• contact@verticalserve.com
  • Part of the Insight broker suite — InsightDealDesk · InsightDueDiligence
InsightDealDesk · Transaction risk placement

Agents read the data room. Brokers place the deal

The AI-first workbench for R&W, tax and contingent liability insurance on M&A deals. Eleven agents read the agreement and the diligence, predict the exclusions, compare the quotes and check the binder — every finding cited to its page. Nothing goes to a market, and nothing is bound, until a broker decides.

8deals in the demo book, mandate to post-close
$1.9Benterprise value in play on the live deals
11agents; every placement action capped at ask-first
~1¢model cost of a full Analyze on a demo deal

Illustrative: figures from the fictional demo book shown on this site. Your deals will differ.

Deal Workspace · Project Alder
The Deal Workspace for an illustrative software buyout at the quotes stage: the seven-stage strip, findings, diligence gaps, three quotes and the premium range, the agents' narrative and the deal rail
Prepared for you · Quote agentThe quote agent's prepared recommendation on an illustrative deal: select Harborline Specialty's $24M over $1.2M for $780K, why, what to negotiate out, with Accept, Change, Reject and Ask
Project Alder (illustrative): three quotes in, and the quote agent’s recommendation waiting for a broker to accept, change, reject or ask.

Built for a broker’s transaction risk practice

Representations & warranties Tax insurance Contingent liability Buyouts and carve-outs Mid-market to large-cap Primary and excess towers
Proof, not promises

Three screens from the demo book, as they are

The running product on fictional deals. The agents’ text is what Claude wrote for them; every figure is the engines’. Click a screen for the full page.

Markets & quotes · Quotes side by side
Three illustrative quotes side by side with limit, retention as a share of EV, premium, rate on line, total cost and deal-specific exclusions, and where the wording differs
1

Three quotes, one grid nobody re-keyed

The comparison engine computes retention as a share of EV, rate on line and total cost, and lines up the deal-specific exclusions and the wording that differs.

The quote agent reads that grid — not the PDFs — and prepares a recommendation for a broker to accept, change, reject or ask.

A software buyout, three quotes
Reps & schedules · No diligence
Reps in an illustrative purchase agreement with their qualifiers and source pages, four flagged red for having no diligence behind them
2

The reps with nothing behind them

Every rep with its qualifier, its schedule, the diligence report behind it and the page it was read from. A data privacy rep with no cyber diligence shows red.

Those are the reps an underwriter narrows — so the diligence gap agent raises them before the submission, not at the indication.

Where placements slip
Decision · Binder agent
The binder agent's finding on an illustrative deal: the binder differs from what was agreed in two places, with Accept, Change, Reject and Ask
3

A binder checked field by field

The agreed $1.3M drop-down retention is missing and a product recall exclusion appeared that nobody agreed. The binder agent says so, cites both, and prepares the send-back.

Catching a binder that differs
The problem

A placement is three weeks of reading and an afternoon of judgement

The judgement is what the client pays for. The reading — every rep, schedule and diligence report, every market’s appetite, every quote revision — is what the day is spent on.

Surprises

The exclusion nobody predicted

A disclosed audit or a quantified diligence finding becomes a deal-specific exclusion the client hears about at the indication.

Coverage

Reps with nothing behind them

A data privacy rep with no cyber diligence is the rep an underwriter narrows — and the one most likely to be claimed on.

Speed

A grid rebuilt per revision

Limit, retention, premium, fees, rate on line and exclusions, re-keyed every time a market revises.

E&O

A binder nobody checked

A drop-down retention missing, an exclusion added — and the binder becomes the policy.

How a run works

From Analyze to a signed decision

Every step after the button is something a broker can open and check.

1

You press Analyze

Or What would they do? — a dry run that writes nothing.

2

The data room is read

Into page-level pieces (InsightXtract in production), so every citation points at a page.

3

Facts gathered

Reps, schedules, diligence, InsightPulse intelligence, the engines’ market fit and quote grid.

4AI

Claude reasons

Over the cited facts, in a fixed shape: findings, a recommendation, what to negotiate.

5

Grounding check

A number that is not in the facts throws the answer away; the agent’s own findings stand.

6

A broker decides

Accept, change, reject or ask — signed on the decision record and counted as evidence.

Capped at ask-first. Sending a submission, approaching markets, selecting a quote, accepting a binder and signing off always wait for a person with the authority.
The agents prepare; the engines compute every figure; a broker decides. Each step is traced with its model, tokens and cost.
The agents

Eleven deal agents, each with one job and a hard limit

Every action that commits the client or the brokerage — sending a submission, approaching markets, selecting a quote, accepting a binder, signing off — is capped at ask first, whatever the dial says.

Data room

The data room against the placement checklist; drafts the request for what is missing.

Reps

Every rep, its qualifier — knowledge, materiality or flat — and the schedules it relies on.

Diligence gap

The reps no diligence report covers: the ones an underwriter will narrow.

Exclusion

Disclosed and quantified issues turned into the deal-specific exclusions to expect.

Intelligence

The target’s company 360, subsidiaries, leadership and news from InsightPulse.

Submission

The package and cover note, disclosing known issues and gaps.

Prepares the send; never sends

Market

Fit per market — lines, sector, size, capacity, retention — with every reason.

Capped at ask first

Quote

The engine’s comparison, a recommendation and what to negotiate out.

Capped at ask first

UW call

The questions underwriters will ask, answered from the data room.

Binder

Binder against the quote, policy against the binder, field by field.

Capped at ask first

Supervisor

Runs the placement controls; prepares the practice leader’s sign-off.

Prepares; a leader signs

The autonomy dial

The level for any action is the lowest of three: the agent’s level for that action and sector, the action’s hard cap, and the tenant ceiling. Here the cap and the ceiling are both L2.

L1

Draft

The agent drafts; the broker does the work. Where every action starts.

Cap and ceilingL2

Ask first

The agent prepares the decision; a broker with the authority accepts, changes, rejects or asks.

L3

Act and tell

Locked for placement actions.

L4

Act

Locked for placement actions.

AI & Agents · Catalogue
The agent catalogue: eleven deal agents across diligence, placement and bind, with their actions and starting levels
Catalogue. Eleven agents in three steps, their actions and what each never invents.
Controls

Placement policy as rules you can read, and a record of every decision

Rules Studio

Controls as sentences

Conflicts not cleared blocks quote selection; a limit above the broker’s authority refers to a practice leader; a retention outside 0.5%–1% of EV warns. Tested against a real deal and versioned.

Binder check

Field by field

Insurer, limit, retention, drop-down, premium, exclusions, enhancements and conditions — missing, added or different, graded by severity.

Decision record

Signed answers

Every control hit, every prepared decision, every answer with its note and every sign-off — replayable for a client, an auditor or an E&O reviewer.

Decision · Binder agent
An illustrative binder check: the authority control referring the bind, and the binder agent finding two differences from the agreed quote
Binder check. The drop-down retention missing and an exclusion nobody agreed — send it back.
Connected

Insurance diligence in, company intelligence in, documents read

InsightDealDesk works on its own data room. With its sibling products, the target’s diligence, intelligence and documents arrive as data, each exchange signed.

FAQ

What practice leaders ask first

Does an agent send a submission or bind on its own?

No. Sending, approaching markets, selecting a quote, accepting a binder and signing off are capped at ask-first whatever the dial says. Outside production every send is a dry run.

What stops the model inventing a premium?

The engines compute every figure. An agent answer with a number that is not among its cited facts is discarded and the agent’s own findings stand, labelled as such.

What does a run cost?

The trace records tokens and cost on every step. On the illustrative deals shown here a full Analyze of eleven agents cost about one cent; your deals will differ.

Are the markets on this site real?

No. Every market, deal, target, person and figure on this site is fictional and labelled illustrative.

See the agents work a deal

Bring an anonymised data room from a deal you have placed and watch the agents predict the exclusions, compare the quotes and check the binder in a 30-minute walkthrough.

Request a demo