Skip to content
EXETER AUTO RECEIVABLES TRUST 2026-2AUTO ABS · AS OF 2026-05-31 · $MSOLID = NAMED · HATCHED = RETAINED · PALE = UNKNOWNCLASS A-1 · P-1 / A-1+$16.293MCLASS A-2 · Aaa / AAA$149.650MCLASS A-3 · Aaa / AAA$156.220MCLASS B · Aaa / AA$82.210MCLASS C · Aa3 / A$85.230MCLASS D · Baa3 / BBB$110.620MCLASS E · NR$75.390MCLASS N · NR$14.850MPIMCO ABS 33% OF CLASSMFS INTERMEDIATE 0.5% OF CLASSNUVEEN ENHANCED 0.0% OF CLASSVANGUARD ULTRA-SHORT-TERM 4.2% OF CLASSSTRATEGIC ADVISERS 0.7% OF CLASSFS MULTI-STRATEGY 4.5% OF CLASSOAKMARK EQUITY + AMERICAN FUNDS95% OF CLASSCLASS D HOLDERS · $110.6MPAYDEN 0.4% · AAM/WILSHIRE 0.2% · SPONSOR 5.0%424B5 + 10-D + N-PORT + REG RR · FIGI BBG0210FK1N199.4% NOT PUBLICLY NAMEDTOTAL STRUCTURE$690.463M · COMPLETE · PER 10-D

Applied AI for decisions where being wrong gets expensive.

A surprising amount of the evidence is already public.

Public filings, regulatory records, and market data can get us a long way. This bond is one example: we show what the public record supports, link the sources, and make the gaps visible.

Your data makes the workflow specific to you.

Add your contracts, positions, policies, invoices, or internal records. We turn the combined evidence into a workflow your team can use again.

Another example

The same approach works in procurement.

The bond above is one example. The same approach works when the decision is about spend, contracts, or vendors.

Procurement

Should we accept this vendor increase?

Compare the contract, market prices, past pricing, public alternatives, and your own purchase data. Then show what the vendor uniquely provides and what you could replace.

Keep it. Renegotiate it. Knock it out.

How we work

Command. Reveal. Run. Keep.

Name the decision. Show the evidence. Run the workflow. Keep what we build.

  1. 01 / Command

    Name the decision.

    Pick the question, workflow, report, or spend decision that matters now.

  2. 02 / Reveal

    Show the evidence.

    Bring the sources together and make clear what is known and what is still missing.

  3. 03 / Run

    Run the workflow.

    Turn the evidence into a report, monitor, review flow, or other working process.

  4. 04 / Keep

    Keep what we build.

    The data work, rules, workflow, and documentation stay with your team.

A fair question

“Can I just ask Claude?”

Yes. We do too. The problem is knowing what to trust when a model sounds confident.

We connect the model to the right evidence, checks, and workflow so your team can use it in a real decision and run the same process again.

See the proof

Proof you can inspect.

Two examples built on real data, with the sources visible.

  • Source check

    Ask a question, then check the sources.

    Loom uses public data to respond to questions. Each response shows the source, what changed, and the date the data was true. Ask for evidence and you can open the full source trail.

    Source
    Public data, traced to source on request
    As of
    2026-07-29
    Read the case study
  • Procurement signal

    When does a resin price increase actually stick?

    An agent built this signal from public data in days. It combines energy prices, producer prices, and industry use to show when an increase may hold. A team's contract data makes it more precise.

    Source
    Henry Hub, BLS PPI, and WTI, all public
    As of
    2026-07-06
    See the signal
We have been using the words “quick win” a lot recently but I think what we just witnessed here redefines what a quick win is.
Operations Director at a global commercial real estate firm

Start with one decision.

We agree on the decision, show you real progress within 48 hours, and keep going only if the work is useful.

  1. Choose the decision.

    Tell us what needs to be decided, when, and what information is available.

  2. See progress in 48 hours.

    By the end of day two, you will have real work to review.

  3. Decide what comes next.

    Continue only if there is a useful next step.