Founders ask Collective 54 this 6 times in our records, 3 of them in 2026. It arrives as a tool question and the useful answer is a sourcing question.
In a boutique the economic engine is billable talent, and anything that does not directly sell or deliver the work is overhead. Finance is squarely overhead, and the published position is that this is correct rather than a flaw: all full-time employees should be billable, overhead should be kept deliberately lean, and non-billable functions should be fractionalized and outsourced wherever possible. Finance is one of four functions, alongside IT, HR and legal, that the most disciplined firms outsource entirely.
That settles the software question before it is asked. A firm that has outsourced finance does not pick the ledger, the invoicing tool or the bill-pay platform any more than it picks the tax software its accountant runs. The provider brings the stack, integrates it, and owns its operation. The build-versus-buy answer on this site makes the same point from the other side: software you run the firm on should be bought and its operation outsourced, and the one thing you must not outsource is the thinking.
The finance material adds why founders end up asking anyway. They do not wake up wanting a bookkeeper, a controller and a chief financial officer. They want finance to work: payroll to run, invoices to go out, cash to come in, taxes to be prepared, reports to arrive on time. They buy finance as a total service, which is rational, because a firm of five to fifty million dollars does not benefit from disaggregating finance into component parts. The mistake was never outsourcing it. The mistake was who delivered it and in what era.
Any modern cloud platform will keep a clean ledger, send an invoice, pay a vendor and reconcile a bank feed. The published account of the last era is that software made finance faster and cheaper, closes happened in days instead of weeks, dashboards appeared, and quality did not improve at all. Finance still reported what already happened. It still lacked benchmarking. It still failed to turn numbers into decisions. Dashboards multiplied and the presence of data was mistaken for understanding.
So the software decides speed and cost, which were solved a decade ago, and it decides nothing about the part that matters, which is whether anyone tells you your margin is leaking before the quarter closes. A founder comparing two accounting platforms is comparing two things that will produce the same reports with the same silence about what to do next.
The question is which provider, and the published material is specific about the test. The old fractional finance model, built on human labor and generalist providers who served restaurants, contractors and boutiques with the same templates, was slow, expensive and strategically useless, and it left founders financially illiterate about their own economics: unclear how gross margin is calculated in a services firm, what belongs in overhead versus sales, and how to reverse-engineer the profit and loss statement to improve margins on purpose. Traditional fractional chief financial officer models are called structurally obsolete unless paired with AI.
What replaces them is a provider that has installed the AI Finance Manager, which is described as a capability rather than a person. The AI owns roughly eighty percent of the function: data ingestion and reconciliation, continuous reporting, variance detection, forward-looking projections and pattern recognition across financial, operational and commercial data. The human layer, which should be outsourced and fractional rather than internal, owns the twenty percent: judgment, interpretation, benchmarking, strategic tradeoffs and accountability. And that human layer must specialize in boutique professional services, because without deep vertical specialization the human contribution collapses back into compliance and commentary.
The workflow inverts. In the old model, data went out to the provider, was processed on their timeline and came back as reports weeks after the decisions were made. In the new one, the AI lives inside the business, ingesting sales, delivery, payroll and cash continuously and producing decision-ready output in real time, and the fractional firm quality-controls that output, applies benchmarks from comparable firms and adds judgment where tradeoffs matter.
The requirements below are drawn from the finance material; the framing of them as a checklist is an inference.
Specialization in firms like yours. Ask how many boutique professional services firms in your revenue band they serve, because benchmarking is impossible without enough of them. The published examples of what founders believe when nobody benchmarks them are precise: that being paid in 45 days is acceptable when peers are paid in advance and run negative working capital, that a two-week close is efficient when it should take 24 hours, that a 50 percent gross margin is strong and a 25 percent EBITDA margin is impressive when both leave money on the table.
AI doing the mechanical work. Ask what share of the reconciliation, reporting and variance detection runs without a person, and how quickly the books close. A provider still staffing that work with contractors is passing an inefficiency on to you.
Judgment rather than reports. Ask what they told their last client to change, and whether it was an accounting answer or an economic one. The test is whether they can say what is driving your margin up or down and what to do about it.
Activity-based costing. The material describes finance as a force multiplier once every hour carries a fully burdened cost: an analyst spending 25 hours on a task becomes a 2,500 dollar delivery cost, and that translation turns costing into an operating decision about automating, shifting to AI, moving offshore or keeping senior. Ask whether the provider can produce cost per project, because the project is the unit of profit in a boutique.
Whatever platform sits underneath, the outputs the published material treats as non-negotiable are few. Gross margin calculated correctly for a services firm, with the benchmarks of 75 percent gross margin and 40 percent EBITDA that leave roughly 35 points of revenue for overhead and sales together. Cash flow per project, built from fee, hours per staff member, fully loaded cost and allocated overhead, which is the diligence formula Capital 54 uses and the one that reveals whether the delivery model is standardized. Collections against fee terms, because a boutique paid in advance has high fee quality and one with aging receivables does not. And a forward view, since cash flow problems hide wherever forward visibility is missing.
If your current setup produces those four and someone reads them with you monthly, the software is fine. If it does not, changing the software will not fix it.
Collective 54 publishes no vendor recommendation for accounting, invoicing or payables software, no integration guidance, no chart of accounts template and no provider list. The position published is about the operating model of the finance function, and it is that a boutique should not be running finance internally at all.
If the firm is below the five million dollar band and the founder is still doing the books personally, any reputable cloud platform with a bookkeeper behind it is adequate for now, and the provider question becomes urgent at the point the founder stops understanding the numbers, which usually arrives with the first real hiring wave.
If the firm already has an internal finance person, the material would rather that person were billable or that the role were fractionalized, but the software they run is not the problem and replacing it will not change the economics.
And if a buyer or lender is about to look at the books, the priority is not the platform but a coherent financial story, because firms are discounted at sale for incoherent financials far more often than for the software that produced them.
You should not be choosing it. Finance is overhead by design and should be fractionalized and outsourced, so the software belongs to the provider and the decision that belongs to you is which provider. Any modern cloud platform keeps the ledger, sends invoices and pays vendors; the last era made finance faster and cheaper without making it better, and no platform will tell you your margin is leaking. Require a provider that specializes in boutique professional services firms, runs reconciliation, reporting and variance detection on AI so the books close in a day, benchmarks you against peers who are paid in advance and run 75 percent gross margins and 40 percent EBITDA, and returns judgment rather than reports. Whatever sits underneath must produce correct gross margin, cash flow per project, collections against fee terms and a forward view. Collective 54 publishes no vendor recommendation, and the position that matters is that a boutique should not be running finance internally at all.
No. Finance is overhead by design, and the published position is that all full-time employees should be billable and non-billable functions should be fractionalized and outsourced. Finance is one of four functions, with IT, HR and legal, that disciplined firms outsource entirely. A full-time finance department rarely makes economic sense at five to fifty million dollars: the volume does not justify the cost and the opportunity cost of diverting dollars from billable talent is too high.
Require specialization in boutique professional services firms, because generalists cannot benchmark you and their judgment collapses into compliance. Require AI running the mechanical work so the books close in a day rather than two weeks. Require judgment rather than reports: ask what they told their last client to change. And require costing per project, because the project is the unit of profit and every hour should carry a fully burdened cost.
The published benchmark is that a close should take about 24 hours, and that founders who believe a two-week close is efficient are flying blind because no one has shown them a comparison. The same list of unbenchmarked beliefs includes being paid in 45 days when peers are paid in advance, and treating a 50 percent gross margin or a 25 percent EBITDA margin as strong when the published benchmarks are 75 and 40.
Not the software, the story. Firms are sold at discounts and on unfavorable terms when their financial story is incoherent, not because the platform was wrong. What a buyer wants is correct gross margin, cash flow per project that shows a standardized delivery model, collections that show high fee quality, and a forward view. A provider that specializes in firms like yours produces that; the ledger underneath is interchangeable.
Sources: Greg Alexander, The AI-Native Boutique Firm (Advantage Books, January 2027), specifically The AI Finance Manager for the position that finance is overhead by design, that all full-time employees should be billable, that non-billable functions should be fractionalized and outsourced and that finance is one of four functions with IT, HR and legal that disciplined firms outsource entirely; for the finding that founders buy finance as a total service rather than as a bookkeeper, controller and chief financial officer; for the account of the first era as slow, expensive and strategically useless and of generalist providers who left founders financially illiterate about gross margin, overhead and EBITDA; for the second era making finance faster and cheaper without making it better; for the unbenchmarked beliefs about 45-day payment, two-week closes, 50 percent gross margin and 25 percent EBITDA; for the AI Finance Manager as a capability owning data ingestion, reconciliation, continuous reporting, variance detection, projections and pattern recognition, with an outsourced and specialized human layer owning judgment, interpretation, benchmarking, tradeoffs and accountability; for the inverted workflow in which AI works inside the firm and humans add value at the edge; for the statement that traditional fractional chief financial officer models are obsolete unless paired with AI; for activity-based costing and the 25-hour, 2,500 dollar translation; and for the finding that firms are valued at discounts when their financial story is incoherent. Related Collective 54 answers on this site: should we build AI tools ourselves or buy existing software; what should I count as overhead; how do I calculate the true cost of delivering a service; how do I make sure I always have enough cash on hand. Greg Alexander, The Boutique: How to Start, Scale, and Sell a Professional Services Firm (Advantage, 2020), chapter 12 for cash flow per project built from fee, hours per staff member, fully loaded cost and allocated overhead as the Capital 54 diligence method, and for cash flow problems hiding where forward visibility is missing; chapter 30 for the 75 percent gross margin and 40 percent EBITDA benchmarks; chapter 32 for fee quality including payment in advance and aging receivables. Note on scope: Collective 54 publishes no accounting, invoicing or payables software recommendation, integration guidance, chart of accounts or provider list. The provider checklist and the list of four required outputs are inferences used here to organize the source material rather than published Collective 54 positions.
Collective 54 is the private community for founders and executives of boutique professional services firms between $5M and $50M in revenue. Members work these answers against their own numbers.