The F&I Office Is the Bottleneck: Rebuilding the Last Hour of Car Buying
What the last hour of a car purchase is actually for, and what F&I process improvement looks like when the buyer arrives already verified.
Alfred BEditorial Reviews
Buyers rate the car well and the chair badly. The car gets chosen over weeks. The chair arrives at the end of a long afternoon, after the decision is made and the energy is gone. Most attempts to fix that hour aim at the clock.
F&I process improvement works better as a redesign than as a speed-up. When identity, income and consent are verified before arrival, the last hour stops being a collection exercise and becomes two things that need a person in the room: judgment about which products fit a specific household, and making sure the buyer understands the obligation they signed.
What does the published evidence say about the last hour of a car purchase?
Published research separates the wait from the meeting. CDK Global's 2026 F&I Buyer Experience study, published 20 February 2026, found 79% of buyers were satisfied with the time spent with the F&I manager, while 41% waited more than 30 minutes to reach one. The queue is the disliked part.
The same CDK Global study, which the dealership software vendor publishes as market research, recorded Net Promoter Score falling from 28 to 8 once the wait reached 30 to 45 minutes, and found 58% of buyers trusted the F&I manager more than the salesperson for product recommendations. Those findings describe different rooms. One measures the corridor, the other the desk.
Sentiment is improving overall. The 2025 Cox Automotive Car Buyer Journey Study, released 13 January 2026 from about 2,300 buyers, found 71% highly satisfied and 63% preferring a mix of online and in-person steps. Cox Automotive sells retailing software to dealerships.
Satisfaction with the financing splits by the buyer's own financial position. J.D. Power's 2025 U.S. Automotive Financing Satisfaction Study, published 13 November 2025 from 13,150 financed customers, scored financially healthy borrowers at 743 on a 1,000-point scale against 593 for financially vulnerable ones, and classified 29% of auto finance customers as financially vulnerable. The people who most need the hour to work well rate it worst. The itemised version of the hour is set out in the anatomy of a floor-time collapse.
What does F&I process improvement look like when the file arrives verified?
A pre-verified deal changes what the desk opens with. Instead of collecting identity, income and consent, the finance manager confirms a file that is already decided, resolves whatever the lender still conditions, presents products, and signs. The hour becomes a review of a known state rather than the construction of one.
Three properties define the rebuilt version. It starts at a known point, so the desk knows before the buyer sits down which lender holds the file, at what tier, and with what conditions. It is interruptible, so a buyer can go home and come back without losing their place. It carries slack on purpose, because a schedule with no gap turns every small surprise into a queue.
Where the file starts is its own design decision, covered in starting the application on the buyer's phone. Whether released minutes turn into money is covered in what saved floor time is worth, and the answer there is usually no, which is a reason to redesign the hour rather than shorten it.
What has to be true operationally for pre-verification to work?
Six things. Identity proofed remotely to a standard a lender accepts, income confirmed at source rather than by document, consent captured once and carried forward, a freshness stamp on every verified fact, a lender able to contract electronically, and an attended path at the desk that runs just as fast for buyers who do not finish on a phone.
Remote identity proofing has a published specification. NIST Special Publication 800-63A revision 4, published 1 August 2025, sets out what an unattended remote proofing process does: collect evidence of fair or strong strength, validate it by automated document validation, validate core attributes against authoritative or credible sources, and verify that the person presenting the evidence is the person it describes. Knowledge-based verification, the security-question pattern, is excluded. Any process calling a vehicle applicant verified is claiming some version of that list.
Income is the harder half. A pay stub photographed on a phone is a picture of an assertion, and confirming employment at source turns it into a fact. Freshness is the requirement nobody writes down alongside it. A lender accepting proof of income from eleven days ago may refuse the same proof at forty, so a verified state with no timestamp is a claim about the past presented as one about the present.
The lender's side has been moving for years. Wolters Kluwer's Automotive Finance Digital Transformation Index for the first quarter of 2026, reported 18 June 2026, put eContracting adoption up more than 61% since the first quarter of 2022. That is vendor research, so read the trend rather than the decimal.
What breaks when pre-verification is only partly true?
Partial verification costs more than none, because the desk has to work out which parts are done. Five failures recur: buyers who cannot finish remote proofing, verifications that have gone stale, income left unverified while identity is complete, co-applicants who never started, and lenders that re-request the artifact anyway.
Coverage is the first failure, and it is measurable in other industries. The U.S. Government Accountability Office, reporting on Login.gov on 16 October 2024, noted the Small Business Administration's report of a 30% to 40% failure rate during account creation. A remote path that works for most people still leaves a group at the door, and a store without an equally quick attended path sorts its customers by how well their phone cooperates.
Scope is the second. A file with identity proved and income unproved looks finished from the front and reopens when the lender's conditions come back, except nobody was expecting it. Fraud pressure follows the same seam, and Point Predictive, a fraud analytics vendor, attributed 45% of a record $10.4 billion in annual auto lending fraud exposure to income misrepresentation on 8 April 2026.
The expensive failure is the lender that re-requests the document regardless. A buyer who verifies income on a phone on Thursday and is asked for a pay stub on Saturday learns the first step was decorative, and the store loses their willingness to repeat it.
The number worth tracking is the share of deals arriving fully verified, counted whole file or not at all. Average completion hides the files that break, because a file at 80% is a file with a hole in it and no label saying where.
What is the F&I hour actually for once paperwork is not the constraint?
The hour does four jobs at once and only two of them are administrative. Collection and verification can move off the desk. The lender's decision mostly moves with it. What cannot move is judgment about which products suit a particular household, and making sure a buyer understands a six-year obligation before signing it.
Table 1: the four jobs the F&I hour performs simultaneously, and what pre-verification changes about each. Rows one and two can largely move off the desk. Rows three and four cannot, and run on whatever time the first two leave behind.
| # | Job | Whose interest it serves | What pre-verification changes | What it needs instead |
|---|---|---|---|---|
| 1 | Collection and verification | The lender's file | Most of it leaves the desk | A verified state that arrives with the buyer, timestamped |
| 2 | Decision and conditioning | The lender's risk | The approval moves earlier; conditions still land in the room | Conditions that name the field they doubt |
| 3 | Product presentation | The dealership's margin | Nothing, except more room to do it | Sequence and timing chosen deliberately |
| 4 | Comprehension and commitment | The buyer | Nothing, except more room to do it | Time that is scheduled rather than left over |
Rows three and four answer the heading. Row three is a real business with real margin: Penske Automotive Group reported finance and insurance gross profit of $1,807 per retail unit on 125,401 vehicles delivered, in second-quarter results published 29 July 2026. Audited money, attached to a conversation held in a room with a person in it.
Row four is the one nobody schedules. Experian's State of the Automotive Finance Market report for the first quarter of 2026, published 28 May 2026, put the average new-vehicle loan term at 69.48 months, with an average payment of $770, with 35.55% of new-vehicle loans running longer than six years. Nearly six years of obligation, originated at the end of an afternoon, in the residue of three other jobs.
Set that against the J.D. Power gap between vulnerable and healthy borrowers and the design problem states itself. The hour is the only point in the transaction where a person can check whether the buyer has followed what is happening. Treating it purely as throughput optimises away the one job with no other home.
How does presentation change in an hour that is no longer a race?
Presentation becomes a design question about sequence and timing rather than a compression exercise. Field experiments show the order in which options appear changes how often buyers accept a default. A menu opened at minute ten, to a buyer who previewed it at home, is a different artifact from the same menu at minute fifty-five.
The order effect has been measured on car buyers directly. Jonathan Levav, Mark Heitmann, Andreas Herrmann and Sheena Iyengar, in field experiments published in the Journal of Political Economy in April 2010, varied the order of attributes during customisation of suits and automobiles, and found order affected both the design and the final price of the car people bought, by changing how likely they were to accept the firm's suggested default.
Trade evidence points the same way. CDK Global's February 2026 study found F&I managers increased attempts to present three or four products from 23% to 27%, with the closing rate on multiple products doubling from 10% to 20%, while most buyers still took one or two items. It also recorded more presentations built from a buyer's questionnaire answers rather than a fixed script, which is a sequencing decision.
Two design choices follow, and neither touches what is presented. Products can be previewed before arrival, so the room is spent on questions rather than first exposure, and comprehension can sit apart from price.
The strongest argument for keeping the menu in the room is the CDK Global finding that 58% of buyers trust the F&I manager's product recommendation more than the salesperson's. Trust built by a format does not transfer to a screen.
What does the Canadian version of this look like?
Canada publishes the role and not the clock. CADA's 2024 Data Report profiles 3,783 franchised dealerships and records average F&I manager compensation of $170,052 in 2023 against a median of $157,410, which describes a retained specialist function rather than an administrative one. No Canadian source publishes finance-office timing.
The obligation signed in Canada is comparable in size. The same CADA report put the average amount financed on a new vehicle at $55,703 in the fourth quarter of 2024, at an average rate of 5.05%.
What Canada measures well is the seam between dealer and lender. J.D. Power's 2026 Canada Dealer Financing Satisfaction Study, published 12 May 2026 from 6,953 evaluations, found 65% of dealers expect a lender's funding staff to respond within 30 minutes. The seam has its own article: the dealer-lender handoff and the journey around it.
What we couldn't verify
No published source gives the share of vehicle deals arriving at a finance desk with identity and income already verified, in either country. We looked at Cox Automotive, CDK Global, J.D. Power, NADA, CADA, DesRosiers, Experian and the filings of five public dealer groups. The number this design question turns on is not public.
Remote identity proofing pass rates for vehicle retail are unpublished too. The Login.gov figure above comes from federal identity services and a different population, so it shows a coverage gap exists without sizing the one in a showroom. The Levav, Heitmann, Herrmann and Iyengar effect sizes sit behind a journal paywall and are described here qualitatively, and their automobile experiments covered vehicle configuration rather than finance products.
Stipulation data is missing everywhere. Nobody publishes conditional approvals as a share of approvals, documents re-requested per deal, or how often a verified fact is re-requested by the lender who received it. The Wolters Kluwer adoption index is vendor research whose methodology is not public. NADA's Data 2025 full-year report counts 16,990 franchised light-vehicle dealers selling 16.2 million light-duty vehicles, and publishes no measure of time inside the finance office.
Common questions
What is F&I process improvement in a car dealership?
F&I process improvement is the redesign of the finance and insurance stage of a vehicle purchase so collection and verification happen before the buyer sits down. The change is in what the desk opens with, rather than in how fast the same sequence runs.
Do car buyers dislike the F&I office itself?
Mostly they dislike the wait. CDK Global's 2026 F&I Buyer Experience study, published 20 February 2026, found 79% of buyers satisfied with time spent with the F&I manager while 41% waited more than 30 minutes to reach one, and Net Promoter Score fell from 28 to 8 at waits of 30 to 45 minutes.
What has to be verified before a buyer arrives for pre-verification to work?
Identity, income and consent, each carrying a timestamp. NIST Special Publication 800-63A revision 4, published 1 August 2025, sets out what unattended remote identity proofing requires: fair or strong evidence, automated document validation, attribute validation against authoritative sources, and biometric or code-based verification.
What goes wrong when only part of the file is verified in advance?
The desk cannot tell which parts are done. Common failures are buyers who cannot complete remote proofing, stale verifications, income left unverified while identity is complete, absent co-applicants, and lenders re-requesting documents already verified, which teaches buyers the first step was decorative.
Is there Canadian data on the dealership finance office?
Very little. CADA's 2024 Data Report profiles 3,783 franchised dealerships and records average F&I manager compensation of $170,052 in 2023. DesRosiers Automotive Consultants reported 500 average units per dealer franchise for 2025 on 18 February 2026. No Canadian source publishes finance-office timing.
Related reading: the anatomy of the F&I hour, what saved floor time is worth, starting the application on the buyer's phone, the auto finance journey map.
Carousel builds the verification layer that lets a file arrive at the desk already finished. See the auto intake suite
