Grant application : Modern Bureaucracy

[polykrate]

Project Abstract

Modern Bureaucracy brings the first AI-assisted compliance infrastructure to the Polkadot ecosystem, solving the fundamental problem that complex bureaucratic procedures exclude honest participants while advantaging those with resources for administrative compliance. We provide a translation layer between natural language and blockchain operations through local AI assistance, maintaining cryptographic proof of every interaction for regulatory audit requirements.

Our solution combines a specialized Substrate parachain with MCP (Model Context Protocol) servers to enable AI-guided workflows. Building upon our existing proof-of-concept (Human Context Protocol with 3 pallets), this grant will deliver the production-ready Modern Bureaucracy parachain featuring pallet-pki, pallet-crypto-trails, and pallet-compliance. The system will launch with a treasury proposal workflow optimized for AI assistance, demonstrating how honest teams can achieve professional compliance without expensive consultants, while maintaining end-to-end encryption and decentralized coordination.

Grant level

• Level 1: Up to $10,000, 2 approvals
• Level 2: Up to $30,000, 3 approvals
• Level 3: Unlimited, 5 approvals (for >$100k: Web3 Foundation Council approval)

Application Checklist

• The application template has been copied and aptly renamed (project_name.md).
• I have read the application guidelines.
• Payment details have been provided (Polkadot AssetHub (USDC & DOT) address in the application and bank details via email, if applicable).
• I understand that an agreed upon percentage of each milestone will be paid in vested DOT, to the Polkadot address listed in the application.
• I am aware that, in order to receive a grant, I (and the entity I represent) have to successfully complete a KYC/KYB check.
• The software delivered for this grant will be released under an open-source license specified in the application.
• The initial PR contains only one commit (squash and force-push if needed).
• The grant will only be announced once the first milestone has been accepted (see the announcement guidelines).
• I prefer the discussion of this application to take place in a private Element/Matrix channel. My username is: @_______:matrix.org (change the homeserver if you use a different one)

[github-actions]

CLA Assistant Lite bot All contributors have signed the CLA ✍️ ✅

[polykrate]

I have read and hereby sign the Contributor License Agreement.

[polykrate]

recheck

[polykrate]

Hello team

The CLA Assistant bot has confirmed "CLA Assistant Lite bot All contributors have signed the CLA ✍️ ✅",
but the status check CLA Assistant / CLAAssistant (pull_request_target) is still failing.
Tried to repush it by adding an empty line at the end at the application, CLA still fail and enforce label have now an issue with the admin review label.

Could someone with access relaunch the failed run please ?

Thanks,

[Noc2]

Thanks a lot for the application. Do you have any proof or research that shows that your solutions actually help to solve the problem that you are describing (regarding the treasury)? Also, if it helps, why is it necessary to put this on-chain? I would first try to find product-market fit with a centralized solution before I start developing a decentralized solution here.

[polykrate]

Also, if it helps, why is it necessary to put this on-chain? I would first try to find product-market fit with a centralized solution before I start developing a decentralized solution here.

A centralized solution can't provide what regulatory compliance requires: immutable audit trails with cryptographic proof of who did what when. This isn't a "nice-to-have" - it's a legal requirement for EU compliance (why EBSI exists).
The product-market fit isn't in Web3 - it's in bridging Web3↔TradFi compliance. EU institutions are already deploying deconcentrated solution (Hyperledger Besu and Eidas v2) precisely because centralized solutions failed at cross-border compliance coordination. My propoasal provides the missing citizen-facing layer.
TypeForm has existed since 2012. If centralized form-builders could solve regulatory compliance, they would have. The question isn't "does it work?" but "is it legally valid?"
The interest of puting the compliance onchain is the accessibility. Once a compliance is published, everyone can use it through the MCP.
For compliance, decentralization isn't a technical choice - it's a juridical necessity.

Do you have any proof or research that shows that your solutions actually help to solve the problem that you are describing (regarding the treasury)?

The proof is in Polkadot's own 2025 data: 30% of treasury spending flows through intermediary structures (departments, bounties, collectives, State of Polkadot Q1 2025), while rejection rates hit 40% under OpenGov (vs 9% in Gov V1). This isn't a voting problem - it's a structural compliance crisis
With 30% of the $18.6M quarterly spending flowing through intermediary structures, and assuming standard 5-10% management fees, the ecosystem already pays $500k-1M per quarter just for proposal formatting and administration. Add the hidden costs of 40% rejection rates and 2-week resubmission cycles, and the true cost of compliance inefficiency likely exceeds $2M quarterly.
By publishing compliance templates and community priorities on-chain as structured schemas, then using AI-driven maieutic questioning to guide proposers, Modern Bureaucracy transforms natural language into formal proposals with cryptographic proof.
We're automating what intermediaries currently charge $20k to do manually, but for 0.01 DOT per transaction.

P.S. - "Try centralized first" is like asking why we need HTTPS when HTTP works fine. Some problems require cryptographic guarantees from day one. Compliance is one of them.

[Noc2]

Thanks for the quick reply. Do you have any additional (market) research? Why do you think "publishing compliance templates and community priorities on-chain as structured schemas, then using AI-driven maieutic questioning to guide proposers" will improve the treasury spending? To be honest, the link isn't clear to me. In any case, I will share it with the rest of the team.

[Noc2]

Also, any previous work that you can share? Given that your GitHub account has no history: https://github.com/polykrate

[polykrate]

Thanks for the quick reply. Do you have any additional (market) research? Why do you think "publishing compliance templates and community priorities on-chain as structured schemas, then using AI-driven maieutic questioning to guide proposers" will improve the treasury spending? To be honest, the link isn't clear to me. In any case, I will share it with the rest of the team.

I should clarify the mechanism: on-chain, we store CIDs and metadata for RAG alongside compliance schemas (both data and forms). These resources are sent to the AI via MCP to align it with specific tasks. The added value is that the AI becomes a specialized assistant with deep domain knowledge.
For complex multi-dimensional compliance like treasury proposals, the AI acts as an intelligent simplifier - using maieutic questioning to guide proposers through intricate requirements they might otherwise miss or misunderstand. This structured guidance should improve proposal quality and reduce rejection rates due to compliance gaps.
The link to improved treasury spending comes through better-informed proposals that meet compliance requirements from the start, reducing iteration cycles and administrative overhead. However, you're right to question this - I'd be interested in your team's perspective on whether this mechanism would actually deliver measurable improvements in your context.

[polykrate]

Also, any previous work that you can share? Given that your GitHub account has no history: https://github.com/polykrate

You're absolutely right about the GitHub history. I can address this in two ways:

Public repository: I can publish the basic version of the parachain implementation to GitHub to demonstrate the core architecture and approach.
MCP demonstration: The MCP integration is fuctional but still in early development. I'd prefer to show you the current prototype privately first to get your feedback before cleaning it up for public release.

[takahser]

@polykrate thanks for submitting your grant proposal, I have a couple of questions:

Unlike oracle solutions that inject external data on-chain, we provide an infrastructure where humans retain full control and legal responsibility for content while AI handles structural compliance.

Could you elaborate on what you mean by "structural compliance" here?

Small teams can't afford $20k+ for professional compliance writers

Do you have any evidence that compliance-related issues is a driving factor for the rejection of OpenGov proposals?

• https://github.com/polykrate/human-context-protocol (polyvalent prototype)
• https://github.com/polykrate/gosseyn (experimentation)

If these repos don't exist or are private, pls remove them from the application doc.

• https://www.linkedin.com/in/jean-francois-meneust-a488bb121/

Could you fix the LinkedIn link? I'd also be curious to learn more about your background, since there is no github trackrecord.

[polykrate]

If these repos don't exist or are private, pls remove them from the application doc.

• https://www.linkedin.com/in/jean-francois-meneust-a488bb121/
  ** fixed https://www.linkedin.com/in/jean-fran%C3%A7ois-meneust-a488bb121/**
  Could you fix the LinkedIn link? I'd also be curious to learn more about your background, since there is no github trackrecord.

Hello,

Thank you for your time.

I fixed linkedin, the repo and the TPS to be pragmatic.

Let me introduce myself after a university license in History I joined the french army 18 years ago, left it at the end of last year.

I was initially recruited then formed as programmer, did that during 4 years, I really liked programming however it was mostly java enterprise during the start of 2010's. After learning Rust, Frame and Typescript, I think I don't like java.
Because of restructuring (outsourcing to civilan) I passed an exam then did the school to become an officer in finance in logistic, did that during 11 years. The first 7 years were interesting, I was in regiments, doing operations and stuff,
After this few years and since I knew how to use excel and a sql database, I was relocated to do complex administrative et finance task between the defense ministry and the EU, After 3 years,I decided it was time to move one.

I am not very familiar with linkedin or github culture, I am an army veteran with the goal to become a blockchain engineer and develop what I think are sub-exploited primitives of a blockchain, mostly for compliance and audit.
I never contributed to anything in github because I am not familiar with it and know how to write descent code since a few month ago
All my times have been on teaching myself, write code, experiment, then prompt Shawn Tabrizi and Gavin Wood to brutally audit my code as if the future of web 3 depends on it.

I launched my company in january 2025 for entreprise AI automaton, I finished building the prototype on August, a substrate solochain for private deployment, that allow agents to coordinate, send RAG, execute complex computation off chain then report back ,leaving crypto trail, and using shared rag carge.fr.
I want to adapt my blockchain for public deployment and make it useful for all.

[polykrate]

Unlike oracle solutions that inject external data on-chain, we provide an infrastructure where humans retain full control and legal responsibility for content while AI handles structural compliance.

Could you elaborate on what you mean by "structural compliance" here?

The AI acts as an intelligent interface that transforms human input into a format where all the needed data are present and in a usable form.

I restructured the traditional way of doing compliance in a compatible format for AI's RAG :

• the instruction : what the agent has to do, for example interrogate the human to fil the schema,
• the resources : how the agents has to do it and what it know about the task, for example a list of questions to clarify the economic model and metrics to compare to other blockchain,
• the format : a Json schema

For handling the structural compliance, the AI is using a tool, it inject in it the 3 parts of the compliance, the AI then interrogate the humain, cross references its answer with the ressources, and when the Agents has enough human data, it can construct a Json object validate by the schema, show it to the human for validation, then send it onchain.

In case of a workflow, it is a guarantee for Bob that Alice has send all the data needed for him to execute its step.

[polykrate]

Small teams can't afford $20k+ for professional compliance writers

Do you have any evidence that compliance-related issues is a driving factor for the rejection of OpenGov proposals?

It was discussed here https://forum.polkadot.network/t/order-out-of-chaos-a-guide-to-opengov-treasury-proposals/1618, and I'm gonna use the 2023 template in the discussion as a proof :
40+ mandatory fields with precise formatting required
Multiple cross-references (GitHub, IPFS, previous proposals)
3 levels of nested tables (Milestones/Tasks/Deliverables)
Mandatory quantified metrics with specific targets

Rejection pattern analysis ;
Proper milestone structuring
Comprehensive competitive analysis
Detailed budget breakdown
Complete context documentation

My solution allow :
Context research (via web scraping + onchain resources)
Similar solutions analysis (via web scraping + onchain resources)
Budget formatting (personal documents and conversational inputs)
Milestone structuring (personal documents and conversational inputs)

[polykrate]

I checked in web3 syntax what my project is it's about creating RWA in a collaborative way.

I'm moving the pallets from the parachain to the Tanssi.template.