The Acme.Foundations LLC Chief AI Risk Officer opens her Q2 2026 portfolio review with a stack of five invitation letters received in one calendar week: the AI Office GPAI Code Transparency Working Group sub-working-group on training-data summary templates; the MLCommons Responsible AI Working Group quarterly steering call; the BSA AI Policy Working Group enterprise-deployer subgroup chartered to draft model contract language; the AI Verify Foundation cross-Atlantic assurance pilot; and the Frontier Model Forum systemic-risk evaluation sharing forum. The CAIO turns to her: "Which do we actually need? We have €240K in coalition budget against eight live invitations across the year. The board wants a ratified portfolio. Walk me through the selection framework, the twelve-element engagement standard, the anti-trust safe-harbor discipline, and how you intend to quantify return on something with no short-term ROI signal. And tell me what happens if we say yes to all five, because I suspect the failure mode is dilution and free-rider perception, not under-engagement." This lesson is the L5 executive-tier orchestration view of GPAI Code working groups and the 2026 industry-coalition portfolio: the Article 56 Code working-group structure (Transparency, Copyright, Safety & Security chapters with sub-working-groups); eight active coalitions selected for the 2026 enterprise CAIRO portfolio; the coalition selection framework; the twelve-element CAIRO engagement standard; anti-trust safe-harbor + IP discipline; ROI quantification across long-horizon indicators; seven common 2026 coalition mistakes; the Acme.Foundations 2026 portfolio of four active engagements with €240K annual budget and quarterly AIGC briefing cadence.

Why a Coalition Portfolio Matters in 2026 (And Why the CAIRO Owns It)

The 2026 AI governance environment runs on three reinforcing rails, regulators, standards bodies, and industry coalitions, and an L5 enterprise CAIRO who orchestrates only the first two has a visible coherence gap. Lessons 101 and 103 established the regulator-engagement layer: proactive engagement with the EU AI Office, Member State authorities, FTC, CFPB, OCC, NIST, CAISI, and the AISIC consortium. Lesson 095 established the standards-engagement layer at ISO JTC1 SC42 and CEN-CENELEC. The coalition layer sits between both and amplifies both: coalitions translate regulator priorities into industry-implementable patterns, channel collective industry feedback into regulator dialogues, and provide the venue where competitors converge on the operational baseline that becomes the de facto compliance template.

The 2026 trigger for coalition prominence is Article 56 of the EU AI Act. Signatories of the GPAI Code of Practice (lesson 076 operational sign-on workflow; lesson 095 board-level posture decision) are not passive, signature confers participation rights in the Code working groups, which are the venue where the AI Office iterates the operational templates that translate Article 53 and Article 55 statutory obligations into concrete measures. The Transparency chapter working group iterates the training-data summary template, the Annex XI documentation reference architecture, and the KPI taxonomy for measure operating effectiveness. The Copyright chapter working group iterates the rightsholder-engagement SLA model, the TDMRep machine-readable opt-out detection requirements, and the periodic-refresh cadence on opt-out-convention evolution. The Safety & Security chapter working group iterates the model-evaluation methodology, the systemic-risk identification framework, and the serious-incident reporting taxonomy aligned to Article 73. Participation in these working groups is not optional for the firm that takes the signatory posture seriously. It is the mechanism through which the firm's operational reality shapes the template it will then be measured against.

The non-Article-56 coalitions matter for different reasons. Partnership on AI (PAI) provides the multi-stakeholder forum where civil society and academic researchers test industry positions; participation signals openness to scrutiny and provides early signal on reputational risk vectors. MLCommons provides the benchmark suite (MLPerf, AILuminate, the Responsible AI Working Group taxonomy) that increasingly shows up in procurement DDQs and ISO 42001 audit evidence. The BSA AI Policy Working Group provides the enterprise-deployer perspective in policy advocacy with US federal agencies, state legislatures, and the EU Commission. ITI (Information Technology Industry Council) provides the broader tech-sector policy advocacy channel. Business at OECD (BIAC) provides the OECD AI Principles implementation forum that anchors the global trustworthy-AI baseline. The AI Verify Foundation (Singapore-led) provides the assurance-framework venue that bridges Asia-Pacific governance approaches with EU and US frameworks. The Frontier Model Forum (Anthropic, OpenAI, Google, Microsoft) provides the foundation-model-provider-only venue for systemic-risk evaluation sharing. The AI Alliance (IBM, Meta, 50+ enterprise + academic members) provides the open-innovation counterweight to Frontier Model Forum's frontier-closed posture.

The CAIRO owns the coalition portfolio because no one else can. The CAIO owns operational delivery; the General Counsel owns regulatory engagement letter-of-the-law; the Chief Privacy Officer owns the data-protection cross-walk; the AIGC owns governance ratification of specific decisions. None of those roles is chartered to orchestrate the multi-coalition portfolio against the firm's policy-posture and influence-objectives. The 2026 mature pattern places the CAIRO as the single accountable executive for coalition selection, engagement-standard enforcement, anti-trust safe-harbor compliance, coordinated public-position approval, and quarterly AIGC briefing on coalition signal. A firm without that accountability assignment defaults to the failure modes catalogued later in this lesson: joining too many, joining for marketing without contribution, public positions not pre-cleared, IP leakage, anti-trust friction.

The Article 56 Code Working-Group Structure and the Eight 2026 Coalitions

The Article 56 Code of Practice working-group structure organizes around the three chapters, with sub-working-groups under each. Signatory CAIRO representatives participate either directly or via designated technical staff; non-signatory participation is invitation-only at the AI Office's discretion. The Transparency chapter working group hosts sub-working-groups on training-data summary templates (the most active in Q1-Q2 2026), Annex XI technical documentation reference architecture, Annex XII downstream-deployer information templates, and KPI taxonomy for measure operating effectiveness. The Copyright chapter working group hosts sub-working-groups on rightsholder-engagement SLA models, TDMRep machine-readable opt-out detection requirements, and periodic-refresh cadence. The Safety & Security chapter working group hosts sub-working-groups on model-evaluation methodology, systemic-risk identification framework (CBRN + cyber + election integrity + child safety + market manipulation + autonomous-replication), and serious-incident reporting aligned to Article 73. Cadence is typically monthly per sub-working-group with quarterly chapter-level consolidation; deliverables flow to the AI Office for incorporation into the next Code revision cycle (the practitioner-tracked revision cycle is approximately every 18 months).

The eight 2026 industry coalitions form the CAIRO's external-engagement candidate set:

• (1) Partnership on AI (PAI). Multi-stakeholder including civil society, academic researchers, foundation-model providers, deployers, and journalism organizations. Working groups on synthetic media disclosure, ABOUT ML model documentation, fairness benchmarks, and AI/labor impacts. 2026 enterprise value: civil-society perspective integration; reputational risk early-warning; multi-stakeholder credibility signaling for procurement and ESG-AI overlay. Membership tiers: Partner (significant influence + ~$250K annual contribution) and Affiliate (working-group access + ~$50K). Annual board meeting + quarterly working-group cadence.
• (2) MLCommons. Industry consortium producing MLPerf training/inference benchmarks, AILuminate safety benchmark (launched 2024-2025), and Responsible AI Working Group taxonomy. 2026 enterprise value: benchmark results increasingly cited in procurement DDQs and ISO 42001 audit evidence; AILuminate participation signals voluntary safety-evaluation commitment that maps to GPAI Code Safety & Security chapter measures. Membership: General Member (full benchmark participation + working groups) ~$25K-$75K. Quarterly working-group cadence + biannual benchmark publication rounds.
• (3) BSA | The Software Alliance - AI Policy Working Group. Enterprise-software-vendor coalition with strong DC and Brussels presence. AI Policy Working Group drafts model contract language for enterprise AI procurement, advocates for harmonized US federal AI framework, and provides Member State engagement in EU AI Act implementation. 2026 enterprise value: enterprise-deployer policy voice (often missing from foundation-model-provider-dominated coalitions); model contract templates that converge enterprise procurement language; US federal advocacy as the OSTP/NIST AI policy stack matures. BSA membership ~$50K-$200K depending on revenue tier. Monthly AI Policy Working Group + biannual member summit.
• (4) ITI - Information Technology Industry Council. Broader tech-sector advocacy with US federal agencies, state legislatures (notably California, Colorado, Texas, New York), and international counterparts. ITI AI policy positions influence congressional staff briefings, federal agency Request for Information responses, and state attorney-general AI enforcement framing. 2026 enterprise value: state-level AI law engagement (Colorado AI Act, Texas TRAIGA, NYC LL 144) where ITI provides aggregated industry positioning. Membership ~$30K-$150K. Monthly policy committee + quarterly board cadence.
• (5) Business at OECD (BIAC). Official business-stakeholder voice at the OECD, where the OECD AI Principles (2019, updated 2024) anchor the global trustworthy-AI baseline that 47+ jurisdictions reference. BIAC AI Committee influences OECD AI Policy Observatory work, Hiroshima AI Process implementation, and G7/G20 AI ministerials. 2026 enterprise value: global-baseline shaping that propagates into national frameworks; OECD AI Principles citations in ISO 42001 audit evidence and ESG-AI overlay frameworks. BIAC membership via national federation (e.g., USCIB in the US, MEDEF in France); AI Committee participation typically requires national-federation membership ~$25K-$80K + BIAC contribution.
• (6) AI Verify Foundation (Singapore-led). AI assurance-framework consortium launched by the Singapore IMDA and Personal Data Protection Commission, with rapidly growing international participation. Produces the AI Verify testing framework (cross-walked to NIST AI RMF, ISO 42001, EU AI Act), Project Moonshot (LLM evaluation toolkit), and the Global AI Assurance Pilot. 2026 enterprise value: assurance-framework convergence between Asia-Pacific, EU, and US approaches; testing-framework participation that strengthens ISO 42001 + SOC 2+AI audit evidence; Singapore IMDA relationship as a cooperative-regulator engagement anchor for APAC market entry. Membership tiers ~$15K-$60K. Quarterly steering + monthly working-group cadence.
• (7) Frontier Model Forum (FMF). Foundation-model-provider-only coalition (Anthropic, OpenAI, Google, Microsoft as founding members; ~10 additional members by Q1 2026). Charters: systemic-risk evaluation sharing, safety-research coordination, government and civil-society engagement on frontier-model governance, $10M+ AI Safety Fund. 2026 enterprise value: foundation-model-provider peer convening for systemic-risk discussion; cross-walk to Article 55 systemic-risk-evaluation requirements and the GPAI Code Safety & Security chapter. Membership criteria: deploys/develops frontier models per FMF threshold definition. Membership cost not publicly disclosed; estimated $500K+ annual contribution for full members. Quarterly forums + working-group cadence.
• (8) AI Alliance. IBM + Meta-led open-innovation coalition with 50+ members spanning enterprise (Intel, Oracle, ServiceNow, Dell), academic (Cornell, Yale, Stanford HAI), and research institutions (CERN, NASA, A*STAR). Charters: open foundation-model development, evaluation tooling, hardware enablement, skills/education, advocacy for open-source AI in regulatory frameworks (including Article 53(2) carve-out preservation). 2026 enterprise value: open-source posture amplification; academic-research access; counterweight perspective to Frontier Model Forum's closed-frontier posture in policy dialogues. Membership free or modest annual contribution depending on tier. Monthly working-group cadence + annual summit.

The eight coalitions are not mutually exclusive, a 2026 enterprise CAIRO portfolio commonly spans 3-5 active engagements across the set, plus Article 56 Code working-group participation if the firm is a Code signatory. The selection discipline is the load-bearing input: joining all eight dilutes influence, exceeds the practitioner-tested attention budget of any CAIRO function, and exposes the firm to coordination friction across coalition positions that may diverge. The next section formalizes the selection framework.

The Coalition Selection Framework and the 12-Element CAIRO Engagement Standard

The mature 2026 CAIRO selection framework runs against four filters applied sequentially. Filter 1, enterprise profile alignment. Is the firm a foundation-model provider, a deployer, a vendor-platform, or a sector-specific deployer (financial, healthcare, public-sector)? Foundation-model providers cannot ignore the Frontier Model Forum (if eligible) or the Article 56 Code working groups (if signatory) without visibility gap. Enterprise deployers cannot ignore BSA AI Policy Working Group without losing the procurement-language convergence influence. Sector-specific deployers prioritize sector-aligned coalitions (e.g., financial-services firms often add Bank Policy Institute AI Working Group; healthcare deployers often add Healthcare AI Standards Coalition). Filter 2, budget envelope. The €240K-€500K typical 2026 enterprise CAIRO coalition budget accommodates 3-5 active engagements at typical $30K-$100K per engagement, plus designated-staff time at ~0.25-0.5 FTE per active engagement. Joining beyond budget produces dilution; joining below budget leaves influence-objectives unmet. Filter 3, influence-objective fit. Three policy-posture objectives drive 2026 coalition selection: (a) shape regulator templates (Article 56 Code working groups, OECD/BIAC, NIST consortium); (b) shape industry baseline (MLCommons, BSA, ITI, Frontier Model Forum); (c) shape multi-stakeholder credibility (Partnership on AI, AI Alliance, AI Verify Foundation). The CAIRO selects across the three objectives proportional to the firm's strategic-posture priorities. Filter 4 - IP-disclosure tolerance. Working-group participation requires discussing operational practices, evaluation methodologies, and sometimes pre-publication results. Firms with high IP-protection sensitivity favor coalitions with strong NDAs and limited public output (Frontier Model Forum); firms with high open-publication tolerance favor coalitions with public deliverables (AI Alliance, MLCommons).

The 12-element CAIRO coalition engagement standard formalizes the discipline for every active engagement. The standard exists as a CAIRO standard document, ratified by the AIGC, referenced in the CAIRO charter, and reviewed annually. Each element is non-negotiable; deviation requires AIGC ratification on a documented exception basis:

• (1) Scope. Coalition name, charter purpose, working-group(s) the firm joins, observation-only vs voting participation, and the influence-objective category (regulator template / industry baseline / multi-stakeholder credibility).
• (2) Membership criteria. The coalition's stated criteria the firm satisfies (foundation-model provider status, deployer status, revenue tier, sector designation) plus the internal governance approval the firm requires to apply (AIGC vote, board AI subcommittee notification, GC sign-off on membership agreement).
• (3) Designated representatives. CAIRO direct involvement (typically reserved for Article 56 Code working groups + 1-2 strategic coalitions) or designated technical staff (named individuals with role, seniority, AIGC-ratified mandate to represent the firm). Backup designate named for continuity. Required AIGC briefing before participation begins.
• (4) Time commitment. Estimated FTE allocation per quarter (e.g., 0.25 FTE for monthly working-group cadence + quarterly steering); calendar visibility for AIGC; reallocation triggers if commitment exceeds estimate by 25%.
• (5) IP rules. Explicit list of firm IP categories that may NOT be discussed in coalition fora (proprietary model architectures, training-data composition beyond public summary, evaluation methodology details that confer competitive advantage, customer-specific deployment data). Explicit list of categories that MAY be discussed (operational practice patterns, public-domain evaluation results, regulatory-posture decisions already publicly committed).
• (6) Anti-trust safe-harbor. Pre-meeting briefing standard for every coalition meeting: Sherman Act Section 1 + EU Article 101 TFEU + UK Competition Act prohibit coordinated pricing, output, market-allocation, or boycott discussions. Coalition meetings must follow safe-harbor rules: published agenda in advance, antitrust counsel present or on-call, minutes recorded with safe-harbor confirmation language, exit protocol if any discussion crosses the safe-harbor line. Designated representatives trained annually on safe-harbor discipline.
• (7) Public-position approval. Any public statement made on behalf of the firm in a coalition context (signed letter, joint position, public-comment submission, conference panel) requires pre-approval by AIGC + General Counsel. Pre-approval standard: 48-hour minimum lead time, AIGC-approved talking points, GC sign-off on legal exposure, CAIO consultation on operational consequences. Coalition-attributed positions (where the coalition speaks, not the firm) require firm sign-off on the firm's specific contribution to the position.
• (8) Conflict-of-interest declarations. Designated representatives disclose any personal financial interest, prior employment, or advisory relationship that may affect coalition participation. Coalition leadership roles (chair, working-group convener) require additional AIGC ratification given visibility and time commitment. Annual conflict-of-interest refresh.
• (9) Financial contribution. Annual membership cost, working-group surcharges, event sponsorship contributions, in-kind contributions (staff time, hosted meetings, contributed research). Total financial exposure tracked in CAIRO budget envelope; approval threshold per AIGC contribution-tier rules.
• (10) Withdrawal triggers. Pre-defined conditions that trigger CAIRO withdrawal recommendation: coalition charter shift away from firm's influence-objectives; persistent anti-trust safe-harbor friction; coordination misalignment with other coalitions in the portfolio; budget reallocation pressure; coalition reputational risk event affecting firm.
• (11) Reporting to AIGC. Quarterly briefing standard: coalition signal (what was discussed, what was decided, what is pending), firm contribution (positions advanced, deliverables produced), regulator and peer-set reaction, influence-objective progress against the year's plan, recommended portfolio adjustments. Annual coalition-portfolio review presented to AIGC and board AI subcommittee.
• (12) Annual review. Full coalition-engagement review on a fixed annual calendar slot. Renew, expand, reduce, or withdraw decision for each active engagement. CAIRO recommendation with AIGC ratification; new-engagement candidates evaluated against the four-filter selection framework; portfolio rebalanced for the coming year.

The 12-element standard is the documentary backbone that distinguishes a disciplined CAIRO coalition portfolio from an ad-hoc network of working-group attendances. ISO 42001 auditors increasingly request the standard document and recent quarterly briefings as evidence of Clause 5 leadership engagement (Annex A.3) and Clause 9.1 performance evaluation. NIST AI RMF auditors map coalition participation to Govern 1.5 (continual improvement processes) and Govern 6.1 (third-party engagement and oversight). The standard document is not optional infrastructure. It is the artifact regulators, auditors, and the board look for when testing whether the coalition portfolio reflects governance discipline rather than passive networking.

ROI Quantification, Anti-Trust Discipline, and Common 2026 Mistakes

Coalition ROI is hard to quantify in short horizons: the input-to-outcome lag is typically 12-36 months, the contribution attribution is shared across coalition members, and the counterfactual ("what would have happened without our participation?") is unobservable. The 2026 mature pattern measures coalition ROI across five long-horizon indicators rather than attempting short-horizon dollar attribution. Indicator 1, co-authored guidance citations. When a coalition deliverable (working-group output, position paper, benchmark, framework) is cited by a regulator (AI Office guidance, NIST publication, FTC enforcement statement), Member State authority, ISO standard, or major audit firm, the firms credited as contributing authors realize multi-year influence return. The CAIRO tracks citation counts annually. Indicator 2, regulator-engagement quality. Firms active in coalitions on a regulator's priority topic typically experience higher-quality regulator engagement on that topic (more cooperative tone, faster pre-decisional dialogue, more nuanced enforcement framing). The CAIRO tracks regulator-engagement-quality signal across the engaged regulators (lessons 101 + 103). Indicator 3, talent attraction and retention. Coalition leadership roles (working-group chair, board seats) attract senior governance talent who join the firm specifically for the influence platform. The CAIRO tracks talent signal across recruitment and retention metrics. Indicator 4, buyer trust signaling. Procurement DDQs increasingly cite coalition memberships (Partnership on AI affiliate, MLCommons benchmark participation, AI Verify pilot) as positive signals, though typically as Tier-B reference rather than Tier-A pass/fail. The CAIRO tracks DDQ-citation patterns through the sales pipeline. Indicator 5, supply-chain partnership leverage. Coalition co-participation creates partnership-conversation venues with potential customers, vendors, and integrators that would otherwise require dedicated business-development overhead. The CAIRO tracks partnership-pipeline signal originating from coalition relationships.

The five indicators are reported quarterly to the AIGC and annually to the board AI subcommittee. None is a short-horizon revenue lever; together they form the basis for the CAIRO to defend the €240K-€500K annual coalition budget as strategic-posture investment rather than discretionary spending. A firm that cannot articulate at least three of the five indicators in coherent metrics typically experiences coalition-budget pressure during cost-cutting cycles, and the operational consequence is loss of engagement continuity, which is the variable that determines coalition influence outcomes more than any other.

The anti-trust safe-harbor discipline is the rare direct-exposure dimension. Sherman Act Section 1, EU Article 101 TFEU, and UK Competition Act prohibit coordinated pricing, output, market-allocation, and boycott discussions among competitors. Coalition meetings are convening venues for competitors and therefore high-risk for inadvertent anti-trust friction. The disciplined CAIRO operates a five-element safe-harbor protocol: (a) published agenda in advance with antitrust-counsel review for any agenda item that could veer into pricing/output territory; (b) antitrust counsel present or on-call for any meeting addressing pricing-adjacent topics (training cost, compute pricing, model-access pricing, professional-services pricing); (c) meeting minutes recorded with explicit safe-harbor confirmation language; (d) exit protocol, any participant identifying a safe-harbor concern raises it in the meeting and may leave without prejudice; (e) post-meeting CAIRO briefing on safe-harbor compliance with any flagged items escalated to AIGC + GC. The 2024-2025 DOJ Antitrust Division activity in tech-sector coalition matters and the European Commission's preliminary inquiries into model-access pricing practices in early 2026 make this discipline non-optional. Anti-trust missteps are the rare direct-exposure dimension where coalition participation can produce DOJ / EC investigation that overshadows all other ROI considerations.

Seven common 2026 coalition mistakes recur in CAIRO failure-mode analysis:

• (1) Joining too many. The firm joins eight or more coalitions on the theory that more presence equals more influence. The result: dilution across all engagements, no coalition deliverable cites the firm as substantive contributor, designated-staff capacity exhausted, no quarterly AIGC briefing depth on any coalition. The corrective: the four-filter selection framework, 3-5 active engagements, depth over breadth.
• (2) Joining for marketing without technical contribution. The firm joins a coalition to use the membership in marketing material without committing technical staff to substantive working-group work. The result: free-rider perception within the coalition; loss of working-group invitations as coalition leadership channels participation to contributing members; reputational damage as peer firms identify the pattern. The corrective: every active engagement carries a contribution commitment (working-group deliverable, benchmark participation, public-comment co-authorship) tracked in quarterly AIGC briefing.
• (3) Staff participation without executive sponsorship. Technical staff attend coalition working-group meetings without CAIRO oversight or AIGC awareness. The result: positions advanced inconsistently with firm strategic posture; public statements made without pre-approval; coalition leadership unaware of how to escalate firm-relevant matters. The corrective: every designated representative reports to CAIRO; AIGC receives quarterly briefing on coalition signal; coalition leadership has named CAIRO escalation contact.
• (4) Public positions not pre-cleared. A designated representative co-signs a coalition position letter, joint comment submission, or panel statement without 48-hour AIGC + GC pre-approval. The result: legal exposure on positions inconsistent with regulator filings; reputational exposure on positions inconsistent with public commitments; coordination friction across the coalition portfolio if positions diverge across coalitions. The corrective: element (7) of the 12-element standard, enforced without exception.
• (5) IP leakage in working-group discussions. A designated representative discusses proprietary model architecture details, training-data composition specifics, or evaluation methodology beyond what the firm has publicly committed. The result: competitor learning that erodes firm advantage; potential trade-secret protection loss; internal pressure to withdraw from coalition. The corrective: element (5) of the 12-element standard with explicit category lists; annual representative briefing on IP discipline.
• (6) Anti-trust violations. A coalition meeting strays into pricing-coordination, output-coordination, market-allocation, or boycott discussion without exit protocol. The result: DOJ or EC investigation with multi-year duration and reputational damage that dwarfs coalition value. The corrective: element (6) of the 12-element standard with antitrust-counsel on-call for pricing-adjacent agenda items; CAIRO-trained designated representatives; documented exit protocol.
• (7) Ignoring Frontier Model Forum if a foundation-model provider. A foundation-model provider eligible for FMF declines to engage on cost or philosophical grounds. The result: visibility gap on systemic-risk evaluation patterns that other frontier-model providers converge on; missed opportunity to influence Article 55 systemic-risk evaluation methodology through the venue where frontier-model providers actually coordinate; potential exclusion from informal AI Office-FMF dialogues that shape Code Safety & Security chapter iterations. The corrective: foundation-model providers ratify FMF engagement (or documented non-engagement rationale) explicitly at the AIGC; revisit annually.

Acme.Foundations LLC - The 2026 Coalition Portfolio Outcome

Acme.Foundations entered 2026 with no formal coalition portfolio: historical participation had grown organically across MLCommons, an inactive PAI affiliate membership, and ad-hoc working-group attendance by technical staff without CAIRO oversight. The Q1 2026 CAIRO appointment (lesson 094) brought coalition orchestration into scope; the Q2 2026 portfolio review produced the ratified four-engagement portfolio that the board AI subcommittee approved at the June 2026 meeting. The portfolio applied the four-filter selection framework: enterprise profile (foundation-model provider with EU regulated-sector deployer customer base); budget envelope (€240K annual + 1.5 FTE designated-staff capacity); influence-objective fit (Article 56 Code template shaping + industry-baseline shaping + multi-stakeholder credibility); IP-disclosure tolerance (medium, willing to discuss operational practice patterns and public-committed posture but not proprietary model architecture or training-data composition).

The four ratified engagements:

• Engagement 1 - Article 56 Code Transparency Working Group. CAIRO direct involvement (not delegated) given strategic centrality. Sub-working-group on training-data summary templates as primary participation (matches Acme's June 1, 2026 Transparency + Copyright signature commitment per lesson 095). Monthly cadence; quarterly chapter consolidation; deliverable contribution to the Q3 2026 Transparency template revision cycle. Designated backup: VP Compliance Engineering. Budget allocation: €0 direct cost (Code participation requires no membership fee for signatories); ~0.30 FTE CAIRO time + 0.20 FTE compliance engineering.
• Engagement 2 - MLCommons Responsible AI Working Group. Designated representative: Director, AI Evaluation Engineering (named individual with AIGC-ratified mandate). Participation in AILuminate benchmark rounds + Responsible AI Working Group taxonomy contributions. Monthly cadence; biannual benchmark publication participation. Budget allocation: €50K General Member fee + 0.35 FTE designated representative.
• Engagement 3 - BSA AI Policy Working Group. Designated representative: VP Policy & Government Affairs. Enterprise-deployer subgroup chartered to draft model contract language for AI procurement. Monthly working-group + biannual member summit. Budget allocation: €90K BSA membership tier + 0.25 FTE designated representative.
• Engagement 4 - AI Verify Foundation cross-Atlantic assurance pilot. Designated representative: Senior Director, Audit & Assurance. Participation in the Global AI Assurance Pilot bridging Singapore IMDA framework with EU and US assurance approaches; aligned with Acme's ISO 42001 certification path (lesson 077). Quarterly steering + monthly working-group. Budget allocation: €40K membership tier + 0.20 FTE designated representative + €30K external assurance-engagement support.

The portfolio explicitly excluded four coalitions on documented rationale: Partnership on AI (deferred to 2027 portfolio reconsideration given budget envelope; civil-society engagement pursued through alternative bilateral channels in 2026); ITI (BSA AI Policy Working Group provides sufficient enterprise-deployer policy advocacy for 2026; ITI revisited if state-law engagement intensity rises); Business at OECD (BIAC) (national-federation membership preserved at minimal cost; AI Committee participation deferred to 2027); Frontier Model Forum (Acme's models below the FMF eligibility threshold in Q1 2026; revisit on the Q1 2027 model crossing 10^25 FLOPs that triggers the lesson 095 board ratification on Safety & Security chapter signature); AI Alliance (open-source posture limited to selected Apache 2.0 model releases; AI Alliance engagement revisited if open-source strategy expands).

The 12-element engagement standard was ratified at the same AIGC meeting that approved the portfolio. Each of the four engagements has a 12-element record in the CAIRO standard document, reviewed annually. The quarterly AIGC coalition-signal briefing began at the Q3 2026 AIGC meeting; the first briefing reported: Engagement 1, Acme contribution to the Transparency template sub-working-group on training-data summary publication format adopted into the Q3 2026 working-group draft for the Code revision cycle; Engagement 2, AILuminate Round 4 participation completed with Acme results cited in two academic publications and one MLCommons quarterly report; Engagement 3, BSA model-contract-language draft incorporated Acme-specific proposals on Annex XII downstream-deployer information references that strengthen Acme's procurement DDQ positioning; Engagement 4, AI Verify pilot Q3 milestone delivered with Acme as one of three foundation-model providers in the cross-Atlantic cohort, with Singapore IMDA engagement opening APAC deployer-customer conversations through the firm's bilateral channel.

The €240K annual budget envelope held; the 1.5 FTE designated-staff capacity tracked within ±10%; the anti-trust safe-harbor protocol activated twice (MLCommons benchmark pricing discussion that was redirected to public benchmark-publication context; BSA member summit panel on enterprise model-access licensing that was reframed to public-procurement contract language). Both incidents were documented in CAIRO records and reported to AIGC + GC; neither produced material exposure. The Q1 2027 annual portfolio review (calendared per element 12 of the engagement standard) will reconsider the Frontier Model Forum on the systemic-risk threshold crossing, Partnership on AI on the broadened civil-society engagement objective, and AI Alliance on the open-source posture decision.

The Acme.Foundations 2026 coalition portfolio exemplifies the L5 CAIRO orchestration pattern: four ratified engagements with documented selection rationale; €240K budget envelope held to with quarterly AIGC visibility; 12-element engagement standard enforced per engagement; anti-trust safe-harbor discipline operated continuously; ROI tracked across the five long-horizon indicators with the citation, regulator-engagement, and DDQ-positioning evidence accumulating quarter over quarter. The portfolio complements lesson 101's regulator-engagement program and lesson 103's NIST/CAISI/AISIC engagement to produce a coherent multi-rail policy posture that the CAIRO can defend to the board, the regulator, the auditor, and the buyer as deliberate orchestration rather than passive coalition shopping.

Key Takeaways

• The 2026 AI governance environment runs on three reinforcing rails, regulators, standards bodies, and industry coalitions, and an L5 enterprise CAIRO who orchestrates only the first two has a visible coherence gap. The CAIRO owns the coalition portfolio because no other role (CAIO, GC, CPO, AIGC) is chartered to orchestrate the multi-coalition portfolio against the firm's policy posture.
• The Article 56 Code working-group structure organizes around three chapters (Transparency, Copyright, Safety & Security) with sub-working-groups under each. Signatory firms participate via CAIRO direct involvement or designated technical staff; deliverables flow to the AI Office for incorporation into the ~18-month Code revision cycle.
• The eight 2026 industry coalitions: (1) Partnership on AI (multi-stakeholder including civil society); (2) MLCommons (MLPerf + AILuminate benchmarks + Responsible AI WG); (3) BSA AI Policy Working Group (enterprise-deployer policy); (4) ITI (broader tech-sector advocacy); (5) Business at OECD/BIAC (OECD AI Principles implementation); (6) AI Verify Foundation (Singapore-led assurance frameworks); (7) Frontier Model Forum (Anthropic/OpenAI/Google/Microsoft + ~10 others, foundation-model-provider-only); (8) AI Alliance (IBM/Meta + 50+ open-innovation members).
• The coalition selection framework: four sequential filters, (a) enterprise profile alignment; (b) budget envelope (€240K-€500K typical 2026 enterprise CAIRO budget accommodates 3-5 active engagements); (c) influence-objective fit (regulator template + industry baseline + multi-stakeholder credibility); (d) IP-disclosure tolerance.
• The 12-element CAIRO engagement standard: (1) scope; (2) membership criteria; (3) designated representatives; (4) time commitment; (5) IP rules; (6) anti-trust safe-harbor; (7) public-position approval; (8) conflict-of-interest declarations; (9) financial contribution; (10) withdrawal triggers; (11) reporting to AIGC; (12) annual review. Standard document mapped to ISO 42001 Clause 5 / Annex A.3 and NIST AI RMF Govern 1.5 + 6.1.
• Anti-trust safe-harbor is the rare direct-exposure dimension. Sherman Act Section 1 + EU Article 101 TFEU + UK Competition Act prohibit coordinated pricing, output, market-allocation, boycott discussions. Five-element protocol: published agenda; antitrust counsel present or on-call for pricing-adjacent agenda; minutes with safe-harbor confirmation; exit protocol; post-meeting CAIRO briefing.
• ROI quantification across five long-horizon indicators: (1) co-authored guidance citations; (2) regulator-engagement quality; (3) talent attraction and retention; (4) buyer trust signaling in procurement DDQs; (5) supply-chain partnership leverage. Reported quarterly to AIGC, annually to board AI subcommittee.
• Seven common 2026 coalition mistakes: (1) joining too many (dilution); (2) joining for marketing without contribution (free-rider perception); (3) staff participation without executive sponsorship; (4) public positions not pre-cleared; (5) IP leakage in working-group discussions; (6) anti-trust violations; (7) ignoring Frontier Model Forum if a foundation-model provider (visibility gap).
• Acme.Foundations 2026 portfolio: 4 ratified engagements (Article 56 Code Transparency WG with CAIRO direct involvement; MLCommons Responsible AI WG; BSA AI Policy WG; AI Verify Foundation cross-Atlantic assurance pilot); €240K annual budget; 1.5 FTE designated-staff capacity; quarterly AIGC briefing; explicit deferral rationale for PAI/ITI/BIAC/FMF/AI Alliance with Q1 2027 revisit triggers.
• Cross-walks: EU AI Act Article 56 (Code of Practice); Partnership on AI Tenets; MLCommons benchmark suite; OECD AI Principles + Recommendation; BSA AI Policy Framework; NIST AI RMF Govern 1.5 + 6.1 (coalitions as governance evidence); ISO/IEC 42001:2023 Clause 5 + Annex A.3 (leadership engagement). Penalty exposure indirect (Article 99(7) good-faith effort signal) with rare direct exposure via anti-trust missteps.