyone agrees is interesting and no one acts on.
- Ask the question in plain English. Comp is parsed from the conversation, not a field.
- Results return with linked transcripts and scorecards so the evidence is one click away.
- The view refreshes in real time as new interviews complete and offers close.
- Alerts fire when new candidates hit the comp threshold, not at the next quarterly review.
- The output is a templated report you can save, share, and rerun, with zero maintenance.
How this fits with the agentic recruiting platform
AI Reports 2.0 is one layer of the agentic recruiting platform, not a standalone analytics product. The Notetaker captures the conversation; Application Review triages inbound against the same ICP context; AI Filters surfaces the candidates that match a query; AI Sourcing acts on the signal. Reports is the layer that lets you ask questions across all of it.
Practically, that means three things matter: the data inheritance, the permissions model, and the integrations surface. AI Reports 2.0 inherits the Multi-Source context that the rest of the platform builds across panels, so a single report can pull from screening calls, debriefs, and scorecards without you having to join anything manually. Permissions are the same scope-by-job, scope-by-role model that Notes and Application Review use. And the ATS field passthrough means custom fields, employment type, seniority, anything you have configured, are filterable from day one on Greenhouse, Ashby, SmartRecruiters, and Lever.
We’ve completed over 1,900 calls using this platform, saving 77 full workdays. We’re not just automating note-taking, we use the Multi-Source feature so each interviewer goes in unbiased but informed enough to cover new ground.”
See it live: natural-language query in action
The query interface is the most visible 2.0 change. Recruiters type the question the way they would ask a colleague, and the system parses the interview corpus, the linked scorecards, the ATS context, and the comp signals into a single answer with sources attached.
Ask the question in plain English. The query layer parses intent, not keywords.
Filters stack on top of the query. ATS custom fields, AI columns, and Scorecard Submission Time are all combinable.
Results return with linked transcripts and scorecards. The evidence is one click away, not buried.
Save the query as a templated report or set it as an alert so new matches fire in real time.
Reporting is where that alignment becomes visible. When a recruiter and a hiring manager can both query the same interview corpus and look at the same evidence, the conversation moves from gut feel to shared signal.
What is available today, what is next
AI Reports 2.0 has shipped iteratively across the past 12 months. The capabilities below are live in production for every workspace today.
- Natural-language query interface: replace tagged-field reports with plain-English questions over the full interview corpus.
- AI columns inside grouped reports: numerical columns return averages, yes or no columns return percentages, iterating every conversation in the group rather than a sample.
- ATS custom field filters: filter Reports by any custom job field on Greenhouse, Ashby, SmartRecruiters, and Lever.
- Scorecard Submission Time column: show how long it takes interviewers to submit scorecards, on every ATS that supports scorecard write-back.
- Self-serve CSV export: download any report under 1,500 rows immediately from the report's actions tab; larger reports arrive via email.
- Notes Posted to ATS column: yes-or-no visibility on whether anyone posted notes to the connected ATS for a given session.
Two threads are still in beta and worth flagging because they hint at where 2.0 is going. The Reports MCP connects Metaview Reports to Claude, ChatGPT, or any MCP-compatible AI client, so the natural-language query layer can run inside whatever interface a recruiter already uses. And the Custom AI columns path makes the column-definition step itself queryable, so a workspace can build a column for any signal the interview corpus contains without engineering work.
Nate WylieSee the natural-language query layer, real-time alerts, and templated reports running on a real interview corpus.
