Back office automation is sold three ways: hourly consulting, fixed price builds, and share of savings. As an illustrative 2025 to 2026 US market observation, hourly work tends to land around 100 to 300 dollars an hour, fixed price builds around 15,000 to 150,000 dollars per scope, and share of savings around 10 to 30 percent of documented savings. Those ranges move with specialty, payer mix, PHI scope, and payer portal count. The real question is not the number. It is who carries the risk if it does not work.
Last updated: July 2026
Anyone quoting you a single figure for “automating the back office” is either guessing or selling a product that will not fit. What is knowable is the shape of each pricing model, what it costs you when the project succeeds, and what it costs you when it does not. That second column is the one buyers skip.
What Are the Three Pricing Models?
Ranges in the table below are illustrative 2025 to 2026 US market observations, not survey data.
| Model | Illustrative range | You pay when | Vendor risk | Your risk |
|---|---|---|---|---|
| Hourly consulting | 100 to 300 dollars per hour | Continuously, regardless of result | None | All of it |
| Fixed price build | 15,000 to 150,000 dollars per scope | Milestones, usually 50 percent up front | Scope overrun only | Outcome risk |
| Share of savings | 10 to 30 percent of documented savings | Only after savings are measured | Full outcome risk | Measurement risk |
| Per seat SaaS platform | 50 to 500 dollars per user monthly | Every month, forever | None | Adoption and fit |
To repeat the caveat because it matters: these are general observations of 2025 to 2026 US mid-market pricing, not survey data, and they vary by specialty, payer mix, PHI scope, and how many payer portals or clearinghouse connections are involved. Treat them as a way to sanity check a quote, not as a benchmark to negotiate against.
Per seat SaaS platforms are included because many vendors package automation as a recurring subscription rather than a project. The monthly fee is predictable, but adoption risk is entirely yours: if staff do not use the platform, you pay the same.
How Does Hourly Consulting Actually Price Out?
You buy hours. A consultant or an agency scopes the work, bills time, and delivers whatever the hours produced.
The honest case for it: it is the right model for genuinely exploratory work, where nobody can define the deliverable yet, and for one off analysis. If you need someone to spend three weeks figuring out why your clean claim rate dropped, hourly is fine.
The problem is structural. The vendor’s revenue increases with the number of hours consumed, and yours decreases. That is not an accusation of bad faith, it is just how the incentive points. Complexity is rewarded. Finishing early is not. In a billing back office, where the work is genuinely intricate and the client usually cannot audit the technical decisions, that incentive gap is wide.
Watch for: no cap, no defined deliverable, and a discovery phase that keeps extending. Ask for a not to exceed number in writing.
What Does a Fixed Price Build Cost?
You agree on a scope, a price, and a delivery date. The vendor eats the overrun if the build takes longer than they estimated.
Rough shape, as an illustrative 2025 to 2026 US market observation rather than a quote. These move sharply with specialty, payer mix, PHI scope, and how many payer portals are in play:
- Single workflow automation. One thing, well defined. Automated eligibility checks, or claim status monitoring against one clearinghouse. Often in the 15,000 to 40,000 dollar range.
- Multi workflow project. Three to five connected automations across intake, eligibility, and denials. Often 40,000 to 100,000 dollars.
- Full back office program. Broad scope across the revenue cycle with multiple integrations and payer portal work. Frequently above 100,000 dollars and often phased.
Fixed price is a real improvement over hourly, because it moves timeline risk to the vendor. But notice what it does not move: outcome risk. The vendor is paid for delivering the thing, not for the thing working. If the automation is delivered exactly as specified and it recovers nothing because the scope was wrong, you paid in full. Fixed price transfers the risk of a late build. It does not transfer the risk of a useless one.
The other failure mode is scope rigidity. Six weeks in, you learn the real leak is somewhere else. Under a fixed price contract, that discovery becomes a change order.
How Does Share of Savings Work, and Where Does It Break?
The vendor is paid a percentage of savings that are documented after the fact. No savings, no fee. Percentages observed in the 2025 to 2026 US market generally run 10 to 30 percent, though that is an observation rather than a published figure.
This is the model where the vendor carries the outcome risk, which is the right place for it if the vendor is the one claiming to know what will work. It is also the model with the most ways to go wrong, and you should interrogate it harder than the other two, not less.
What to pin down before you sign:
- How is savings defined? Recovered revenue, reduced labor hours, avoided hires, reduced write offs, or some combination. Vague definitions here are the single biggest problem with this model.
- What is the baseline, and who measures it? If the baseline is set by the vendor after the work starts, it is not a baseline.
- How long does the fee run? A percentage that continues indefinitely on savings from a build finished two years ago is a subscription wearing a costume. A defined term, often 12 to 24 months, is more reasonable.
- What happens if savings are disputed? There should be a documented measurement method both sides agreed to in advance.
- Are soft savings counted? “Time saved” that does not translate into a reduced headcount, reduced overtime, or increased throughput is not a saving you can spend. Be careful about paying real dollars for hypothetical hours.
To be concrete about our own version, since it is easier to evaluate a model when someone shows their actual terms: XClear AI runs a free 30 minute Waste Audit, documents where the waste is, and takes 15 percent of documented savings. The client keeps 85 percent. If there are no savings, there is no fee. That is one implementation of the third model, not the only defensible one, and the questions above apply to us the same as to anyone else. Ask them.
A Savings Formula Worth Having
Before you sign anything, agree on how savings will be calculated. A reusable formula that covers the main categories:
Documented savings = (baseline write-offs minus post write-offs) + (hours eliminated x loaded hourly cost) minus (new software and maintenance costs)
What this formula excludes on purpose: soft “time saved” where no headcount changed and no additional throughput was produced. If the same staff run the same volume and nobody was redeployed or reduced, the hours saved are real but the dollar value is not yet realized. Count them when the business acts on them, not before.
Illustrative Worked Example
The following is explicitly illustrative. Numbers are round figures chosen to show the method. They are not quotes, survey data, or claims about any specific practice.
Scope: Eligibility verification (270/271 transactions), claim status monitoring (276/277 transactions), and denial detection from 835 remittances for a five-provider practice running athenaCollector as the PM system with Availity as the primary clearinghouse.
Assumptions:
- Region: US mid-market
- Specialty: primary care
- PHI scope: eligibility and claim status data only, covered under BAA
- Payer portal count: three portals plus one clearinghouse connection
- Baseline labor: front desk spending estimated time weekly on manual eligibility lookups and claim status checks
| Model | Year 1 | Year 2 | Two year total |
|---|---|---|---|
| Hourly consulting, 200 per hour for 300 hours, plus 60 hours of year two upkeep | 60,000 | 12,000 | 72,000 |
| Fixed price build of 40,000, plus 4,800 software and transaction costs and 3,000 maintenance annually | 47,800 | 7,800 | 55,600 |
| Share of savings at 15 percent of 80,000 documented savings | 12,000 | 12,000 | 24,000 |
Every figure above is illustrative and chosen to show the method. The hourly row assumes the build is delivered in 300 hours and then needs ongoing time in year two, which hourly engagements usually do. The fixed price row carries maintenance and software forward because portals change and transactions cost money. The share of savings row assumes 80,000 dollars in documented savings each year; if savings are lower, the fee is lower, and if there are no savings, there is no fee.
Note what the two year column hides. Share of savings looks cheapest here only because the savings assumption is fixed. If this scope produced 200,000 dollars in documented savings, the same 15 percent would cost 30,000 dollars a year, and the fixed price build would be the cheaper deal. That is the actual trade: fixed price caps your cost and keeps your outcome risk, share of savings caps your risk and leaves your cost open ended. Run the arithmetic against a savings figure you believe rather than one a vendor supplies.
What Costs Do All Three Models Leave Out?
The quote is not the total. Budget for these regardless of model:
- Your team’s time. Somebody internal has to answer questions, grant access, validate outputs, and make decisions. Plan on real hours from a person who knows your billing operation, not an intern.
- Software and API costs. Clearinghouse transaction fees, LLM API usage, RPA licensing, and any middleware. These are usually small next to labor but they recur.
- Maintenance. Payer portals change. Rules change. Anything touching a portal needs ongoing upkeep, and a proposal that shows zero maintenance line is incomplete.
- Compliance review. Anything touching PHI needs a BAA and a security review. If your vendor did not raise HIPAA in the first conversation, that is a signal.
- Change management. The automation that nobody uses costs the same as the one everybody uses.
Switching Costs and Exit Planning
Ask these three questions before you sign, not after.
- Who owns the data? Your patient and claims data should be exportable in a standard format at any time. Confirm this in the contract.
- Who hosts the automations? Bots and workflows running inside vendor-controlled infrastructure may not be portable. Understand what you keep if the relationship ends.
- What survives vendor exit? If the vendor disappears or you terminate, is there a documented handoff, are credentials and credentials returned, and can your team operate without them for a transition period?
Where This Does Not Work
Share of savings does not work when savings genuinely cannot be measured. If your practice has no reliable baseline data, no clean reporting, and no way to isolate the effect of the automation, the model produces an argument instead of an invoice. In that situation a small fixed price engagement to establish measurement is more honest.
None of these models work well below a certain volume. A two provider practice with a part time biller usually does not have enough repetitive volume to pay back any of them. The right answer there is frequently better use of the practice management system you already own, not a build.
And no pricing model saves a project where the underlying process is broken. If your intake data is unreliable, automating downstream just moves the error faster. Fix the process, then automate what remains.
How Should You Compare Two Quotes?
Ignore the headline number and compare on four things:
- What happens if it does not work? Who is out money.
- What exactly is the deliverable? If you cannot state it in one sentence, the scope is not defined.
- What is the measurement method? Written down, agreed before the work starts.
- What is the exit? How you stop, what you keep, and whether the automation survives the vendor relationship ending.
That last one matters more than people expect. Ask whether the workflows live in systems you control.
You can see how this gets scoped for revenue cycle work on our medical billing page, and the general approach is under solutions. Nothing here requires replacing your practice management system or clearinghouse. Automation should sit on top of what you already run.
FAQ
How much does it cost to automate medical billing? It depends on the model. As an illustrative 2025 to 2026 US market observation, hourly consulting runs around 100 to 300 dollars an hour, fixed price builds around 15,000 to 150,000 dollars depending on scope, and share of savings around 10 to 30 percent of documented savings, costing nothing if savings do not appear. These vary by specialty, payer mix, PHI scope, and payer portal count, so use them to sanity check a quote rather than as a benchmark.
Is share of savings better than a fixed price? It puts outcome risk on the vendor, which is usually the right place for it. It is only better if savings can be measured credibly against a baseline set before the work begins. If measurement is fuzzy, a fixed price scope is cleaner.
What is a reasonable percentage in a share of savings deal? Percentages observed in the 2025 to 2026 US market generally fall between 10 and 30 percent, as a rough observation rather than a published figure. The percentage matters less than the definition of savings, the baseline, and the term length. A low percentage on a vague definition can cost more than a higher percentage on a tight one.
Are there ongoing costs after the build? Yes. Expect transaction or API fees, and maintenance for anything touching payer portals, since portals change. A proposal with no maintenance line is incomplete.
Can a small practice afford this? Sometimes not, and that is a legitimate answer. Below a certain claim volume the repetitive work is not large enough to pay back a build. Better use of your existing practice management system is often the cheaper fix.
If you want a specific number for your operation rather than a range, book the free 30 minute Waste Audit. If the honest finding is that there is not enough waste to justify a project, that is the finding you will get.