IEI / Tokenfrastructure — Original Proposal
IEI_RFP09_proposal_TKNFRA.pdf
PROPOSAL CARDANO PRODUCT COMMITTEE
RESEARCH GRANTS.
These grants are designed to support focused research initiatives that help Cardano make better product, market, ecosystem, and roadmap decisions.
RFP 09 - AI Commercial Positioning.
An institutional, evidence-based assessment of whether AI can become
a real adoption vertical for Cardano, tested at its sharpest point: the
integrity, auditability, and governability of autonomous agents that move
value.
Field Entry
Proposer Intelligence Economy Institute (IEI), part of Tokenfrastructure (TKNFRA)
Primary contact Zakaryae Boudi - boudi@tokenfrastructure.com
RFP applied for RFP 09 - AI Commercial Positioning (single initiative)
Proposal title Where the Story Holds - AI Commercial Positioning for Cardano
Requested budget (modular, scenario-based - see Section 8)
76,000 ADA core scope; up to 100,000 ADA including optional add-ons.
Proposed duration 14 weeks from contract start
Conflicts of interest No direct Cardano conflicts; potential perceived conflict disclosed in Section 10.
TKNFRA | RFP 09 proposal Page 1
Executive summary
The Intelligence Economy Institute (IEI), the research arm of Tokenfrastructure (TKNFRA), studies the convergence of autonomous AI and tokenized value; where software agents do not merely advise but transact, settle, and bear consequence. Our team also founded and leads the Tokenized Economies Institute (TEI), a public–private research platform that convenes financial institutions, regulators, researchers, and technologists. We bring to this commission exactly the three capabilities RFP 09 is hardest to satisfy: **institutional buyer **access, **formal-verification **engineering credibility, and a maturing academic framework on AI control in finance.
RFP 09 sets a deliberately hard test: not whether Cardano can tell an AI story, but whether AI can become a real adoption vertical with identifiable buyers, delivery partners, revenue pathways, and on-chain activity. Most blockchain "AI positioning" work fails that test because it argues from technical properties outward ("we have formal methods, therefore we fit AI") instead of from buyer behaviour inward. This study is built to fail honestly where the story does not hold, and to commit only where the evidence supports a position the committee can defend in front of a sceptical institutional buyer.
Our central hypothesis, stated up front so it can be tested rather than assumed, is that Cardano's credible AI wedge is not the decentralised-intelligence layer (model markets, GPU compute, data marketplaces), where Bittensor, the Artificial Superintelligence Alliance, Render and NEAR already hold the narrative and the liquidity. It is the narrower, more defensible layer of institutional-grade agent integrity at the moment of action:** the guarantee that when an autonomous agent executes a consequential transaction (e.g. a payment, a settlement, a tokenized-asset transfer), what it did was authorised, bounded, deterministic, and auditable before the fact rather than after. **
The agent-payments protocol war of early 2026 (Google AP2, Coinbase x402, Visa TAP, the OpenAI–Stripe ACP, Mastercard Agent Pay) has made this the most contested layer in AI commerce, and it is precisely the layer where Cardano's eUTXO determinism, formal-methods heritage, and native auditability translate into a checkable product property rather than a slogan.
Cardano is not starting from zero. Masumi (Serviceplan Group, NMKR, the Cardano Foundation), with named enterprise pilots, already runs agent identity, escrow-based agent-to-agent payment, on-chain decision logging, and stablecoin settlement on Cardano; a live beachhead in exactly this wedge, and the right object of study rather than a slide.
The work delivers every decision output RFP 09 names, and adds what most positioning consultancies cannot: an engineering-grade verification of the technical claims themselves, performed by a team that has published formal verification of financial smart
TKNFRA | RFP 09 proposal Page 2
contracts and is currently studying AI control in tokenized finance with Université Paris Dauphine-PSL.
1 The proposing institute - capacity to execute
RFP 09 will be won or lost on three things a generic research vendor cannot easily assemble: access to the institutional buyers whose evidence the brief demands; the engineering depth to judge technical claims rather than repeat them; and a method mature enough to produce a defensible go/no-go. IEI brings all three, with a documented track record.
**1.1 Institutional research access and track record **
The buyers, risk functions, and delivery partners RFP 09 requires us to reach are reachable because the team already works with them. Through TEI and TKNFRA, the team has produced institution-grade research in direct collaboration with named financial and public institutions, including:
•** Caisse des Dépôts (CDC): **multi-chapter research on scalability and interoperability in tokenization systems.
•** CDC × Université Paris Dauphine–PSL:** research on what tokenization can do for development finance.
•** Casper Association:** a B-method framework for smart-contract verification applied to ACTUS financial contracts (see 1.3).
•** École des Ponts Business School:** a one-year research program on tokenization for sustainable finance.
•** Broader institutional engagement:** the team's track record includes work with SWIFT, Caisse des Dépôts, Crédit Agricole, and other institutional players.
**1.2 An AI-control framework already in development **
IEI is currently conducting a study with the Chaire Fintech of Université Paris Dauphine–PSL on AI control in tokenized finance, focusing on the governance, constraint, and verification of autonomous agents acting on financial systems. This provides relevant academic grounding for the proposed work, while the Cardano engagement remains a distinct commercial positioning study, focused on buyer demand, competitive positioning, adoption pathways, and strategic decision support.
**1.3 Formal-verification depth: the credibility differentiator **
The hardest task in RFP 09 is separating a real technical edge from a sayable one; deciding whether "formal methods" is a product property or a reputation. That judgement requires people who do formal verification, not people who cite it. The team has published a B-method framework for verifying financial smart contracts (the ACTUS
TKNFRA | RFP 09 proposal Page 3
standard), the same family of methods used to verify safety-critical systems such as metro signaling.
This is combined with detailed architectural work on the control layer between a large language model and a deterministic execution rail in agentic payments: the model proposes, but does not execute; a deterministic layer resolves and checks actions; a verified mediator is the sole approval authority; mandates are enforced as hard constraints; and every transition is auditable before the fact. This lets us audit Cardano's AI claims against working reference artifacts. It is the basis for the credibility teardown in Workstream C.
**1.4 Codified institutional method **
The team's methodology for assessing institutional capability (talents, end-to-end tooling, capabilities, and track records) is codified in a forthcoming book under a signed contract with Palgrave Macmillan: Digital Asset Strategy for Financial Institutions - Talents, End-to-End Tooling, Capabilities, and Track Records. RFP 09's questions about adoption pathways, delivery partners, and production readiness are exactly the questions that book's framework is built to answer; this commission is an applied instance of a discipline we have already formalised.
2 Understanding of the brief
RFP 09 sits in a portfolio designed to close evidence gaps and enable concrete decisions. Three features of the brief shape the study:
•** The test is commercial, not technical. **Technical fit (formal methods, deterministic processing, privacy) is the assumption to interrogate, not the conclusion. The deliverable stands on identifiable buyers, delivery partners, revenue pathways, and on-chain activity.
•** The comparison is against specialised AI chains. **The committee wants to know where Cardano wins, loses, and should not compete versus chains purpose-built for AI.
•** The output must be a defensible go/no-go. **A conditional recommendation is acceptable only if the conditions are explicit, observable, and tied to a funding trigger.
We read the brief as asking: is there a version of "Cardano for AI" that an institutional buyer would sign for, a delivery partner would staff for, and the chain would show transactions for, and if so, exactly which version, for whom, and on what evidence?
3 The positioning hypothesis we will test
TKNFRA | RFP 09 proposal Page 4
Good positioning research starts from a falsifiable claim. Ours has three parts, each of which the methodology in Section 4 is designed to confirm or break.
**3.1 Where Cardano does not have a credible AI position **
On current evidence the decentralised-intelligence layer is likely difficult for Cardano to contest economically, given the existing mindshare, developer gravity, and liquidity of specialised AI networks already held by players like Bittensor (model/subnet markets), the Artificial Superintelligence Alliance, Render and Akash (GPU compute), and NEAR (AI-native UX). A Cardano "we also do decentralised AI compute" narrative would likely be a weak follower story with limited defensible advantage. The study will test whether this should become an explicit do-not-compete conclusion, which we expect.
**3.2 The wedge where it plausibly does **
The defensible wedge is institutional-grade agent integrity, accountability, and settlement: the layer that answers "can we prove this agent was allowed to do that, within these bounds, and reconstruct exactly what happened?" Here Cardano's genuine properties become commercially legible:
•** Deterministic execution (eUTXO). **Transaction outcomes are knowable before submission, the property an institutional risk function actually wants behind an autonomous agent, and the property Masumi already cites.
•** Formal-methods heritage. **Credible only when it attaches to a checkable artifact (a verified mediator, a specified mandate check) rather than to reputation. We test exactly that attachment point.
•** Native auditability and identity. **On-chain decision logging and per-agent identity map directly to the accountability and compliance requirements institutions and regulators impose on agent deployments.
The timing is the reason this matters now. In early 2026 every major payments platform shipped an agent-payment protocol, and the shared primitive, AP2 "mandates," signed agent identity, scoped single-use tokens, is precisely a constraint-and-authorisation layer around an LLM that must be enforced deterministically and proven after the fact. The named open problems in those protocols (agent identity, mandate revocation, compliance enforced at execution time) are the wedge restated. Cardano does not need to win "AI"; it needs to be the most credible place to enforce and audit agent mandates for institutions that cannot afford to be wrong.
**3.3 Why IEI can test this credibly **
Distinguishing a real edge from a sayable one requires institutional buyer access (Section 1.1), formal-verification depth (1.3), and an AI-control framework (1.2), the three assets that turn this hypothesis from an opinion into a tested finding. Workstream C audits each
TKNFRA | RFP 09 proposal Page 5
technical claim against working reference artifacts; Workstream A tests the wedge against the buyers who would actually sign.
4 Methodology
Four workstreams run partly in parallel. Each produces one or more of the RFP's named decision outputs (mapped in Section 5). The method is evidence-first: every positioning claim in the final report is traceable to an interview, an on-chain measurement, a competitive teardown, or a documented protocol fact.
**Workstream A: Demand and institutional buyer pipeline **
Who would actually buy "Cardano for agent integrity," and what evidence do they require before signing?
• 15–25 structured interviews across institutional buyers (regulated-finance risk, payments, settlement, procurement), AI-agent builders and platforms, and prospective delivery partners, drawn from IEI/TEI's institutional network and from existing Cardano AI deployments and their pilot clients, plus a non-Cardano control set to avoid an echo chamber.
• Per-segment evaluation map: decision criteria, trust signals, blocking objections, and the specific evidence threshold each segment needs before taking Cardano seriously.
• Demand pipeline: a ranked, named (where permitted) list of sectors and use cases with realistic 12–24 month conversion potential, distinguishing genuine pull from polite interest.
**Workstream B: Competitive position **
Where does Cardano win and lose versus specialised AI chains, by layer?
• Layered competitive teardown - intelligence/model markets (Bittensor, ASI), compute (Render, Akash), agent coordination/identity, and agent settlement/integrity - scoring Cardano per layer against a repeatable rubric.
• Explicit win / lose / do-not-compete verdict per layer, with the operator- or buyer-attributed reason in each case.
• Agent-payment protocol landscape (AP2, x402, Visa TAP, ACP, Mastercard Agent Pay) mapped against Cardano's settlement and mandate-enforcement story: complement, substitute, or irrelevant.
**Workstream C: Technical-fit credibility teardown **
Do the technical claims survive contact with people who do formal verification?
TKNFRA | RFP 09 proposal Page 6
• Claim-by-claim audit of Cardano's AI-relevant properties (determinism, formal methods, privacy, auditability) against working reference artifacts, for each: real, conditionally real (and on what condition), or narrative-only.
• Reference deployment review (Masumi and comparable systems): what is genuinely enforced on-chain today, what is logged after the fact, and what is still aspirational.
• Output: a credibility ledger grading each claim and supplying the defensible phrasing the committee can use with each audience, and the phrasings to retire.
**Workstream D: Narrative, go/no-go, and action pathway **
Which positioning holds up, what is the recommendation, and what happens next?
• Three to four candidate positioning statements, each stress-tested against institutional, developer, and ecosystem audiences using the Workstream A base.
• Go / no-go / conditional-go recommendation for AI as a strategic Cardano vertical, with conditions that are explicit, observable, and tied to funding triggers.
• Research-to-action pathway with impact tiers, from narrative credibility, through signed pilot, to recurring on-chain transactions, naming the next 2–3 engagements worth funding and the KPI each must hit.
TKNFRA | RFP 09 proposal Page 7
**Evidence access, ethics, and data handling **
The research will combine public source review, source-dated desk research, structured interviews, expert calls, and Cardano/non-Cardano ecosystem evidence. Respondents will be informed of the purpose of the research, who the work is for, how their input may be used, and whether their comments are attributable, anonymized, or confidential. Recording will only occur with consent. Raw interview notes and commercially sensitive information will not be published. The final report will distinguish public findings, confidential findings, source-backed claims, and vendor judgment. Any evidence that cannot be independently inspected or cited publicly will be labelled with the appropriate limitation and confidence level.
5 Deliverables, mapped to RFP 09 decision outputs
Every decision output named in RFP 09 maps to a concrete deliverable and the workstream that produces it.
RFP 09 decision output IEI deliverable WS
AI commercial positioning assessment
Positioning report: where Cardano has a credible story, by layer, and where it does not
B, C
Audience-specific credibility map Per-segment map of claims that land, evidence required, objections — with approved phrasing
A, C
AI adoption pathway Buyer segment × use case × technical requirement × delivery partner × revenue/transaction pathway × go/no-go threshold
A, D
Narrative options and stress-tests 3–4 positioning statements tested against institutional, developer, ecosystem audiences
D
Go/no-go recommendation Decision memo with explicit, observable conditions and funding triggers
D
Research-to-action pathway with impact tiers
Next-engagement shortlist, impact tiers from narrative credibility to sustained on-chain activity
A, D
Final artifacts:
(1) a decision memo;
(2) a full evidence report with credibility ledger, competitive teardown, current activity evidence review, interview synthesis, and limitations;
(3) adoption-pathway, audience-credibility, and use-case prioritization matrices as working spreadsheets;
(4) a public summary;
(5) a 30-minute readout with Q&A.
TKNFRA | RFP 09 proposal Page 8
(6) A short cross-RFP handoff memo identifying findings relevant to adjacent Product Research Initiative workstreams, such as stablecoins/payments, enterprise/RWA, delivery partners, brand, L2/interoperability, customer segmentation, or broader use-case strategy.
6 Workplan and timeline
Fourteen weeks from contract start, with a mid-point checkpoint so the committee can redirect before synthesis.
Phase Weeks Milestone / gate
Inception & scoping 1–2 Confirmed interview targets, rubric, claim list; kickoff with committee
Evidence gathering (WS A–C) 3–8 Interviews complete; competitive teardown drafted; credibility ledger v1
Mid-point checkpoint 8 Interim findings; committee may redirect scope
Synthesis (WS D) 9–12 Positioning options, stress-tests, draft go/no-go
Review & delivery 13–14 Final memo, evidence report, matrices, readout
7 Team
The engagement is staffed from IEI/TKNFRA and TEI, combining institutional research, formal-methods engineering, AI-control expertise, and ecosystem analysis. Further team credentials are provided in the appendix.
•** Engagement lead: Zakaryae Boudi.** Leads overall research quality, strategic synthesis, and the final decision memo. Brings experience across formal verification, AI systems control, institutional tokenization, and public–private research.
•** Formal-methods / technical lead: Zakaryae Boudi.** Owns Workstream C and the technical-fit credibility teardown, including assessment of Cardano claims around determinism, formal methods, auditability, identity, and agent-payment integrity.
•** Institutional research lead: Jiulin Teng.** Owns Workstream A, including institutional interview design, buyer-need mapping, respondent analysis, and synthesis of proof requirements, objections, and adoption pathways.
•** Ecosystem and market research lead: Sara Kawas.** Owns the source-dated market scan, Cardano-side activity review, competitive landscape, and Cardano-side value analysis across AI-agent, payment, identity, auditability, and settlement use cases.** **
8 Budget and cost efficiency
TKNFRA | RFP 09 proposal Page 9
We present a modular budget: a focused core scope plus optional add-ons.
Component Effort Indicative (ADA)
Core: WS A–D, all named decision outputs, memo + report + readout
≈10 weeks 76 000 ADA
Add-on 1: extended institutional interview base (+10) 1-2 weeks 8 000 ADA
Add-on 2: deep verification of one named deployment
1-2 weeks 8 000 ADA
Add-on 3: 6-month follow-on tracking of go/no-go conditions
1-2 weeks 8 000 ADA
Total (Core) ≈10 weeks 76 000 ADA
Total (Core + add-ons) ≈14 weeks 100 000 ADA
Cost-efficiency rationale: IEI's existing institutional access, formal-methods capability, and AI-control framework remove the largest discovery costs in this kind of study, building buyer access and enough engineering context to judge the claims, so the core scope is delivered lean. We recommend funding the core first and triggering add-on 3 only on a conditional go.
9 Risks and mitigations
Risk Mitigation
Interview access to institutional buyers
Anchor on IEI/TEI's existing relationships and Cardano deployment pilot clients; offer anonymity; non-Cardano control set; target list at 2× the needed sample
Confirmation bias toward IEI's own frameworks/stack
Published grading rubric; non-Cardano control interviews; recommendation framed around the wedge, not a vendor; method transparency in the appendix; COI disclosed (Section 10)
Fast-moving market: protocols shift mid-study
Competitive section dated and versioned; conclusions stated as conditions, not snapshots; mid-point checkpoint absorbs material changes
A conditional go the committee cannot act on
Every condition observable and tied to a funding trigger and an owner; no open-ended conditions
Perceived overlap with the Paris Dauphine study
Clear scope separation: the Dauphine work provides academic grounding, while this engagement produces Cardano-specific commercial research, evidence, and recommendations.
10 Conflicts of interest
TKNFRA | RFP 09 proposal Page 10
No direct Cardano conflicts; potential perceived conflict from TKNFRA's broader work in tokenization/agent infrastructure is disclosed and managed through independent evidence standards.
Appendix: Team Biographies
Team Biographies
**Zakaryae Boudi, PhD **Zakaryae Boudi, PhD is an engineer, entrepreneur, and researcher whose work sits at the intersection of formal verification, artificial intelligence, blockchain technology, and institutional finance. He leads initiatives dedicated to designing advanced architectures for programmable financial systems, while also advancing public–private research efforts focused on the future of digital and tokenized economies. His work further extends to the study of machine-driven economic systems and the verification of agent-based, autonomous infrastructures. He has founded and led ventures specializing in institutional-grade tokenization, and has been involved in the development of advanced technological frameworks for financial and economic applications. Earlier in his career, he collaborated on projects with major international institutions, including multilateral organizations and public-sector bodies.
He holds an engineering degree from École Mohammadia d'Ingénieurs, a Master's degree from Sciences Po Paris, and a PhD focused on formal methods for complex and safety-critical systems.
**Jiulin Teng, PhD **Jiulin Teng is a strategy architect, entrepreneur, and researcher with experience spanning advisory, venture creation, and systems design. He has committed to the development of ventures at the intersection of blockchain, artificial intelligence, and information systems for the last decade, being one of the early proponents of tokenization. He combines a strong foundation in strategy, economics, game theory, and evolutionary biology with hands-on expertise in AI training, database architecture, and full-stack development, focusing on the application of structured frameworks to complex technological and organizational challenges. Since the early days of his career, Jiulin has consulted clients on corporate and business-unit strategy; he continues to conduct theoretical research in strategy.
He holds a PhD in Strategic Management and Business Policy from HEC Paris, a Master's degree in Bioentrepreneurship from Karolinska Institute, and a Bachelor's degree in Biological Sciences from Nankai University.
**Sara Kawas **Sara Kawas is an entrepreneur working on blockchain infrastructure, tokenized finance, and development finance. She is Growth & Research Project Manager at the Tokenized Economies Institute and previously worked at FeverTokens on financial standards, tokenization protocols, and institutional research initiatives.
TKNFRA | RFP 09 proposal Page 11
She is also co-founder of DebLeb and Athareum, two initiatives focused on programmable financial infrastructure for fragile economies. Her earlier experience includes blockchain startups, the Permanent Mission of Lebanon to the United Nations in Geneva, where she explored humanitarian blockchain use cases, and CMA CGM, where she worked on blockchain applications for trade finance and supply chains. She studied Economics and Business, with a focus on international economics, at Université Paris-Panthéon-Assas.
TKNFRA | RFP 09 proposal Page 12
CARDANO PRODUCT COMMITTEE RESEARCH GRANTS.
These grants are designed to support focused research initiatives that help Cardano make better product, market, ecosystem, and roadmap decisions.
RFP 09 - AI Commercial Positioning.
An institutional, evidence-based assessment of whether AI can become a real adoption vector for Cardano, tested at its sharpest point: the integrity, auditability, and governability of autonomous agents that move value.
| Field | Entry |
|---|---|
| Proposer | Intelligence Economy Institute (IEI), part of Tokeninfrastructure (TKNFRA) |
| Primary contact | Zakaryae Boudi - boudi@tokenfrastructure.com |
| RFP applied for | RFP 09 - AI Commercial Positioning (single initiative) |
| Proposal title | Where the Story Holds - AI Commercial Positioning for Cardano |
| Requested budget (modular, scenario-based - see Section 8) | 76,000 ADA core scope; up to 100,000 ADA including optional add-ons. |
| Proposed duration | 14 weeks from contract start |
| Conflicts of interest | No direct Cardano conflicts; potential perceived conflict disclosed in Section 10. |
Executive summary
The Intelligence Economy Institute (IEI), the research arm of Tokenfrastructure (TKNFRA), studies the convergence of autonomous AI and tokenized value; where software agents do not merely advise but transact, settle, and bear consequence. Our team also founded and leads the Tokenized Economies Institute (TEI), a public-private research platform that convenes institutional, regulators, researchers, and technologists. We bring to this commission exactly the three capabilities RFP 09 is hardest to satisfy: institutional buyer access, formal-verification engineering credibility, and a maturing academic framework on AI control.
RFP 09 sets a deliberately hard test: not whether Cardano can tell an AI story, but whether AI can be a real adoption vehicle with identifiable buyers, delivery partners, revenue pathways, and on-chain activity. Most blockchain "AI positioning" work fails that test because it argues from technical properties outward ("we have formal methods, therefore we fit AI") instead of from buyer behaviour inward. This study is built to fail honestly where the story does not hold, and to commit only where the evidence supports a position the committee can defend in front of a sceptical institutional buyer.
Our central hypothesis, stated up front so it can be tested rather than assumed, is that Cardano's credible AI wedge is not the decentralised-intelligence layer (model markets, GPU compute, data marketplaces), but Bittensor, the Artificial Superintelligence Alliance, Render and NEAR already hold the narrative and the liquidity. It is the narrower, more defensible layer of institutional-grade agent integrity at the moment of action: the guarantee that when an autonomous agent executes a consequential transaction (e.g. a payment, a settlement, a tokenized-asset transfer), what it did was authorised, bounded, deterministic, and auditable before the fact rather than after.
The agent-payments protocol war of early 2026 (Google AP, Coinbase x402, Visa TAP, the OpenAI-Stripe ACP, Mastercard Agent Pay) has made this the contested arena in AI commerce, and it is precisely the layer where Cardano's eUTXO determinism, formal-methods heritage, and native auditability translate into a checkable product property rather than a slogan.
Cardano is not starting from zero. Masumi (Serviceplan Group, NMKR, the Cardano Foundation), with named enterprise pilots, already runs agent identity, escrow-based agent-to-agent payment, on-chain decision logging, and stablecoin settlement on Cardano; a live beachhead in exactly this wedge, and the right object of study rather than a slide.
The work delivers an every decision output RFP 09 names, and adds what most positioning consultants cannot: an engineering-grade verification of the technical claims themselves, performed by a team that has published formal verification of financial smart
contracts and is currently studying AI control in tokenized finance with Université Paris Dauphine-PSL.
1 The proposing institute - capacity to execute
RFP 09 will be won or lost on three things a generic research vendor cannot easily assemble: access to the institutional buyers whose evidence the brief demands; the engineering depth to judge technical claims rather with than repeat them; and a method mature enough to produce a digestible go/no-go. IEI brings all three, with a documented track record.
1.1 Institutional research access and track record
The buyers, risk functions, and delivery partners RFP 09 requires us to reach are reachable because the team already works with them. Through TEI and TKNFRA, the team has produced institution-grade research in direct collaboration with named financial and public institutions, including:
- Caisse des Dépôts (CDC): multi-chapter research on scalability and interoperability in tokenization systems.
- CDC x Université Paris Dauphine-PSL: research on what tokenization can do for development finance.
- Casper Association: a B-method framework for smart-contract verification applied to ACTUS financial contracts (see 1.3).
- École des Ponts Business School: a one-year research program on tokenization for sustainable finance.
- Broader institutional engagement: the team's track record includes work with SWIFT, Caisse des Dépôts, Crédit Agricole, and other institutional players.
1.2 An AI-control framework already in development
IEI is currently conducting a study with the Chaire Fintech of Université Paris Dauphine-PSL on AI control in tokenized finance, focusing on the governance, constraint, and verification of autonomous agents acting on financial systems. This provides relevant academic grounding for the proposed work, while the Cardano engagement remains a distinct commercial positioning study, focused on buyer demand, competitive positioning, adoption pathways, and strategic decision support.
1.3 Formal-verification depth: the credibility differentiator
The hardest task in RFP 09 is separating a real technical edge from a sayable one; deciding whether 'formal methods' is a product property or a reputation. Judgement requires people who do formal verification, not people who cite it. The team has published a B-method framework for verifying financial smart contracts (the ACTUS
technical claim against working reference artifacts; Workstream A tests the wedge against the buyers who would actually sign.
4 Methodology
Four workstreams run partly in parallel. Each produces one or more of the RFP's named decision outputs (mapped in Section 5). The method is evidence-first: every positioning claim in the final report is traceable to an interview, an on-chain measurement, a competitive teardown, or a documented protocol fact.
Workstream A: Demand and institutional buyer pipeline
Who would actually "Cardano for agent integrity," and what evidence do they require before signing?
- 15–25 introduced interviews across institutional buyers (regulated-finance risk, payments, settlement, procurement), AI-agent builders and platforms, and prospective delivery, drawn from IE/TEI's institutional network and from existing Cardano AI deployments and their pilot clients, plus a non-Cardano control set to avoid an echo chamber.
- Per-segment evaluation map: decision criteria, trust signals, blocking objections, and the specific evidence threshold each segment needs before taking Cardano seriously.
- Demand pipeline: a ranked, named (where permitted) list of sectors and use cases with realistic 12–24 month conversion potential, distinguishing genuine pull from polite interest.
Workstream B: Competitive position
Where does Cardano win and lose versus specialised AI chains, by layer?
- Layered competitive teardown - intelligence/model markets (Bittensor, ASI), compute (Render, Akash), agent coordination/identity, and agent settlement/integrity - scoring Cardano per layer against a repeatable rubric.
- Explicit win / loss / do-not-compete verdict per layer, with the operator- or buyer-attributed reason in each case.
- Agent-payment protocol landscape (AP2, x402, Visa TAP, ACP, Mastercard Agent Pay) mapped against Cardano's settlement and mandate-enforcement story: complement, substitute, or irrelevant.
Workstream C: Technical-fit credibility teardown
Do the technical claims survive contact with people who do formal verification?
Evidence access, ethics, and data handling
The research will combine public source review, source-dated desk research, structured interviews, expert calls, and Cardano/non-Cardano ecosystem evidence. Respondents will be informed of the purpose of the research, who the work is for, how their input may be used, and whether their comments are attributable, anonymized, or confidential. Recording will only occur with consent. Raw interview notes and commercially sensitive information will not be published. The final report will distinguish public findings, confidential findings, source-backed claims, and vendor judgment. Any evidence that cannot be independently inspected or cited publicly will be labelled with the appropriate limitation and confidence level.
5 Deliverables, mapped to RFP 09 decision outputs
Every decision output named in RFP 09 maps to a concrete deliverable and the workstream that produces it.
| RFP 09 decision output | IEI deliverable | WS |
|---|---|---|
| AI commercial positioning assessment | Positioning report: where Cardano has a credible story, by layer, and where it does not | B, C |
| Audience-specific credibility map | Per-segment map of claims that land, evidence required, objections — with approved phrasing | A, C |
| AI adoption pathway | Buyer segment x use case x technical requirement x delivery partner x revenue/transaction pathway x go/no-go threshold | A, D |
| Narrative options and stress-tests | 3-4 positioning statements tested against institutional, developer, ecosystem audiences | D |
| Go/no-go recommendation | Decision memo with explicit, observable conditions and funding triggers | D |
| Research-to-action pathway with impact tiers | Next-engagement shortlist, impact tiers from narrative credibility to sustained on-chain activity | A, D |
Final artifacts:
- a decision memo;
- a full evidence report with credibility ledger, competitive teardown, current activity evidence review, interview synthesis, and limitations;
- adoption-pathway, audience-credibility, and use-case prioritization matrices as working spreadsheets;
- a public summary;
- a 30-minute readout with Q&A.
(6) A short cross-RFP handoff memo identifying findings relevant to adjacent Product Research Initiative workstreams, such as stablecoins/payments, enterprise/RWA, delivery partners, brand, L2/interoperability, customer segmentation, or broader use-plane strategy.
6 Workplan and timeline
Fourteen weeks from contract start, with a mid-point checkpoint so the committee can redirect before synthesis.
| Phase | Weeks | Milestone / gate |
|---|---|---|
| Inception & scoping | 1–2 | Confirmed interview targets, rubric, claim list; kickoff with committee |
| Evidence gathering (WS A–C) | 3–8 | Interviews complete; competitive teardown drafted; credibility ledger v1 |
| Mid-point checkpoint | 8 | Interim findings; committee may redirect scope |
| Synthesis (WS D) | 9–12 | Positioning options, stress-tests, draft go/no-go |
| Review & delivery | 13–14 | Final memo, evidence report, matrices, readout |
7 Team
The engagement is staffed from IEI/TKNFRA and TEI, combining institutional research, formal-methods engineering, AI-control expertise, and ecosystem analysis. Further team credentials are provided in the appendix.
- Engagement lead: Zakaryae Boudi. Leads overall research quality, strategic synthesis, and the final decision memo. Brings experience across formal verification, AI systems control, institutional tokenization, and public-private research.
- Formal-methods / technical lead: Zakaryae Boudi. Owns Workstream C and the technical-fit credibility teardown, including assessment of Cardano claims around determinism, formal methods, auditability, identity, and agent-payment integrity.
- Institutional research lead: Jiulin Teng. Owns Workstream A, including institutional interview design, buyer-need mapping, respondent analysis, and synthesis of proof requirements, objections, and adoption pathways.
- Ecosystem and market research lead: Sara Kawas. Owns the source-dated market scan, Cardano-side activity review, competitive landscape, and Cardano-side value analysis across AI-agent, payment, identity, auditability, and settlement use cases.
8 Budget and cost efficiency
We present a modular budget: a focused core scope plus optional add-ons.
| Component | Effort | Indicative (ADA) |
|---|---|---|
| Core: WS A-D, all named decision outputs, memo + report + readout | ≈10 weeks | 76 000 ADA |
| Add-on 1: extended institutional interview base (+10) | 1-2 weeks | 8 000 ADA |
| Add-on 2: deep verification of one named deployment | 1-2 weeks | 8 000 ADA |
| Add-on 3: 6-month follow-on tracking of go/no-go conditions | 1-2 weeks | 8 000 ADA |
| Total (Core) | ≈10 weeks | 7 6 000 ADA |
| Total (Core + add-ons) | ≈14 weeks | 100 000 ADA |
Cost-efficiency rationale: Iei's existing institutional access, formal-methods capability, and Al-control framework remove the largest discovery costs in this kind of study, building buyer access and enough engineering context to judge the claims, so the core scope is delivered lean. We recommend the core first and triggering add-on 3 only on a conditional go.
9 Risks and mitigations
| Risk | Mitigation |
|---|---|
| Interview access to institutional buyers | Anchor on Iei/TEI's existing relationships and Cardano deployment pilot clients; offer anonymity; non-Cardano control set; target list at 2x the needed sample |
| Confirmation bias toward Iei's frameworks/stack | Published grading rubric; non-Cardano control interviews; recommendation framed around the wedge, not a vendor; method transparency in the appendix; COI disclosed (Section 10) |
| Fast-moving market: protocols shift mid-study | Competitive section and versioned; conclusions stated as conditions, not snapshots; mid-point checkpoint absorbs material changes |
| A condition go the committee cannot act on | Every condition observable and tied to a funding trigger and an owner; no open-ended conditions |
| Perceived overlap with the Paris Dauphine study | Clear scope separation: the Dauphine work provides academic grounding, while this engagement produces Cardano-specific commercial research, evidence, and recommendations. |
10 Conflicts of interest
No direct Cardano conflicts; potential perceived conflict from TKNFRA's broader work in tokenization/agent infrastructure is disclosed and managed through independent evidence standards.
Appendix: Team Biographies
Team Biographies
Zakaryae Boudi, PhD
Zakaryae Boudi, PhD is an engineer, entrepreneur, and researcher whose work sits at the intersection of formal verification, artificial intelligence, blockchain technology, and institutional finance. He leads initiatives dedicated to designing advanced architectures for programmable financial systems, while also advancing public-private research efforts focused on the future of digital and tokenized economies. His work further extends to the study of machine-driven economic systems and the verification of agent-based, autonomous infrastructures. He has founded and led ventures specializing in institutional-grade tokenization, and has been involved in the development of advanced technological frameworks for financial and economic applications. Earlier in his career, he collaborated on projects with major international organizations, including multilateral organizations and public-sector bodies.
He holds an engineering degree from École Mohammadia d'ingénieurs, a Master's degree from Sciences Po Paris, and a PhD focused on formal methods for complex and safety-critical systems.
Jiulin Teng, PhD
Jiulin Teng is a strategy architect, entrepreneur, and researcher with experience spanning advisory, venture creation, and systems design. He has committed to the development of ventures at the intersection of blockchain, artificial intelligence, and information systems for the last decade, being one of the early proponents of tokenization. He combines a strong foundation in strategy, economics, game theory, and evolutionary biology with hands-on expertise in AI training, database architecture, and full-stack development, focusing on the application of structured frameworks to complex technological and organizational challenges. Since the early days of his career, Jiulin has consulted clients on corporate and business-unit strategy; he continues to conduct theoretical research in strategy.
He holds a PhD in Strategic Management and Business Policy from HEC Paris, a Master's degree in Bioentrepreneurship from Karolinska Institute, and a Bachelor's degree in Biological Sciences from Nankai University.
Sara Kawas
Sara Kawas is an entrepreneur working on blockchain infrastructure, tokenized finance, and development finance. She is Growth & Research Project Manager at the Tokenized Economies Institute and previously worked at FeverTokens on financial standards, tokenization protocols, and institutional research initiatives.
She is also co-founder of DebLeb and Athareum, two initiatives focused on programmable financial infrastructure for fragile economies. Her earlier experience includes blockchain startups, the Permanent Mission of Lebanon to the United Nations in Geneva, where she explored humanitarian blockchain use cases, and CMA CGM, where she worked on blockchain applications for trade finance and supply chains. She studied Economics and Business, with a focus on international economics, at Université Paris-Panthéon-Assas.