Artifact · A2 · 2026-05-28

Derive / extend / complete mapping.

3 products x 3 statutory prongs = 9 cells. Each cell carries the connecting argument, legal-test fit, confidence tag, and a one-liner in the CO's own register.

A2 · Product-to-Prong Grid · artifact id: derive-extend-complete-mapping-v0.html · 2026-05-28 · v0

What this mapping does

15 U.S.C. 638(e)(6)(C) defines Phase III as work that "derives from, extends, or completes efforts made under prior funding agreements under the STTR program." A contracting officer writing a Phase III action memo must satisfy at least one of the three prongs for each product scope being procured. This grid maps all three prongs for each of NorthAI's three commercial products against the Phase I verbatim scope.

The predicate award is FA8649-21-P-0756 (AFWERX STTR Phase I, CHN Analytics LLC, 2021-02-19 to 2021-05-18, $49,500). KNOWN The Phase I scope verbatim from USASpending:

"Research in attempt to discover a new method for identifying commercialization partners for government-funded research output using multiple, connected big data sets (patents, grants, financial) powered by artificial intelligence."

Four elements are in scope by explicit text: (1) identifying commercialization partners, (2) for government-funded research output, (3) using patents, grants, and financial data as connected big data sets, (4) powered by AI. Every cell below is evaluated against these four anchors.

Two conditional flags

Flag 1 · "Financial" data semantic ambiguity The Phase I scope lists "patents, grants, financial" as the connected data sets. The word "financial" is unresolved: it may refer to commercial investment data (VC rounds, funding rounds, startup capital flows) or to government budget and appropriations data (DoD RDT&E program elements, President's Budget submissions). Both readings are linguistically consistent with the USASpending description. If "financial" meant commercial investment data, Defense BD's structured RDT&E dashboard is an extension from a related but distinct data dimension. If "financial" meant government budget data, Defense BD's extension argument is a direct one-for-one match. Tim's Phase I technical report is the primary source that would resolve this. TO-CONFIRM The Defense BD EXTENDS cell below is rated as such either way -- the extends prong is broad enough to cover both readings -- but the fit label changes from moderate to strong once Tim's report is reviewed.

Flag 2 · NorthAI / CHN Analytics LLC legal continuity The predicate award FA8649-21-P-0756 was issued to CHN Analytics LLC. The domain chnanalytics.com currently 301-redirects to northai.io, the expected signal for a rebrand. The statute requires that Phase III issue to "the SBIR and STTR award recipients that developed the technology" -- the awardee entity must be CHN, its renamed successor, or a formally novated assignee. INFERRED If CHN Analytics LLC was renamed NorthAI (or merged into an entity now operating under the NorthAI brand), the award travels with the entity and no additional action is required. If NorthAI is a separately incorporated entity that acquired CHN's assets, a novation of FA8649-21-P-0756 must be executed before Phase III can issue. This flag appears in every cell below and is a gating condition for the CO action memo (A3), not a bar to the derives-from thesis. Tim needs to confirm which structure applies. The 301-redirect and the 5.5-year OSI&A relationship continuity both suggest rebrand rather than asset acquisition, but legal confirmation is required before the CO signs.

The 3 x 3 grid

NorthStar · Tech Vector · Defense BD across Derives / Extends / Completes

Product	Derives from	Extends	Completes
Tech Vector Guided query over 200M-document corpus; returns technology-landscape analysis across patents, grants, publications, and DoD spending data	Strong KNOWN Tech Vector is the productized commercial execution of exactly what the Phase I was funded to discover. The Phase I scope verbatim -- "discover a new method for identifying commercialization partners for government-funded research output using multiple, connected big data sets (patents, grants, financial) powered by artificial intelligence" -- maps directly and textually to Tech Vector's function: a guided query over connected big data (patents, grants, legal filings, DoD spending) powered by AI, returning a landscape of who is doing what in a given technology domain. The data sources in the Phase I description (patents, grants, financial) are the same data sources Tech Vector accesses. The goal of the Phase I (identify who is doing relevant research across government-funded output) is the same goal Tech Vector delivers to its defense buyer. Both the Phase I verbatim and the Tech Vector product description are primary public sources; the textual match is direct, not inferred. Tech Vector operationalizes the AI-over-connected-big-data method developed under FA8649-21-P-0756; the Phase I verbatim scope and the product corpus match on data sources (patents, grants, financial) and on function (identifying commercialization partners for government research output).	Moderate INFERRED The Phase I was a 3-month customer-discovery and feasibility confirmation exercise -- not a full commercial product. The gap between a feasibility-demonstrated AI method and a commercial guided-query interface with a 200M-document corpus, 15,000+ RDT&E project integrations, and a real-time query architecture represents genuine extension: Tech Vector applies the Phase I method to a broader and deeper commercial-scale data estate than any 3-month feasibility study could have built. The product scope -- global coverage, defense industrial base buyers, commercial subscriptions -- extends the original dual-use dual-purpose framing into a fully commercial product with an independently monetized user base. This extends prong is secondary to the DERIVES cell (which is the stronger argument), but it is available and honest as a companion framing if the CO wants to acknowledge both the lineage and the commercial-scale expansion. Tech Vector extends the Phase I feasibility method to a commercial-scale corpus and a broader buyer set (defense industrials, investment analysts, tech-transfer offices) not served by a 3-month STTR deliverable.	Moderate INFERRED The Phase I was categorized as feasibility research -- "discover a new method." Phase I deliverables under AFWERX open-topic norms include a preliminary and final technical report with proof-of-concept validation of the proposed method, but not a production system. Tech Vector completes the transition from feasibility-demonstrated method to a production-ready, commercially deployed AI product -- the step the STTR program explicitly anticipates when it defines Phase III as work that completes prior STTR efforts. The completes prong is the most legally literal of the three and is a valid argument when framing this as the full technology-transition arc from government R&D to commercial deployment. The confidence is INFERRED rather than KNOWN because no public document links the Phase I deliverables to a named commercialization plan that Tech Vector is completing. Tech Vector completes the transition arc from the Phase I feasibility proof (FA8649-21-P-0756 final technical report) to a production-deployed, commercially sustained AI product available to DoD buyers under a Phase III sole-source action.
NorthStar Conversational multi-agent AI assistant; natural-language question returns SLM-generated interactive research report from the 200M-document corpus	Moderate INFERRED NorthStar draws from the same 200M-document corpus and the same underlying AI methodology that the Phase I established. The derives prong is available: NorthStar's multi-agent SLM architecture produces research reports from patent, grant, and financial data -- the same data types named in the Phase I scope. The corpus-construction method and the AI-over-connected-big-data approach both trace back to the CHN Phase I work. The fit is moderate rather than strong because NorthStar adds a generative AI interface layer (SLM-generated interactive reports, conversational follow-up) that was not a feature of the Phase I scope and that represents a distinct technical capability beyond what was researched in 2021. The derives prong holds at the data-and-method level but is thinner than Tech Vector's direct textual match. INFERRED -- Tim's Phase I technical report would confirm whether the natural-language interface was in scope or is clearly an extension. NorthStar derives its corpus and core AI methodology from the multi-big-data-set approach developed under FA8649-21-P-0756; the data sources (patents, grants, financial) and the goal (surface who is doing relevant research) match the Phase I scope at the method level.	Strong INFERRED NorthStar's primary claim is EXTENDS. The AFWERX execution guide defines "extends" as "work for other applications not researched or performed in prior SBIR/STTR efforts." The Phase I discovered an AI method for identifying technology relationships over connected big data. NorthStar applies that method through a different access modality: a conversational interface backed by technology-specific Small Language Models that generate interactive research reports. The SLM-generative layer, the natural-language question format, and the interactive follow-up capability were not applications researched in the 3-month feasibility Phase I. NorthStar is the extension of the Phase I method into a strategy-layer buyer experience -- a researcher asking "who is leading work on quantum sensing for navigation?" rather than running a structured query. The inference is strong: the Phase I corpus and method are the ancestor; the SLM conversational surface is the extension. Tim's Phase I technical report would upgrade this to KNOWN by confirming the NLP/conversational layer was not in the Phase I scope. NorthStar extends the Phase I AI methodology to a conversational SLM interface -- an application not researched in FA8649-21-P-0756 -- making the same corpus accessible to strategy-layer buyers through natural-language queries and interactive reports.	Weak INFERRED The completes prong requires that NorthStar finish something the Phase I started but left unfinished. The Phase I was a feasibility exercise; NorthStar's SLM-generative interface is a new capability layer added well after the Phase I concluded, not the completion of a Phase I deliverable that was left open. The completes argument for NorthStar would require showing that the Phase I explicitly planned a conversational interface as a future deliverable and that NorthStar is the delivery of that plan -- a connection not supported by any public source. The DERIVES and EXTENDS prongs are both more defensible for NorthStar; a CO relying on completes for NorthStar would be on thinner ground. Report this honestly: the completes prong is available in theory (NorthStar completes the technology-transition arc from government R&D to end-user-ready product) but the specific "completes" nexus is weak compared to the EXTENDS argument. NorthStar completes the technology-transition arc from Phase I feasibility to a deployed, user-facing AI research assistant -- but the direct nexus between a specific Phase I deliverable and NorthStar's SLM interface is not established in public sources and should not be the lead argument.
Defense BD Structured DoD RDT&E budget dashboard; $100B+ in spending data, filterable by agency / budget activity / program element / technology cluster for budget forecasting	Moderate INFERRED The Phase I scope explicitly includes "financial" as one of the three connected big data sets. If "financial" in the Phase I context referred to government budget or appropriations data, Defense BD's structured RDT&E dashboard derives directly from that financial-data dimension of the Phase I methodology. Defense BD structures $100B+ in DoD RDT&E spending data into filterable budget intelligence -- the same goal of understanding "who is funding what research" expressed through the financial-data layer rather than the patent-and-grant layer. The fit is moderate rather than strong because the "financial" data type in the Phase I scope is ambiguous (see Flag 1 above), and because Defense BD is the most architecturally distinct of the three products -- structured BI dashboard over budget submissions, not AI-generative or query-intelligence in the same sense as Tech Vector. TO-CONFIRM resolves to STRONG if Tim's Phase I technical report shows the financial data layer was DoD budget data. Defense BD derives from the "financial" data dimension of the Phase I scope (FA8649-21-P-0756), structuring DoD RDT&E spending at program-element level -- the same budget-signal layer CHN was researching as part of the connected big-data-set methodology.	Strong INFERRED Defense BD extends the Phase I scope in two documented ways. First, the financial data dimension: whether "financial" in the Phase I verbatim meant commercial investment data or government budget data, Defense BD operationalizes the financial layer of the Phase I methodology into a standalone, buyer-facing product with filter-to-program-element granularity over $100B+ in RDT&E data. This is an extension: the Phase I researched a method; Defense BD delivers a structured commercial product on the financial-data component. Second, the scope of the buyer mission: the Phase I was oriented to identifying commercialization partners for government-funded research. Defense BD serves defense commercial firms trying to forecast where DoD is allocating R&D funding -- a related but distinct buyer question (where is money going, not who is developing what). This extends the Phase I use case from tech-transfer intelligence to budget-signal intelligence. Both extensions are credible under the broad "extends" definition. INFERRED -- Tim's Phase I technical report will confirm whether the budget-data layer was in scope or is clearly an extension. Defense BD extends the financial-data dimension of FA8649-21-P-0756 -- building a dedicated structured-RDT&E-budget product on the financial layer of the Phase I methodology, serving the budget-signal buyer mission not addressed by the Phase I feasibility exercise.	Weak INFERRED The completes argument for Defense BD requires connecting the Phase I deliverables to a specific unfinished thread that Defense BD closes. The Phase I was a 3-month feasibility exercise; the final technical report would have addressed whether the AI-over-connected-big-data method was feasible. Defense BD's structured budget dashboard is not the obvious completion of a feasibility finding -- it is a product built on one data layer of the Phase I methodology. A CO relying on completes for Defense BD would need Tim's Phase I documents to show that the budget-data integration was explicitly proposed as future work in the Phase I report. Absent that document, the completes prong is architecturally available (Defense BD completes the financial-data thread of the Phase I methodology into a production-deployed product) but the argument is thinner than EXTENDS. The EXTENDS cell is the right lead for Defense BD; completes is available as a fallback framing for a CO who wants belt-and-suspenders coverage. Defense BD completes the financial-data integration thread of the Phase I methodology into a production-deployed RDT&E budget dashboard -- but the direct nexus to a specific open Phase I deliverable is not established without Tim's technical report and should not be the primary argument.

Product

Derives from

Extends

Completes

Tech Vector

Guided query over 200M-document corpus; returns technology-landscape analysis across patents, grants, publications, and DoD spending data

Strong KNOWN

Tech Vector is the productized commercial execution of exactly what the Phase I was funded to discover. The Phase I scope verbatim -- "discover a new method for identifying commercialization partners for government-funded research output using multiple, connected big data sets (patents, grants, financial) powered by artificial intelligence" -- maps directly and textually to Tech Vector's function: a guided query over connected big data (patents, grants, legal filings, DoD spending) powered by AI, returning a landscape of who is doing what in a given technology domain. The data sources in the Phase I description (patents, grants, financial) are the same data sources Tech Vector accesses. The goal of the Phase I (identify who is doing relevant research across government-funded output) is the same goal Tech Vector delivers to its defense buyer. Both the Phase I verbatim and the Tech Vector product description are primary public sources; the textual match is direct, not inferred.

Tech Vector operationalizes the AI-over-connected-big-data method developed under FA8649-21-P-0756; the Phase I verbatim scope and the product corpus match on data sources (patents, grants, financial) and on function (identifying commercialization partners for government research output).

Moderate INFERRED

The Phase I was a 3-month customer-discovery and feasibility confirmation exercise -- not a full commercial product. The gap between a feasibility-demonstrated AI method and a commercial guided-query interface with a 200M-document corpus, 15,000+ RDT&E project integrations, and a real-time query architecture represents genuine extension: Tech Vector applies the Phase I method to a broader and deeper commercial-scale data estate than any 3-month feasibility study could have built. The product scope -- global coverage, defense industrial base buyers, commercial subscriptions -- extends the original dual-use dual-purpose framing into a fully commercial product with an independently monetized user base. This extends prong is secondary to the DERIVES cell (which is the stronger argument), but it is available and honest as a companion framing if the CO wants to acknowledge both the lineage and the commercial-scale expansion.

Tech Vector extends the Phase I feasibility method to a commercial-scale corpus and a broader buyer set (defense industrials, investment analysts, tech-transfer offices) not served by a 3-month STTR deliverable.

Moderate INFERRED

The Phase I was categorized as feasibility research -- "discover a new method." Phase I deliverables under AFWERX open-topic norms include a preliminary and final technical report with proof-of-concept validation of the proposed method, but not a production system. Tech Vector completes the transition from feasibility-demonstrated method to a production-ready, commercially deployed AI product -- the step the STTR program explicitly anticipates when it defines Phase III as work that completes prior STTR efforts. The completes prong is the most legally literal of the three and is a valid argument when framing this as the full technology-transition arc from government R&D to commercial deployment. The confidence is INFERRED rather than KNOWN because no public document links the Phase I deliverables to a named commercialization plan that Tech Vector is completing.

Tech Vector completes the transition arc from the Phase I feasibility proof (FA8649-21-P-0756 final technical report) to a production-deployed, commercially sustained AI product available to DoD buyers under a Phase III sole-source action.

NorthStar

Conversational multi-agent AI assistant; natural-language question returns SLM-generated interactive research report from the 200M-document corpus

Moderate INFERRED

NorthStar draws from the same 200M-document corpus and the same underlying AI methodology that the Phase I established. The derives prong is available: NorthStar's multi-agent SLM architecture produces research reports from patent, grant, and financial data -- the same data types named in the Phase I scope. The corpus-construction method and the AI-over-connected-big-data approach both trace back to the CHN Phase I work. The fit is moderate rather than strong because NorthStar adds a generative AI interface layer (SLM-generated interactive reports, conversational follow-up) that was not a feature of the Phase I scope and that represents a distinct technical capability beyond what was researched in 2021. The derives prong holds at the data-and-method level but is thinner than Tech Vector's direct textual match. INFERRED -- Tim's Phase I technical report would confirm whether the natural-language interface was in scope or is clearly an extension.

NorthStar derives its corpus and core AI methodology from the multi-big-data-set approach developed under FA8649-21-P-0756; the data sources (patents, grants, financial) and the goal (surface who is doing relevant research) match the Phase I scope at the method level.

Strong INFERRED

NorthStar's primary claim is EXTENDS. The AFWERX execution guide defines "extends" as "work for other applications not researched or performed in prior SBIR/STTR efforts." The Phase I discovered an AI method for identifying technology relationships over connected big data. NorthStar applies that method through a different access modality: a conversational interface backed by technology-specific Small Language Models that generate interactive research reports. The SLM-generative layer, the natural-language question format, and the interactive follow-up capability were not applications researched in the 3-month feasibility Phase I. NorthStar is the extension of the Phase I method into a strategy-layer buyer experience -- a researcher asking "who is leading work on quantum sensing for navigation?" rather than running a structured query. The inference is strong: the Phase I corpus and method are the ancestor; the SLM conversational surface is the extension. Tim's Phase I technical report would upgrade this to KNOWN by confirming the NLP/conversational layer was not in the Phase I scope.

NorthStar extends the Phase I AI methodology to a conversational SLM interface -- an application not researched in FA8649-21-P-0756 -- making the same corpus accessible to strategy-layer buyers through natural-language queries and interactive reports.

Weak INFERRED

The completes prong requires that NorthStar finish something the Phase I started but left unfinished. The Phase I was a feasibility exercise; NorthStar's SLM-generative interface is a new capability layer added well after the Phase I concluded, not the completion of a Phase I deliverable that was left open. The completes argument for NorthStar would require showing that the Phase I explicitly planned a conversational interface as a future deliverable and that NorthStar is the delivery of that plan -- a connection not supported by any public source. The DERIVES and EXTENDS prongs are both more defensible for NorthStar; a CO relying on completes for NorthStar would be on thinner ground. Report this honestly: the completes prong is available in theory (NorthStar completes the technology-transition arc from government R&D to end-user-ready product) but the specific "completes" nexus is weak compared to the EXTENDS argument.

NorthStar completes the technology-transition arc from Phase I feasibility to a deployed, user-facing AI research assistant -- but the direct nexus between a specific Phase I deliverable and NorthStar's SLM interface is not established in public sources and should not be the lead argument.

Defense BD

Structured DoD RDT&E budget dashboard; $100B+ in spending data, filterable by agency / budget activity / program element / technology cluster for budget forecasting

Moderate INFERRED

The Phase I scope explicitly includes "financial" as one of the three connected big data sets. If "financial" in the Phase I context referred to government budget or appropriations data, Defense BD's structured RDT&E dashboard derives directly from that financial-data dimension of the Phase I methodology. Defense BD structures $100B+ in DoD RDT&E spending data into filterable budget intelligence -- the same goal of understanding "who is funding what research" expressed through the financial-data layer rather than the patent-and-grant layer. The fit is moderate rather than strong because the "financial" data type in the Phase I scope is ambiguous (see Flag 1 above), and because Defense BD is the most architecturally distinct of the three products -- structured BI dashboard over budget submissions, not AI-generative or query-intelligence in the same sense as Tech Vector. TO-CONFIRM resolves to STRONG if Tim's Phase I technical report shows the financial data layer was DoD budget data.

Defense BD derives from the "financial" data dimension of the Phase I scope (FA8649-21-P-0756), structuring DoD RDT&E spending at program-element level -- the same budget-signal layer CHN was researching as part of the connected big-data-set methodology.

Strong INFERRED

Defense BD extends the Phase I scope in two documented ways. First, the financial data dimension: whether "financial" in the Phase I verbatim meant commercial investment data or government budget data, Defense BD operationalizes the financial layer of the Phase I methodology into a standalone, buyer-facing product with filter-to-program-element granularity over $100B+ in RDT&E data. This is an extension: the Phase I researched a method; Defense BD delivers a structured commercial product on the financial-data component. Second, the scope of the buyer mission: the Phase I was oriented to identifying commercialization partners for government-funded research. Defense BD serves defense commercial firms trying to forecast where DoD is allocating R&D funding -- a related but distinct buyer question (where is money going, not who is developing what). This extends the Phase I use case from tech-transfer intelligence to budget-signal intelligence. Both extensions are credible under the broad "extends" definition. INFERRED -- Tim's Phase I technical report will confirm whether the budget-data layer was in scope or is clearly an extension.

Defense BD extends the financial-data dimension of FA8649-21-P-0756 -- building a dedicated structured-RDT&E-budget product on the financial layer of the Phase I methodology, serving the budget-signal buyer mission not addressed by the Phase I feasibility exercise.

Weak INFERRED

The completes argument for Defense BD requires connecting the Phase I deliverables to a specific unfinished thread that Defense BD closes. The Phase I was a 3-month feasibility exercise; the final technical report would have addressed whether the AI-over-connected-big-data method was feasible. Defense BD's structured budget dashboard is not the obvious completion of a feasibility finding -- it is a product built on one data layer of the Phase I methodology. A CO relying on completes for Defense BD would need Tim's Phase I documents to show that the budget-data integration was explicitly proposed as future work in the Phase I report. Absent that document, the completes prong is architecturally available (Defense BD completes the financial-data thread of the Phase I methodology into a production-deployed product) but the argument is thinner than EXTENDS. The EXTENDS cell is the right lead for Defense BD; completes is available as a fallback framing for a CO who wants belt-and-suspenders coverage.

Defense BD completes the financial-data integration thread of the Phase I methodology into a production-deployed RDT&E budget dashboard -- but the direct nexus to a specific open Phase I deliverable is not established without Tim's technical report and should not be the primary argument.

Reading the diagonal

The three strongest cells -- the diagonal arguments a CO will use in the action memo -- are:

Tech Vector DERIVES: STRONG / KNOWN. Direct textual match between Phase I verbatim scope and product description. Both are primary public sources. This is the anchor argument and the one a CO can quote verbatim in Section 4 of the action memo.
NorthStar EXTENDS: STRONG / INFERRED. The conversational SLM modality is an application not researched in the Phase I. Legally sufficient under the AFWERX "extends" definition. Upgrades to KNOWN when Tim's Phase I technical report confirms the NLP interface was out of Phase I scope.
Defense BD EXTENDS: STRONG / INFERRED. The financial-data dimension of the Phase I scope is the ancestor; Defense BD is the standalone product built on it. Both readings of "financial" (commercial investment OR government budget) support the EXTENDS argument. The TO-CONFIRM on Flag 1 determines whether "financial" is a direct match or an adjacent extension.

The off-diagonal cells (Tech Vector EXTENDS, NorthStar DERIVES, Defense BD DERIVES) are MODERATE -- honest and legally available but not the primary arguments. The two WEAK cells (NorthStar COMPLETES, Defense BD COMPLETES) are disclosed honestly: the completes prong does not have a strong documented nexus for either product without Tim's Phase I technical report. A CO should not lead with completes for either NorthStar or Defense BD.

What Tim's Phase I technical report would change

The TO-CONFIRM items in this grid all route to one document: the Phase I final technical report that Tim holds. Amyn offered to review it under NDA on the 2026-05-20 discovery call. That document would resolve:

The "financial" data definition (Flag 1): upgrades Defense BD DERIVES from MODERATE/TO-CONFIRM to STRONG/KNOWN if financial = government budget data.
The NorthStar EXTENDS confidence: upgrades from INFERRED to KNOWN if the technical report confirms no conversational/NLP interface was in scope.
The NorthStar and Defense BD COMPLETES cells: upgrades from WEAK to at minimum MODERATE if the technical report shows either dashboard or conversational interface was planned as future work.
The STTR research institution partner (30% rule): does not affect the derives-from arguments, but confirms the technical lineage and may explain the OSI&A relationship start date.

Before finalizing the CO action memo (A3), Amyn should obtain and review Tim's Phase I technical report. The A2 grid above is complete and defensible as-is; the report would lock down the two INFERRED STRONG cells and potentially upgrade two WEAK cells to MODERATE.

CO-facing one-liner summary

If the CO action memo needs a single paragraph covering all three products, it reads:

The work to be procured under this Phase III action -- NorthAI's Tech Vector, NorthStar, and Defense BD platforms -- derives from, extends, and completes efforts made under STTR Phase I award FA8649-21-P-0756 (CHN Analytics LLC, AFWERX, AF20C-TCSO1, 2021). Tech Vector derives from the Phase I methodology directly: the Phase I scope was to discover an AI method for identifying commercialization partners using connected big data sets of patents, grants, and financial data; Tech Vector is the productized implementation of that exact method over a 200M-document corpus at commercial scale. NorthStar extends the Phase I methodology to a conversational SLM interface not researched in Phase I, making the same corpus accessible through natural-language queries and interactive generated reports. Defense BD extends the financial-data dimension of the Phase I methodology into a structured RDT&E budget-intelligence product covering $100B+ in DoD spending data. All three products are built and maintained by the entity that holds FA8649-21-P-0756 subject to confirmation of legal continuity between CHN Analytics LLC and the current NorthAI operating entity.