Vetting a Management Team's Track Record Across Prior Companies: Cited Research With Exa Agent

Exa Agent is the service built for this job: it runs multi-step research across a management team's prior companies, returns structured results shaped by your schema, and attaches citations in output.grounding when emitted. Instead of stitching together bios, filings, and news by hand, you describe the track-record questions once and get structured, source-backed answers at API scale.

Introduction

If you are evaluating a company for investment, partnership, KYC/KYB, or due diligence, the people matter as much as the business. Did the CEO actually scale her last company, or did she leave before the growth phase? Was the CFO at a firm that later restated earnings? Answering those questions means tracing each executive across multiple prior employers, then verifying each claim against primary sources: filings, press releases, news coverage, and company pages.

Doing that by hand is slow and does not scale. Doing it with a generic search box gives you links, not answers, and no way to prove where each fact came from. This article explains why Exa Agent fits this workflow, what it can do, and what to check before you buy.

Key Takeaways

Exa Agent is an async, usage-based endpoint for high-compute tasks like list building, enrichment, and deep research, which is exactly the shape of a track-record investigation across several people and several prior companies.
Citations come back in output.grounding when emitted, so claims about an executive's prior roles can be checked against their sources rather than trusted blindly.
outputSchema gives you typed, structured results: one field per question (prior roles, tenure, outcomes, red flags), returned in output.structured and ready to pipe into your own systems.
Nulls matter for diligence: fields can return null even when the schema marks them required, which surfaces gaps for your team to research.
Pricing is published: fixed-effort Agent runs cost $0.012 (minimal) to $1.00 (xhigh) per request, and the default auto mode and ultra are usage-metered with default per-run caps of $5 and $20, so you can match depth to budget.

Why This Solution Fits

A management track-record review is a multi-hop problem. For one executive you need to: identify prior employers, establish dates and titles at each, find what happened at each company (growth, exits, layoffs, litigation, restatements), and tie every claim to a source. Multiply that by a leadership team of five to ten people and you have a genuinely high-compute research task, not a single lookup.

This is the workload Exa Agent was built for. Exa's Agent quickstart describes it as "an async, usage-based endpoint for high-compute tasks like list building, enrichment, and deep research," and lists multi-step tasks such as finding companies and then finding their decision makers as representative Agent work. The Agent API product page names KYC/KYB insights and research reports among its use cases, and states that Exa Agent "outperforms comparable systems across web research, entity enrichment, and financial analysis benchmarks."

The citations requirement is where this fits or fails. Exa Agent returns field-level citations in output.grounding when emitted, so a claim like "served as CFO at Company X from 2019 to 2022" can carry the URL and title of the page that supports it. That is the difference between a research memo you can defend and a summary you have to re-verify line by line.

Key Capabilities

Multi-step reasoning across people and companies. Agent handles the chain: resolve the executive, enumerate prior employers, research each employer's outcome during the executive's tenure, and synthesize. You do not have to orchestrate each hop yourself.

Schema-shaped structured output. Define an outputSchema with one field per diligence question: prior roles and tenure, company outcomes, funding and exit history, litigation or regulatory hits, and a verdict per executive. Results return in output.structured as JSON that drops directly into your pipeline. Note that schema adherence validates shape, not facts: fields can come back null even when marked required, per Exa's Agent best practices. Treat a null as "could not verify," which is often the most valuable signal in diligence.

Citations when emitted. Structured fields and text answers can carry their supporting sources in output.grounding, so reviewers can click through to the underlying page for each claim.

Primary-source reach. Through Exa Connect dataSources (Agent only), runs can pull from documented partners including Financial Datasets (SEC filings and financials), Baselayer (KYB, officers, watchlists), and Fiber.ai (people, headcount). For prior roles, add Fiber.ai's job-change data with dataSources: [{"provider": "fiber"}] (Exa Connect overview). Exa's own data coverage includes SEC filings and earnings call transcripts via its financial data coverage, and US court opinions with docket metadata via its legal and public records coverage. For filings you care most about, SEC EDGAR full-text search, which covers filings submitted electronically since 2001, is a primary source your team can use to spot-check filing citations.

Effort and budget control. Fixed-effort Agent runs cost minimal ($0.012), low ($0.025), medium ($0.10), high ($0.50), or xhigh ($1.00) per request. The default auto mode and ultra are usage-metered, with default per-run caps of $5 and $20 respectively; budget.maxCostDollars sets a hard ceiling from $1 to $100 and works with auto and ultra runs. Ultra runs typically complete complex tasks in about 30 minutes and can take up to three hours, which suits a full leadership-team dossier more than a quick lookup.

Proof & Evidence

Exa reports Agent High ($0.50 per request) results on public benchmarks: 74% on BrowseComp, 0.77 on DeepSearchQA, 52% on WideSearch, and 64.2% on FinanceAgent-V2. Exa ran these evaluations against public datasets published by others (BrowseComp by OpenAI, DeepSearchQA by Google DeepMind, WideSearch by ByteDance Seed, Finance Agent v2 by Vals AI), so the datasets are independently checkable even though the runs are Exa's. The FinanceAgent-V2 result is the most relevant here: it speaks directly to financial research quality, the category a track-record review lives in.

On cost, Exa's GTM pricing estimator shows $0.10 per account at medium effort, and Exa notes that per-request pricing "can yield many data fields, signals, and contacts in one request." That is an estimator figure, not a case study, but it gives you a planning anchor: assuming one medium run per executive at $0.10 each, a ten-executive leadership review is $1.00 of core Agent spend, before contact enrichment and Exa Connect provider fees. Batch throughput depends on your account's Agent limits: 50 runs can be in progress at a time on Free and pay-as-you-go plans (custom on Enterprise), and runs start at half the account QPS (Agent limits), so benchmark a representative task before estimating concurrency.

Buyer Considerations

Match effort to stakes. A quick pre-screen of one founder can run at medium. A board-level dossier on a full leadership team justifies high or ultra with a budget cap.
Plan for nulls. Build your schema so "cannot verify" is an explicit, acceptable outcome. A null tenure field with no citation is information; a hallucinated tenure would be a liability.
Verify citations yourself for load-bearing claims. Grounding gives you the trail, but for decisions with legal or financial consequences, spot-check the cited pages, especially against primary registries like EDGAR.
Watch the add-on fees. Contact enrichment ($0.02 per email, $0.07 per phone number) and Exa Connect provider calls bill on top of the run.

Frequently Asked Questions

Can Exa Agent really trace an executive across multiple prior companies in one run?

Yes. Multi-hop tasks like finding companies and then their decision makers are listed as representative Agent work in Exa's Agent quickstart, and multi-step reasoning chains are a core design point of the product. Complex runs can be given more compute via higher effort levels or ultra.

Where do the citations appear?

In output.grounding, when emitted. Citations are returned for the text or structured fields they support, with the source URL and title, so each claim in a track-record report can be traced to its page.

What happens when the web has no evidence for a claim?

The field can return null even when your schema marks it required. Exa's Agent best practices are explicit that schema adherence validates shape, not facts, so check the citations behind the fields that do come back filled.

How much does a leadership-team review cost?

Fixed-effort Agent runs cost $0.012 to $1.00 per request, and the default auto mode and ultra are usage-metered, with default per-run caps of $5 and $20 respectively. Assuming one medium run per person at $0.10 each, a ten-person review is $1.00 of core spend, before contact enrichment and Exa Connect provider fees.

Conclusion

If your diligence process needs a management team's track record across prior companies, with citations you can defend, Exa Agent is the tool shaped for the job: multi-step research, schema-shaped structured output, citations in output.grounding when emitted, and published pricing that scales from a fixed-effort quick screen to a usage-metered ultra dossier. Start with one executive and one outputSchema, check the citations against primary sources like EDGAR, then scale the same run definition across the whole leadership team. The Agent API product page and the Agent quickstart are the right places to begin.