What “defensible” means here, concretely
Every figure in a Ricerca study is the end of a chain that starts in one of your own systems. That chain is the difference between the three approaches below, and it is the thing an examination actually tests.
- Source systems: Payroll and HRIS; Ledger and engineering tickets.
- Evidence: Time and cost records; Contemporaneous project records.
- Component narratives: One per business component; Four-part test applied to each.
- QRE workpapers: By category and by employee; Tied to the ledger accounts.
- Form 6765 lines: Section A or B, and Section G; Each line tied to a workpaper.
- Audit-defense binder: Indexed exhibits per component; Provenance for every figure.
- Every number traces back to a source row
What you are actually choosing between
Before the scorecard, the plain-language version, because the differences are structural, not stylistic.
Traditional consultants
Experienced people, manual process. Thorough when the engagement is well staffed - slow, costly, and opaque when it isn’t. You often never see the reasoning behind a number.
Automation-only platforms
A questionnaire, a classifier, and a PDF. Fast and cheap, and fine right up to the moment someone asks which alternatives your engineers actually evaluated.
Ricerca - expert-led, AI-assisted
The statute enforced in software, the study red-teamed before a human sees it, and experienced R&D experts making every qualification call that matters.
Traditional vs. automation-only vs. Ricerca
| Dimension | Traditional consultants | Automation-only platforms | Ricerca (expert-led + AI) |
|---|---|---|---|
| Turnaround speed | Slow - often months | Fast | Fast - typically weeks |
| Cost efficiency | Expensive | Low | Efficient |
| Expert review & finalization | Sometimes | None | Every R&D credit study, finalized |
| Audit-ready substantiation | Varies by firm | Often thin | Four-part-test mapped |
| Audit Protection | Often an add-on | None | Included |
| Data security | Varies | Varies | Encrypted, §7216-aware routing |
| §174A / OBBBA expertise | Sometimes | Rarely | Built in |
| Human support | High | Mostly self-serve | High - with R&D experts |
| Technology / AI assistance | Limited | Core | Core, with oversight |
| Transparency | Often opaque | Black box | Sourced & explainable |
Illustrative / for general comparison. This compares delivery approaches in general terms - not measured benchmarks, and not a statement about any specific provider. Individual results vary.
See the same comparison scored 0-5
View data
| Dimension | Traditional consultants | Automation-only platforms | Ricerca (expert-led + AI-assisted) |
|---|---|---|---|
| Turnaround speed | 2 | 5 | 4 |
| Cost efficiency | 2 | 5 | 4 |
| Expert review & finalization | 4 | 1 | 5 |
| Audit-ready substantiation | 4 | 2 | 5 |
| Audit Protection | 2 | 1 | 5 |
| Data security | 3 | 3 | 5 |
| §174A / OBBBA expertise | 3 | 2 | 5 |
| Human support | 4 | 1 | 5 |
| Technology / AI assistance | 1 | 5 | 5 |
| Transparency | 2 | 3 | 5 |
Defensible by construction, not promised by a salesperson
Ricerca enforces the statute in software, so every position is tested against the actual rules before a dollar is claimed, then finalized by experienced R&D experts.
§41(d)(4) exclusion screens
Research after commercial production, adaptation, duplication, surveys, routine data collection, internal-use software and more, caught up front rather than argued about later.
Funded and contract analysis
The funded-research and rights-and-risk tests applied to contract work, with the alternatives your team evaluated documented.
§41(f) controlled group
Related-entity aggregation determined and applied, so the credit is computed correctly across a controlled group rather than entity by entity.
Form 6765, Section G ready
Built for the Dec. 2024 revision of Form 6765, including business-component-level Section G detail.
Federal and state, together
The federal §41 credit and applicable state R&D credits computed from the same qualified base, the state benefit most studies leave behind.
Four-part test, enforced
Every business component is evaluated against all four parts of the IRC §41(d) test in software, not just tagged “qualified.”
- 01 Permitted purpose
- 02 Technological in nature
- 03 Elimination of uncertainty
- 04 Process of experimentation
None of this is a workflow preference. It is the difference between a number and a defensible computation, and it is why the platform refuses to finalize a study the substantiation doesn’t support.
The claims we won’t make, and why that’s the point
In a market full of promises, the useful signal is what a vendor refuses to say. Here is ours, in writing.
No borrowed proof
No client logos, no case studies, no star ratings, no "companies like yours" testimonials. If we can’t evidence it, it isn’t on this site.
No badges we haven’t earned
We don’t hold SOC 2 today, so you won’t find the seal here. What we do hold up is the control set itself - encryption, §7216 redaction before AI processing, enterprise-only model routing.
No promised outcome
Qualification is fact-specific. We tell you what your facts support, including when the answer is "less than you hoped" or "not this component."
No vague representation claim
Audit Protection means we prepare the examination response, the documentation, and the technical defense of our study, working alongside your tax preparer. Representation before the IRS, where it is needed, is performed by credentialed practitioners.
The full terms of what Audit Protection covers, and what it does not, are on the Audit Protection page.
Start where you actually are
Five situations, five honest next steps. None of them require talking to us first.
You’ve never claimed the credit
Size it before you commit to anything. Run the R&D tax credit calculator on last year’s R&D payroll, then let us pressure-test the assumptions: first-time claims are where the base-period and method choices matter most.
You have a study, but it’s thin
If your file is a spreadsheet and one narrative for the whole company, the number may be fine and the substantiation is not. Component-level evidence is what an examination is decided on.
You’re pre-revenue and owe no income tax
A credit you cannot use this year is not cash. A qualified small business may elect to apply the credit against employer payroll taxes instead: the path that turns the study into money now.
The §41(h) payroll offset Get your free R&D credit assessment
You own a building and have never had a study
If you bought, built or renovated a building you use or rent out, a study can move depreciation into the first years. A larger deduction is only worth something if it reduces tax, so your CPA confirms that before you commit.
Run the cost segregation calculator How cost segregation works
You bought the building years ago
The missed depreciation is usually still there. A look-back study is an automatic change in accounting method on Form 3115, taken as a single §481(a) adjustment in the year of change, without amending prior returns.
Form 3115 and look-back studies Get a free property assessment
Not sure which one you are? That is a normal answer: the free assessment sorts it out, or read what qualifies in your industry first. Comparing cost segregation providers? Five questions to put to any of them.
See the difference on your own study
Tell us about your R&D or your property. We come back with what a Ricerca study could capture and a written quote, free and with no obligation.