Two Different Jobs: A Marketplace That Finds Aircraft vs an Agent Model That Works the Quote

A £31,000 quote for a Farnborough to Nice return can be £4,000 wrong before anyone has opened the PDF. The error is rarely the operator's arithmetic. It is the gap between the number on the page and the trip the client will actually fly.

Avinode is a marketplace and communication layer. It shows which operators have aircraft available, carries the quote request out, and carries the operator's response back. That is a real job, and it is done at scale. Nothing in that description covers what happens once a price lands in the broker's inbox.

villiersOS is built for that second job. It is an agent model: software that receives the quote and acts on it, re-pricing it against the routing, vetting the operator, moving the trip through approval stages and generating the contract. Anyone weighing villiersOS vs Avinode is therefore not comparing two versions of the same product. They are comparing a tool that finds supply with a tool that works the response.

The two models also differ in who does the work between steps. In a listing-based workflow, the broker's staff sit between every stage and move the data by hand. In an agent model, software moves it and staff review the exceptions.

This piece follows one trip to show where the models diverge: EGLF (Farnborough) to LFMN (Nice) and back, flown on a Bombardier Challenger 350 based at Luton (EGGW), which means a positioning leg before the client boards.

The Quote Lands: What Actually Happens in the Next 40 Minutes in a Listing-Based Workflow

The operator's response arrives as a figure, a set of terms and an aircraft registration. In a marketplace-led workflow, the next 40 minutes belong to the broker's team. The pattern below is typical of a small desk, not a measured industry average.

The first ten minutes go on reading. A broker opens the quote, checks it against the original request and notices whether the operator has priced the trip as a single return or as two one-way sectors. Those are different commercial animals, and the response does not always say which it is.

The next fifteen go on retyping. The headline price, aircraft type, crew arrangements and terms move into a spreadsheet or a CRM, then into a client-facing proposal on the broker's own template. Each manual transfer is a chance for a transposed digit or a dropped line.

The last fifteen go on judgement calls made without full information: is this operator one we have used, does its air operator certificate cover the route, and what is its recent safety record? Some brokers check these every time. Others rely on memory, which is cheaper and less reliable.

None of this is a criticism of the marketplace, which was never designed to do it. But the quote has to be competitive within hours, because operators reprice and aircraft get taken. If the broker takes 40 minutes per quote and is comparing four responses, the desk has spent close to three hours before a client has seen anything.

Re-Pricing Against the Real Routing: Positioning Legs, Fees, Fuel and Taxes the Headline Number Hides

The Challenger 350 seats nine to ten passengers in a cabin 7 ft 2 in wide and 6 ft high, with a range of roughly 3,200 nm. Farnborough to Nice is about 1 hour 45 minutes each way, well inside that envelope. The headline return quote for this trip might sit between £26,000 and £34,000, depending on dates and the operator.

The headline hides at least four things.

First, the positioning leg. The aircraft starts at Luton, so it flies EGGW to EGLF empty, about 25 minutes, before the client boards. If the operator stays in Nice for three days, the quote usually carries crew overnights and parking. If the aircraft flies home empty and returns for the client, the quote carries two extra sectors. Both appear as different line items, and the second can add several thousand pounds.

Second, airport and handling fees. Nice in peak season carries higher parking and handling charges than Farnborough, and some operators fold them into the price while others add them afterwards. A quote that excludes them is cheaper on the page and dearer on the invoice.

Third, fuel. A quote may be built on a fuel price from a week earlier, and some carry a fuel adjustment clause. Re-pricing against current uplift costs at EGLF and LFMN is a calculation, not a guess.

Fourth, taxes. UK Air Passenger Duty, French aviation levies, emissions-trading obligations and the VAT treatment of each leg all depend on the routing and the operator's registration. They change the real cost, and a broker who ignores them quotes a number the client cannot rely on.

An agent model re-prices these on arrival. It reads the operator's response, rebuilds the cost against the actual routing and flags any line that differs from the headline by more than a set tolerance. A listing-based desk can do the same by hand, and good ones do. The difference is that software does it every time, in seconds, for every quote.

Acting on a Quote: Operator Vetting, Stage Gates and Contract Generation Without a Retype

Once the price is right, three things have to happen before a client signs: the operator has to be vetted, the trip has to pass internal approvals, and a contract has to exist.

Operator vetting is the step most often compressed under time pressure. An agent model can check the air operator certificate, the aircraft's registration against the quoted type, insurance cover and the operator's history on this broker's own books, then attach the result to the quote. A person reviews it. The check happens before the client sees the number, not after.

Stage gates are the second piece. A trip moves through defined stages: request, quote received, re-priced, approved, contract with client, signed. Software can refuse to advance a trip until the preceding stage is complete. That matters in a business where a half-checked quote that reaches a client becomes an informal commitment.

The third piece is contract generation. Once a quote is approved, the operator's terms, the verified price, the routing and the passenger details can populate the broker's contract directly, with no retyping. The retype is where errors enter: a wrong registration, a mismatched date, a price copied from the superseded version.

What does a broker give up? Control over the in-between moments. When software acts, a person no longer reads every line as it passes through, so the exceptions and the tolerances have to be well designed. A badly set tolerance means a bad quote passes quietly. The broker also depends on the quality of the operator's response data: a quote sent as a scanned PDF is harder to act on than one arriving as structured fields.

What does the broker gain? On a trip like this, the 40-minute handling window shrinks to a review of flagged items, and the person who used to retype spends the time on the client. The same architecture could run in other brokered sectors where a supplier's price must be checked against real conditions before a contract exists, such as yacht charter, where fuel, berthing and crew terms also sit behind a headline rate. That is a possibility, not a present operation.

Where a Marketplace Remains the Right Tool, and How the Two Models Can Sit Side by Side

A marketplace is the right tool whenever the hard problem is finding supply. A broker with an unusual request, such as a Dassault Falcon 8X positioned from the Gulf on a short fuse, needs breadth of operators fast, and a listing platform provides it. An agent model cannot invent availability that nobody has published.

A marketplace also suits desks with low volume. If a broker handles a handful of trips a month, the manual steps cost little, and software built to act on quotes is more machinery than the workload justifies. A rough threshold: below perhaps 20 quotes a week, the retyping is tolerable. Above 50, it becomes the bottleneck.

The two models are not exclusive. A broker can source availability and receive responses through a marketplace, then pass the operator's quote into an agent layer that re-prices, vets and drafts the contract. The marketplace answers "who can fly this?" and the agent model answers "what does this quote really cost, and can we sign it safely?" Treated that way, the comparison of villiersOS and Avinode becomes a question of where in the workflow each belongs, not which one wins.

For the Farnborough to Nice return, the practical test is simple. Take the three quotes received, and ask how long it takes your desk to discover which of them is cheapest once the positioning leg, Nice handling and taxes are included. If the answer is "by the end of the afternoon", the work after the quote lands is where a broker's time is going, and that is the part a marketplace was never built to take on.