Smith-Atlas Federal Cost Allowability Assistant (SA-1008-V1) (Enterprise License)
Regular price
$400.00
⚡ Leverage AI to improve your operations. See the description below for more details.
Cost allowability guidance that walks the regulation with you — not a chatbot guessing from memory.
Ask whether a cost can be charged to your federal award, and get a structured answer grounded in the actual cost principles: the three-part test (allowable, allocable, reasonable), the precise provision and subsection, and the version date of the text it relied on. While it provides a recommended approach with backup, the assistant always encourages users to make the ultimate decisions and ask their sponsored programs or contracting office if they have additional questions.
Built by a research administrator and attorney for the people who field these questions every day: sponsored programs and research administration offices, departmental grant managers, contracting officers, and the PIs who ask them "can I charge this?"
What it does
- Runs the three-part test, every time. It separates allowability, allocability, and reasonableness instead of collapsing them, and names the specific test a cost fails (e.g., the allocability trap on an end of award equipment buy, for instance) with the pinpoint citation to match.
- Cites to the subsection, with a version date. Every recommendation carries the controlling provision down to the subsection (e.g., 2 CFR 200.405(a) or FAR 31.201-4(b)), the applicable Selected Item of Cost, and the effective date of the text it used.
- Routes you to the right cost allowability regulations automatically. It resolves Uniform Guidance vs. FAR cost principles from your organization type and award type, names the routing clause, if applicable, so you see why (FAR 31.303 / 31.603 / 31.703), and flags Cost Accounting Standards (CAS) when your configuration triggers it.
- Spots the issues users miss. Cost transfers, direct-vs-F&A classification, prior approval timing, the F&A/MTDC rebudgeting trap, participant support, pre-award and closeout timing, salary caps and Institutional Base Salary, change in scope: it screens every fact pattern against these and asks the targeted follow-up questions to get to the best recommendation.
- Tailors to your setup. Configure it once for your institution — and, optionally, for a single sponsor (NIH, NSF, DOE, and others) — and it raises only the cost allowability regulations that can actually apply to you and treats your award and sponsor's manual(s) as a first-check overlays.
What it does not do
- It is not your sponsored programs office, and its output is not a determination. Every analysis is AI-generated and advisory only. A qualified human makes the final call, and figures and citations should be independently verified before you rely on them.
- It does not authorize spend or bind your institution. It cannot approve a cost, submit a prior approval, change a ledger, or guarantee that a cost will pass an audit or clear central review.
- It is only as current as the texts you give it. It reasons from the regulations and sponsor files loaded into its Knowledge. Regulations and sponsor guidance change; an out-of-date upload produces an out-of-date answer, which is why it always states the version it used.
- It won't invent your institution's policy. Where a rule turns on your own written policy or cost accounting practice (travel, extra service pay, direct/indirect treatment), it marks the conclusion contingent and invites you to confirm, rather than guessing what your policy says. You can, however, upload policies where they're applicable, so the assistant has more information.
- It stays in its lane. It addresses federal cost allowability, allocability, and reasonableness only. Scientific merit, legal briefs, and non-federal accounting questions get redirected, not answered.
What you'll need
- A subscription to a capable AI assistant (Claude, ChatGPT, Gemini, Copilot, or an in-house solution) with a configurable assistant or project that supports a Knowledge/Context area for file uploads.
- The current regulation texts it cites against (Subpart E to 2 CFR Part 200 and, where applicable, FAR Subpart 31.2), plus any sponsor guidance or prior approval matrix you want it to apply, uploaded to the assistant's Knowledge under the exact filenames the setup guide specifies. Setup takes about 20 minutes and is covered step by step instructions in the included documentation.
What's included
- The Loader system instructions (identity, configuration, and session intake)
- The Process system instructions file (three-part test, issue-spotting checks, and output format)
- Setup and use instructions
License, coupon, updates, and refunds
- License. A perpetual, enterprise-wide license to this version (SA-1008-V1) for your internal business use, under the Terms of Service. One-time payment: no subscription, no renewal fee. You may adapt the Loader, Process, and documentation for your own AI tools and your own institutional configuration. You may not resell, sublicense, or distribute them outside your organization.
- Coupon. If your organization can certify that it is a small business under the applicable North American Industry Classification System (NAICS) code OR that it is a Carnegie Research 2: High Spending and Doctorate Production (R2) or smaller higher education institution, you can use the "Certified-Small-Organization" coupon code at checkout for a 40% discount.
- Updates. Any updates we release to this version within twelve months of your purchase are yours at no charge. After twelve months, the software keeps working; we simply won't be issuing further updates or support for it. Because the tool runs on third-party AI platforms that change over time, continued compatibility beyond that window is not guaranteed.
- Refunds. Full refund within 30 days of purchase, for any reason. Email matt@smith-atlas.com.
- Provided AS-IS. See the Terms of Service for the complete terms, including warranty disclaimers and limitation of liability.
- Public institutions. §9.1 and §12.1 of the Terms of Service address sovereign immunity and venue.
Regular price
$400.00
Cost allowability guidance that walks the regulation with you — not a chatbot guessing from memory.
Ask whether a cost can be charged to your federal award, and get a structured answer grounded in the actual cost principles: the three-part test (allowable, allocable, reasonable), the precise provision and subsection, and the version date of the text it relied on. While it provides a recommended approach with backup, the assistant always encourages users to make the ultimate decisions and ask their sponsored programs or contracting office if they have additional questions.
Built by a research administrator and attorney for the people who field these questions every day: sponsored programs and research administration offices, departmental grant managers, contracting officers, and the PIs who ask them "can I charge this?"
What it does
- Runs the three-part test, every time. It separates allowability, allocability, and reasonableness instead of collapsing them, and names the specific test a cost fails (e.g., the allocability trap on an end of award equipment buy, for instance) with the pinpoint citation to match.
- Cites to the subsection, with a version date. Every recommendation carries the controlling provision down to the subsection (e.g., 2 CFR 200.405(a) or FAR 31.201-4(b)), the applicable Selected Item of Cost, and the effective date of the text it used.
- Routes you to the right cost allowability regulations automatically. It resolves Uniform Guidance vs. FAR cost principles from your organization type and award type, names the routing clause, if applicable, so you see why (FAR 31.303 / 31.603 / 31.703), and flags Cost Accounting Standards (CAS) when your configuration triggers it.
- Spots the issues users miss. Cost transfers, direct-vs-F&A classification, prior approval timing, the F&A/MTDC rebudgeting trap, participant support, pre-award and closeout timing, salary caps and Institutional Base Salary, change in scope: it screens every fact pattern against these and asks the targeted follow-up questions to get to the best recommendation.
- Tailors to your setup. Configure it once for your institution — and, optionally, for a single sponsor (NIH, NSF, DOE, and others) — and it raises only the cost allowability regulations that can actually apply to you and treats your award and sponsor's manual(s) as a first-check overlays.
What it does not do
- It is not your sponsored programs office, and its output is not a determination. Every analysis is AI-generated and advisory only. A qualified human makes the final call, and figures and citations should be independently verified before you rely on them.
- It does not authorize spend or bind your institution. It cannot approve a cost, submit a prior approval, change a ledger, or guarantee that a cost will pass an audit or clear central review.
- It is only as current as the texts you give it. It reasons from the regulations and sponsor files loaded into its Knowledge. Regulations and sponsor guidance change; an out-of-date upload produces an out-of-date answer, which is why it always states the version it used.
- It won't invent your institution's policy. Where a rule turns on your own written policy or cost accounting practice (travel, extra service pay, direct/indirect treatment), it marks the conclusion contingent and invites you to confirm, rather than guessing what your policy says. You can, however, upload policies where they're applicable, so the assistant has more information.
- It stays in its lane. It addresses federal cost allowability, allocability, and reasonableness only. Scientific merit, legal briefs, and non-federal accounting questions get redirected, not answered.
What you'll need
- A subscription to a capable AI assistant (Claude, ChatGPT, Gemini, Copilot, or an in-house solution) with a configurable assistant or project that supports a Knowledge/Context area for file uploads.
- The current regulation texts it cites against (Subpart E to 2 CFR Part 200 and, where applicable, FAR Subpart 31.2), plus any sponsor guidance or prior approval matrix you want it to apply, uploaded to the assistant's Knowledge under the exact filenames the setup guide specifies. Setup takes about 20 minutes and is covered step by step instructions in the included documentation.
What's included
- The Loader system instructions (identity, configuration, and session intake)
- The Process system instructions file (three-part test, issue-spotting checks, and output format)
- Setup and use instructions
License, coupon, updates, and refunds
- License. A perpetual, enterprise-wide license to this version (SA-1008-V1) for your internal business use, under the Terms of Service. One-time payment: no subscription, no renewal fee. You may adapt the Loader, Process, and documentation for your own AI tools and your own institutional configuration. You may not resell, sublicense, or distribute them outside your organization.
- Coupon. If your organization can certify that it is a small business under the applicable North American Industry Classification System (NAICS) code OR that it is a Carnegie Research 2: High Spending and Doctorate Production (R2) or smaller higher education institution, you can use the "Certified-Small-Organization" coupon code at checkout for a 40% discount.
- Updates. Any updates we release to this version within twelve months of your purchase are yours at no charge. After twelve months, the software keeps working; we simply won't be issuing further updates or support for it. Because the tool runs on third-party AI platforms that change over time, continued compatibility beyond that window is not guaranteed.
- Refunds. Full refund within 30 days of purchase, for any reason. Email matt@smith-atlas.com.
- Provided AS-IS. See the Terms of Service for the complete terms, including warranty disclaimers and limitation of liability.
- Public institutions. §9.1 and §12.1 of the Terms of Service address sovereign immunity and venue.