A Binding Workflow Must Know Who Can Bind

A placement portal makes the journey look simple. A broker enters a submission. An agent or MGA returns a price. The client accepts. A document appears. Everybody sees the same green status mark. Everybody assumes the risk became a policy.

It has not travelled through one decision.

It has crossed several, made by people who may represent different principals and hold different permissions.

The broker may owe the client advice.

The MGA may hold a contractual pen to commit a carrier inside a narrow grid.

The carrier may accept an exception outside it.

One partner profile cannot hold those facts.

The cleanest distribution journey is still a chain of people who are permitted to do different things.

The interface is not the authority

That is operating logic, not compliance trivia.

Suppose we record a quote as a bind, or record an MGA as the insurer.

Our error can stay invisible for years.

It surfaces at a loss, an audit, or a reconciliation.

Somebody asks who was allowed to say yes.

So I would hold the boundary here.

We should simplify the work of collecting evidence and moving a file.

We should not simplify the authority chain.

We design these platforms around the customer experience.

The authority chain has a different shape.

It does not get redrawn for convenience.

A licence is not a duty, and a role is not either

“Intermediary” is a useful word on a regulator’s map.

It is dangerous in a data model.

Canadian licensing is provincial and class-specific.

FSRA supervises Ontario life and accident-and-sickness agents. RIBO has its own statute for Ontario general insurance brokers. The AMF regulates in Quebec. Insurance councils operate in several western provinces.

Life, general insurance, and adjusting are separate authorities.

A life agent arranging commercial liability isn’t rescued by a confident interface.

Nor does a licence settle the duty question.

A broker is usually the insured’s agent when arranging coverage. The same broker can be the insurer’s agent for binding risk and collecting premium.

The common-law problem is obvious.

Agency law doesn’t welcome one person serving conflicting principals.

Insurance does it by statutory design.

Then it builds disclosure around the seam.

The courts are clear about the advisory half.

Fletcher v. Manitoba Public Insurance puts the broker or agent under a duty to give information and advice, not merely to execute literal instructions. Fine’s Flowers v. General Accident sharpens the edge: a broker who accepts an open-ended instruction to arrange full coverage takes responsibility for the gaps.

Neither case is about authority to bind.

That is the point.

The client’s request and the authority to promise on a carrier’s behalf are two facts, not one.

We have to record both.

Compensation deserves its own field too.

Flat commission on a policy is not contingent compensation tied to volume, growth, retention, or loss ratio. RIBO’s Ontario code requires general brokers to disclose remuneration including contingent arrangements, and Canadian disclosure expectations also reach client fees and material insurer ownership links. When we call every payment “commission,” we have hidden the conflict from the file.

A price is not yet permission

The chain starts earlier than the quote screen.

On a commercial renewal, an abeyance surfaces from Applied Epic, Power Broker, or Sigxp 90 to 120 days before expiry. The broker gathers the expiring policy, five years of loss runs, an updated statement of values, and a CSIO commercial property application.

Renewal terms arrive 45 to 60 days out.

Every missing answer now costs calendar.

Market clearance decides who may approach which market.

The first brokerage to register a risk with an insurer or MGA can block a later submission at the same civic address.

That is a distribution control.

It is not a coverage declination.

A portal that reports it as failed underwriting gives our user the wrong reason.

The audit trail then can’t explain the result.

Then the price arrives, in stages.

An indication is a price signal, nothing more.

A firm quote is bindable, and it carries the exact subjectivities on which that bind depends. A satisfactory inspection within 30 days. Evidence of a prior electrical update. A signed application. A no-known-losses declaration. A water-loss questionnaire.

They aren’t decoration on a better-looking quote.

They are conditions.

An unsatisfied subjectivity can unwind coverage.

Then comes the client instruction.

The broker presents terms and receives an order to bind.

It should be written.

By telephone, the contemporaneous file note is the defence. The client believed a certain coverage was bought; the file has to show what was actually bound. The ordering is boring, which is another way of saying we can rely on it. Indication. Firm quote. Client instruction. Authority test. Binder. Issued policy. Binding produces an interim binder identifying the insurer, limits, deductibles, forms list, and an effective time of 12:01 a.m. local standard time. In the broker workflow it stays valid pending policy issuance for 30 to 90 days. The certificate sent to a property manager and the evidence sent to a mortgagee may be the documents the outside world actually sees.

Neither confers a power the insurer never delegated.

The MGA’s pen has edges

An MGA creates the most persuasive illusion of continuity, because it can quote and issue on a carrier’s paper. Yet Canada largely doesn’t license an entity as an “MGA.” The firm holds agency or brokerage licences in each province, and its underwriters hold individual licences. On the P&C side in Ontario an MGA is registered with RIBO, while its life and accident-and-sickness business is licensed by FSRA.

The carrier remains answerable for the outsourced outcome under CCIR and CISRO fair-treatment guidance.

So an MGA is not our escape hatch.

The operative document is the binding authority agreement, also called the binder agreement.

It is the carrier’s grant of authority.

It is not the binder sent to a broker after an individual risk is bound.

It sets our business classes and territories, per-risk limits and aggregates, maximum policy period, prohibited risks, rating basis, approved wordings, permitted endorsements, and referral triggers.

It should also prohibit sub-delegation to a producing broker without prior written consent.

I find the distinction less semantic than it sounds.

In one programme, the appetite grid permits defined manufacturing classes up to $10M of insured value per location and $5M of CGL per occurrence, in every province except Quebec, on twelve-month terms.

A Gatineau location needs a written special acceptance from the carrier, naming the deviation and the terms.

Programmes may promise an answer in 2 business days.

The underwriter still waits for the written record.

A phone call that produces “yes,” followed by a bind, produces an audit exception.

No interface improvement fixes it afterwards.

The carrier may later ratify an out-of-authority bind.

Ostensible authority will generally protect an insured or a retail broker acting in good faith.

An undocumented exception is still an undocumented exception.

It makes it an indemnity, a commission clawback, or a reduction in the MGA’s authority.

Two documents called a binder make this worse.

The binder agreement is a standing contract, renewed annually.

The individual binder is interim evidence that coverage is in force before policy issuance.

At its issuance the carrier is on risk, even though the carrier’s own system may not yet know the account exists.

We need separate objects, because they answer separate questions.

Where did the power come from?

What coverage is in force now?

Subjectivities are part of the bind

The temptation is to model subjectivities as tasks that follow a completed transaction.

That is backwards.

The transaction completes with an obligation still attached.

A satisfactory inspection. A sprinkler warranty. A hot-work warranty. Confirmation of an electrical update. An updated water-loss questionnaire.

Each decides whether the quoted terms hold.

So each one needs an owner, a source, a due date, a status, the evidence supplied, and the authority required to waive, amend, or escalate it.

The 30-day inspection belongs in the binding record.

A document store is not the same thing.

Nor is the carrier’s written approval of an out-of-territory location subject to a sprinkler warranty.

At issuance we have to check that the approved warranty actually reached the policy.

Yes, we are adding a screen, and it breaks the flow.

A screen that asks whether a condition was satisfied is less elegant than one declaring the policy complete.

The first preserves the contract.

The second preserves a conversion metric.

The carrier has to see what is on risk

The first bordereau closes the gap that binding created.

A monthly premium bordereau sends the carrier one risk-level record per transaction: policy number, insured name, effective and expiry dates, class, location, sums insured, gross written premium, commission, tax, and transaction type.

The carrier loads that into policy administration and actuarial systems, and uses it for treaty reinsurance cessions.

So bordereau quality is one of our binding controls.

It is not reporting hygiene.

A missing Quebec location or a mis-coded class means the carrier’s reserving, regulatory reporting, and reinsurance do not match the risks actually in force.

Binder agreements commonly require a risk and premium bordereau monthly within 30 days of month end, with a claims bordereau on the same cycle or a tighter one.

Late or malformed reporting breaches the agreement even when the underwriting decision was sound.

Premium follows a separate chain, easily confused with it.

On agency bill, the insurer invoices the brokerage, the brokerage invoices the client, holds the funds in a statutory trust account, and remits net of commission by the account-current date.

On direct bill, the insurer invoices the client and pays the brokerage commission on a statement.

Where an MGA sits in the middle, the broker may remit net of its commission to the MGA, and the MGA then remits to the carrier under its own settlement terms.

These are not interchangeable payment states.

They decide whose money is held in trust.

They decide who carries the credit risk.

Provincial legislation requires intermediaries to keep client premiums separate from operating funds, and trust-account failures are unusually direct evidence in licensing action.

So we should reconcile the premium record, the account current, the trust position, the policy record, and the bordereau row.

We should not infer that a payment settled because the policy status says issued.

Authority does not disappear at loss

The same distinction returns when a claim arrives.

The broker reports the loss and advocates.

The broker does not adjust or settle.

An MGA or administrator may hold a separate claims grant.

It may appoint an independent adjuster, set an initial reserve, and pay from a carrier-funded loss fund.

It does not follow that it can decide coverage.

In the same programme, a $100,000 incurred cap, paid plus outstanding reserve, requires carrier notification and approval before settlement.

A coverage denial, a reservation of rights, a fatality, a class proceeding, or a likely media event goes to the carrier at any dollar value.

The MGA almost never has authority to deny.

“The portal approved it” won’t survive a coverage meeting.

OSFI’s third-party risk management guidance requires a federally regulated insurer to apply proportionate due diligence, a written agreement, ongoing monitoring, and access rights for itself and for OSFI.

In practice the enforcement is a delegated authority audit.

It is annual for a new or material binder, and often biennial once the relationship settles.

An audit may sample 25 to 50 policy files against the authority schedule, the rating basis, the wording version, the referral evidence, the licence, the bordereau, and the premium trust account.

Repeat failures matter.

A carrier can pull the pen by email.

It can reduce a class or a limit.

It can claw back commission.

It can terminate and place the book into run-off.

It can withdraw capacity altogether.

Run-off stops new binding and leaves existing policies on risk until natural expiry.

Cut-off ends insurer liability, and it is rare in direct binders because it would strand live policyholders.

We have to know which state exists.

“Inactive partner” is not an adequate answer.

Ask for the map, not the funnel

The board is shown placement speed, abandonment, quote conversion, and the tidy count of documents generated without human intervention.

Those are real measures of a real thing.

They describe the journey.

They do not describe the permission.

I would bring a binding-control map to that meeting instead.

Per policy, our map names the insured’s representative.

It names the intermediary’s licence and role.

It names the insurer whose paper is used.

It names the authority version and the clearance result.

It names the indication, the firm quote, and the written order to bind.

It names every subjectivity and its evidence.

It names the binder, the policy issuance, the billing path, and the bordereau acknowledgement.

At claim, it names the claims grant, the reportable-event rule, and the carrier’s coverage decision.

That is more than a frictionless journey wants us to show.

It is also information, in the sense we usually claim for the word.

It is connected to the authority it supports and to the evidence that proves it.

No portal of ours can infer authority from an outcome.

It can only reproduce the explicit rule and leave a record of the handoff.

Somebody was permitted to say yes

Every policy in force is the residue of one moment.

A particular person was allowed to commit a particular insurer to particular terms.

Our file has to be able to reconstruct that moment.

Everything else in the journey is administration around it.

We can automate nearly all of it, and we should.

The portal can simplify the placement.

We can’t simplify the authority that makes it valid.