Case 2 — Canadian SR&ED Assessment Agent: Attempted Burial

Domain: Canadian public-subsidy compliance — Scientific Research and Experimental Development (SR&ED) federal R&D tax-credit program, administered by the Canada Revenue Agency under the Income Tax Act

Time period: Operator-observed; specific window left structural pending operator naming decision

Outcome: A compliance-assessment agent designed to make the SR&ED qualification surface transparent and quantifiable was perceived as a threat by named institutional actors with weight on the program-design surface, and those actors moved to bury the agent rather than compete on merit

Methodology classifications: Methodology 14 (lock-in as darkness), Methodology 17 (early-iteration commoditization — institutional-runaway sequence phases 3 and 4)

Naming posture for this case

The public program is named. SR&ED is a federal program; its existence, its statutory basis, its administering agency, and its widely-documented compliance economics are all public record and fair game for educational analysis.

The specific institutional actors who moved to bury the assessment agent are left as structural roles in this draft. The operator (lead operator) decides before any wider distribution whether to fill the named specifics in or leave them at the role level. The structural lesson holds either way. Placeholders used in this draft:

PROGRAM-DESIGN LAYER — the federal program-administration surface where eligibility criteria, documentation requirements, and reviewer discretion are set. Includes CRA SR&ED Directorate and the policy interface above it.
CROWN-UTILITY-ADJACENT ADVISORY LAYER — provincial Crown-corporation-adjacent advisory bodies whose mandate includes shaping which R&D activity counts as eligible in their jurisdiction. The operator observed this layer's behavior in New Brunswick; cognate layers exist in other provinces.
CONSULTANCY ECOSYSTEM — the specialized SR&ED consulting industry that earns contingent fees (typically 15–30% of claim value) for navigating the qualification surface on behalf of claimants. Includes Big Four practices, mid-tier SR&ED specialty firms, and boutique advisors. The industry's economics depend on the qualification surface remaining opaque enough that claimants cannot do the assessment themselves.
NB SCALE TIPPERS — operator-observed actors in the New Brunswick institutional environment whose weight is sufficient to tip program-design decisions and whose interests aligned with burying the assessment agent. Specific identities are the operator's call.

The structural lesson is independent of which names go in which slots.

The setup

SR&ED is, by statutory design, a public subsidy intended to encourage Canadian R&D activity. The program issues refundable and non-refundable tax credits against expenditures on qualifying scientific or experimental development work. In any given fiscal year, the program processes claims in the multi-billion-dollar range and touches every sector from biotech to enterprise software to advanced manufacturing.

The qualification surface — what activity qualifies, what documentation supports a claim, what threshold of technological uncertainty must be demonstrated, what threshold of systematic investigation must be evidenced — is not, in practice, transparent. The program is administered through written guidance, reviewer interpretation, and case-by-case adjudication. Two claims with structurally identical underlying activity can receive different outcomes depending on which reviewer adjudicates, which consultancy prepared the file, and which province's claim culture surrounds the submission.

This opacity is not an accident. It is the structural property that makes the SR&ED consulting ecosystem economically viable. If the qualification surface were transparent and quantifiable — if a claimant could run their R&D activity through a deterministic assessment and know with confidence whether it qualified, at what level, and with what documentation — the contingent-fee consulting industry's value proposition would collapse to the residual value of the document-preparation work, which is far smaller than the contingent fee.

The opacity also benefits the program-design layer. Reviewer discretion is administrative power. Transparency removes administrative power. A program that is administered through reviewer discretion is one whose direction can be tilted, year over year, by quiet adjustments to internal guidance. A program with a transparent qualification surface cannot be tilted that way; its direction is bound to the surface.

Both layers — the consultancy ecosystem and the program-design layer — had completed the institutional-runaway sequence on the SR&ED qualification surface. Useful start: the program was, in its origin, a real instrument for encouraging Canadian R&D investment, and the consulting ecosystem began as a legitimate translation layer between claimants and a complex statutory regime. Asymmetric authority accrual: over decades, the consultancy ecosystem accumulated specialized institutional knowledge that no individual claimant could replicate without disproportionate cost; the program-design layer accumulated reviewer-discretion authority that no individual claimant could push back against without disproportionate cost. Only-path consolidation: by the operator-observed window, sophisticated SR&ED claims were not being prepared by claimants — they were being prepared by the consultancy ecosystem, on contingent fees, with reviewer-relationship-management as part of the value being purchased. Resistance to exit: any claimant who attempted to bypass the consultancy ecosystem and self-prepare faced a higher rejection rate, longer review cycles, and an asymmetric audit posture that priced the bypass attempt above the contingent fee. The lock-in was not contractual. It was structural.

The customer base was not happy. The customer base was locked in. The two were not the same.

What the assessment agent did

The operator (or, more precisely, a team in the operator-adjacent environment) built an SR&ED compliance assessment agent. The technical surface is straightforward: a structured assessment tool that takes a description of an R&D activity, a documentation set, and an expenditure ledger, and returns a deterministic classification across the SR&ED qualification dimensions — technological uncertainty, systematic investigation, technological advancement, and supporting evidence sufficiency. The agent's output is a qualification reading the claimant can read directly, with reasoning traces the claimant can follow.

The agent does not replace the statutory program. It does not replace CRA adjudication. It does what any operator on the methodology side of the lattice would recognize as the early-iteration commoditization move: it makes a previously-opaque qualification surface substrate-portable for the claimant. A claimant who runs the assessment agent walks into the consultancy conversation, or into the CRA submission, with their own classification reading already in hand. The claimant is no longer the dependent party in the conversation; the consultancy is now competing against the claimant's own analysis.

Three properties of the agent matter structurally:

1. Substrate portability for the claimant

The agent's reading travels with the claimant. The claimant does not need the consultancy to access it, does not need the program-design layer's blessing to use it, and does not need a multi-year engagement to internalize its logic. The claimant runs the agent, reads the output, and is now able to evaluate the consultancy's recommendation on the same surface the consultancy is using.

This is the Avalara-equivalent move from Case 1, translated into the public-subsidy domain. Avalara made tax-engine portability travel with the foreign mid-cap. The SR&ED assessment agent makes qualification-surface portability travel with the R&D claimant.

2. Transparency of the qualification reasoning

The agent does not produce a number. It produces a classified reading with reasoning traces. The claimant can see why an activity is being read as qualifying or not, which sub-criterion is the determining factor, and which documentation gap (if any) is moving the reading. This is what makes the agent commoditizing rather than merely competitive — it is not selling a better consultancy service; it is dissolving the opacity that the consultancy service was priced against.

3. Refusal of contingent-fee structure

The agent's distribution model does not depend on a percentage of claim value. The claimant can run it, learn from it, and use it without entering a contingent-fee relationship. The agent treats the qualification surface the way Methodology 17 treats AI substrate access: as something that should be in the claimant's hands by construction, not gated by an institutional layer that earns rent on opacity.

What the burying actors did

The pattern observed is the diagnostic pattern of late-phase institutional-runaway sequence. The burying actors did not engage the agent on its technical merit. They did not produce a competing assessment instrument that was more accurate or better-calibrated. They did not argue that the agent's classification surface was wrong. Each of those moves would have placed them on the merit surface, and on the merit surface the agent's transparency was the disadvantage to argue against — a difficult argument to make publicly without revealing the institutional interest in opacity.

Instead, the observed burying pattern operated through institutional weight on surfaces the agent could not reach:

Program-design-layer signaling. Quiet guidance, through the channels the program-design layer routinely uses, that submissions accompanied by or referencing the assessment agent's output would receive heightened review scrutiny. This is not a published policy. It is a reviewer-culture adjustment. The signal does not need to be loud; it needs only to reach the consultancy ecosystem and the larger claimants.

Crown-utility-adjacent advisory-layer pressure. In the New Brunswick environment specifically, advisory-layer actors with weight on innovation-program design moved to characterize the agent as out-of-scope, premature, or insufficiently aligned with provincial program intent. The substantive merit of the agent was not the conversation. The conversation was about whether the agent should be considered part of the legitimate ecosystem at all.

Consultancy-ecosystem response. The consulting industry's response had two registers. The visible register was technical critique — the agent could not capture the nuance of reviewer relationships, the agent did not account for sector-specific case-history, the agent was a starting point but not a substitute. The invisible register was the more decisive one: the consultancy ecosystem's contingent-fee customers were quietly counseled that referencing the agent's output in a submission would be received unfavorably. The contingent-fee customers had no incentive to dispute this counsel; the contingent fee was already paid against the claim's success, not against the claim's transparency.

The NB scale-tipper layer. Actors in the New Brunswick institutional environment with weight to tip program-design decisions moved on the question of whether the agent should be allowed to operate at all in that jurisdiction. The weight was not exercised through statute. It was exercised through the institutional channels through which Crown-adjacent decisions are routinely made — boardroom-adjacent advisory, ministerial-adjacent signaling, and the quiet alignment of innovation-ecosystem narratives around the position that the agent was unhelpful to provincial R&D growth.

None of these moves are the merit move. Each of these moves is the late-phase runaway-sequence move. An institution in phase 2 of the runaway sequence (asymmetric authority accrual) competes; it can still afford to. An institution in phase 3 (only-path consolidation) defends; it has a position to hold. An institution in phase 4 (resistance to exit) buries; the exit is the threat, and the threat-substrate is the thing to attack rather than the substrate's merit.

What this teaches

The structural lesson of Case 2 is the diagnostic value of attempted-burial behavior. When an institutional actor responds to a substrate-portability instrument by attempting to deny the instrument access to its operating surface rather than by competing against it on merit, the institutional actor is signaling its own phase position in the runaway sequence.

This is the operator-recognizable signal. It does not require inside knowledge of the burying institution. It does not require a leak or a whistleblower. The pattern is observable from outside:

Merit-surface refusal. The burying actor declines to argue the merit of the instrument's output. The argument is always about scope, legitimacy, premature timing, or insufficient ecosystem alignment — never about whether the output is correct.
Channel-substitution attack. The burying actor's moves are routed through institutional channels the instrument cannot answer on. Reviewer-culture signals, advisory-layer characterizations, contingent-fee-customer counseling, and Crown-adjacent narrative alignment all share this property: the instrument cannot defend itself on these channels because the instrument is not a participant on them.
Opacity preservation. The institutional actor's defended position turns out, on inspection, to be the preservation of an opacity that the actor benefits from. The actor's stated interest (program integrity, claimant protection, ecosystem stability) does not predict the actor's behavior; the preservation of opacity does.
Time-asymmetric pressure. The burying institution has time. The instrument's operators do not. The burying move is to make the instrument's continued operation costly enough, over a long enough horizon, that the instrument's operators abandon the substrate before the institution has to defeat them on it. This is the phase-4 hallmark.

Recognition of this pattern is the operator skill. The reason recognition matters before disclosure is that the routing decision for the instrument's substrate must be made before the burying institution sees the substrate. If the substrate is routed through institutional channels the burying actor can reach (a provincial pilot program, a Crown-adjacent funding line, a partnered consultancy distribution), the burying move can cut the substrate at the channel. If the substrate is routed through channels the burying actor cannot reach (direct claimant distribution, open-protocol publication, jurisdictional plurality), the burying move loses its reach.

The pair this case makes with Case 1

Case 1 (Vertex/Sabrix vs. Avalara) and Case 2 (SR&ED assessment agent) are deliberately paired because they sit at opposite ends of a substrate-mediation spectrum.

Case 1 is market-mediated lock-in. The institutional incumbents had no surface above the market through which to bury Avalara. Mid-cap CFOs in the foreign-installed segment could route around the incumbents' Big Four channel, and the lock-in fell when the rerouting cost dropped below the lock-in cost. The free marketplace did the structural work; the methodology operator's job was to recognize which entrant was the substrate-portability competitor.

Case 2 is policy-mediated lock-in. The institutional incumbents had a surface above the market — the program-design layer and its Crown-adjacent satellites — and that surface was beyond the market's reach. The free marketplace could not, by itself, defeat the lock-in, because the lock-in did not live in the market. It lived in the qualification-surface design. A pure-market substrate-portability move did not have anywhere to go; the substrate had to be routed outside the burying institutions' reach and the burying behavior had to be anticipated and pre-empted, not merely competed against.

The two cases together teach the operator that lock-in recognition is not the same problem in both substrates. In market substrates, recognition is sufficient — the market will defeat the lock-in once the substrate-portability entrant is correctly identified. In policy substrates, recognition is necessary but not sufficient — the substrate routing decision must be made with the burying behavior in view, before the burying institution sees the substrate.

The comparative-analysis file (comparative-analysis.md) develops this distinction further, with five recognition prompts the operator should be able to apply to their own working environment.

Three classification questions for the operator

Before moving on to the comparative analysis, the operator should be able to answer these about Case 2:

Which phase of the institutional-runaway sequence is each burying actor in, and on what observable evidence? (Methodology 17 phases 1–4 — useful start, asymmetric authority accrual, only-path consolidation, resistance to exit.) The answer is not the same for all four placeholder layers in this case.

What substrate routing would have made the burying behavior reach-less? Identify, by name or by role, the institutional channel the burying actor used. Then identify the alternative channel that would have placed the substrate outside that reach. This is the pre-emption design.

Is the operator's own current working environment closer to Case 1 or Case 2 in its lock-in substrate? This is the application question. The answer determines whether market recognition is sufficient (Case 1 disposition) or whether pre-emptive routing is required (Case 2 disposition). The operator's own SAP-environment Stage 2 RISE-push observation, captured elsewhere, is a Case 2 cognate; recognizing this is the first step of the application.

Cross-references

eve-glyph-lattice/methodology/14-lock-in-as-darkness.md — the principle.
eve-glyph-lattice/methodology/17-early-iteration-hyperscaler-commoditization.md — the runaway sequence and the commoditization counter-move; phases 3 and 4 are the diagnostic frame for Case 2.
eve-glyph-lattice/methodology/19-sap-inflection-point-finding.md — the named-institution application; the SAP environment is a third substrate cognate to Cases 1 and 2.
case-1-vertex-avalara-commoditizer-won.md — the market-mediated end of the spectrum.
comparative-analysis.md — the structural distillation across both cases.

Pour le bien-être du peuple.

© 2026 Dany Theriault. EVE “digital stem cell” glyph and glyph-based design principles — all rights reserved. Stewardship of rights of use and assignment for large public and institutional usage rests with the Pacific Utilities Design Council. Published as a time-stamped record of authorship and intent.
pour le bien-être du peuple