Harmonica: chatbot for LLM-powered deliberation in Cardano DAOs

[GENERAL] Name and surname of main applicant

Artem Zhiganov

[GENERAL] Email address of main applicant

hello@teamharmoni.ca

Additional applicants

• Will Payne
• Felix Grosvenor-Duffy
• Edward Saperia

[GENERAL] Please specify how many months you expect your project to last (from 2-12 months)

3

[GENERAL] Please indicate if your proposal has been auto-translated into English from another language.

No

[GENERAL] Does your project have any dependencies on other organizations, technical or otherwise?

Yes

[GENERAL] If YES, please describe what the dependency is and why you believe it is essential for your project’s delivery. If NO, please write “No dependencies.” .

We are using OpenAI API. Their language model is essential for the synthesis of viewpoints / opinions / preferences at scale. We also use Discord API for obvious reasons.

[GENERAL] Will your project’s output/s be fully open source?

Yes

[GENERAL] If NO, please describe which outputs are not going to be open source. If YES, please write “Project will be fully open source.”

Project will be fully open source.

[METADATA] Category of proposal

DAO

[METADATA] SDG rating

SDG goals:

Promote peaceful and inclusive societies for sustainable development, provide access to justice for all and build effective, accountable and inclusive institutions at all levels

SDG subgoals:

Develop effective, accountable and transparent institutions at all levels

Ensure responsive, inclusive, participatory and representative decision-making at all levels

Key Performance Indicator (KPI):

Proportion of population satisfied with their last experience of public services

Proportion of population who believe decision-making is inclusive and responsive, by sex, age, disability and population group

#proposertoolsdg

[IMPACT] Please describe your proposed solution.

Harmonica aims to make collective decision-making faster, more accessible and higher quality by using large language models to facilitate sensemaking and deliberation at scale. By synthesizing participants' input into votable options, Harmonica can make governance process faster and more inclusive, enabling DAOs on Cardano to coordinate efficiently and engage more members.

Advanced LLMs can synthesise statements pretty well out of the box, but we use sophisticated heuristics to enable true P2P ideation without people actually talking to each other. By doing so, we drastically reduce the time needed to make a decision in a group. And that in itself makes decision-making easier and more fun, and leads to more open deliberation.

Here’s what a session could look like:

• Admin asks a question in a Discord channel using /ask.
• “Where do we want to go to dinner tonight?”
• “We have several projects which we could award grants. [How much grant money we have, amount of money requested from each project, description of each project.] How should we allocate that money?”
• This summons the Harmonica bot, which puts a JOIN button in the channel.
• When a user clicks JOIN, the bot DMs each user and asks for their answer to the question.
• The admin gets notifications when someone joins the session and submits their answer.
• When enough users have responded, the admin asks the bot to synthesize the answers by clicking SUMMARISE. Harmonica prints a summary to the main channel.
• The admin can either start a poll based on the results or ask another question.

We are focused on building a Discord chatbot because it provides a simple conversational UX for DAOs. Harmonica will demonstrate its impact through case studies showing how our solution decreases time to find consensus / common ground and increases participation in governance.

[IMPACT] How does your proposed solution address the challenge and what benefits will this bring to the Cardano ecosystem?

Harmonica directly addresses Cardano's need for "Effective Collaboration Management Platforms to Organize Community Intentions and Actions." Our chatbot solution enables DAOs to improve their governance processes, enabling faster and more inclusive decision making.

Harmonica will benefit the Cardano ecosystem in several key ways:

1. Attract more DAOs to organize on Cardano. By offering an innovative governance solution, Harmonica makes Cardano an attractive choice for DAO operators and communities.
2. Increase mainstream appeal of DAOs and blockchain. As DAOs represent the evolution of human collaboration, tools that improve governance experience (GX) also broaden their appeal to mainstream audiences. Harmonica aims to decrease barriers to DAO participation, supporting Cardano's goal to drive mainstream DAO and blockchain adoption.
3. Open-source software strengthens the ecosystem. By providing Harmonica as open-source infrastructure, we make it accessible for anyone to build upon and help shape Cardano's ecosystem. Open-source software spurs more rapid innovation and gives the ecosystem access to essential tools to sustain growth.
4. Gain competitive advantage. There are currently few options for DAOs to organize and govern on Cardano. Solutions like Harmonica, which provide innovative governance experience, give Cardano a distinct advantage in attracting DAOs over other blockchains. Supporting Harmonica helps ensure Cardano's position as the leading choice for DAOs.

We hope to reach 10 Cardano-based communities shortly after project completion and then scale to 100+

[IMPACT] How do you intend to measure the success of your project?

• Completion of milestones: successful delivery of chatbot into production and execution of awareness campaign
• Number of Cardano-based communities using the tool: aiming for 10+ (as a starting point)
• Frequency of usage per community: aiming for everyday usage

[IMPACT] Please describe your plans to share the outputs and results of your project?

We are planning to publish a series of articles with research findings in RnDAO blog (rndao.mirror.xyz) and our twitter (https://twitter.com/harmonica__bot).

Additionally, we have strong connections with academic institutions in the UK (Nesta, UCL, University of Manchester, Newspeak House, etc) and we plan to publish research findings in their publications as well.

[CAPABILITY/ FEASIBILITY] What is your capability to deliver your project with high levels of trust and accountability?

We're an experienced team of contributors (C-level experience and MSc) in the relevant areas, who previously worked on dozens of successful IT projects and contributed to leading DAOs (Seed Club alumni). We bring the same quality of execution and commitment to this project and are excited to work with the Cardano community.

Our processes include:

• code quality monitoring and QA
• daily tech sync sessions
• bi-weekly sprint planning sessions
• quarterly retrospectives
• a policy of transparency (you're welcome to read our Notion and join our meetings)

[CAPABILITY/ FEASIBILITY] What are the main goals for the project and how will you validate if your approach is feasible?

The main goals for Harmonica are:

1. Build a Discord chatbot solution that synthesizes natural language input from multiple users into votable options for group decision making. We will validate this goal through launching MVP version of the chatbot within 3 months of funding.
2. Make governance more accessible and inclusive for DAOs on Cardano. We will measure progress toward this goal through case studies with early adopter DAOs showing decreased time to reach decisions and increased participation rates after installing Harmonica.
3. Attract new DAOs to organize and govern on Cardano. We will track the number of new DAOs launching within 6 months of our solution going live, with a target of at least 5-10 DAOs in that timeframe.
4. Provide open-source infrastructure for the Cardano ecosystem. By releasing Harmonica's code as open-source, others can build upon and improve our solution, benefiting both Cardano and the broader blockchain community.
5. Gain competitive advantage over other chains in attracting DAOs. Supporting solutions like Harmonica that meet key DAO needs help ensure Cardano's position as the leading choice for DAO organization and governance.

For goals related to technology implementation (Goal 1), we will follow standard agile software development practices, including iterative releases, continuous testing, and feedback incorporation. For qualitative goals (Goals 2-5), we will rely on metrics described above as well as feedback from DAO users and the broader community to determine if our approach is feasible and impactful. While numbers provide an indication of progress, qualitative feedback offers a more comprehensive view of Harmonica’s value.

[CAPABILITY/ FEASIBILITY] Please provide a detailed breakdown of your project’s milestones and each of the main tasks or activities to reach the milestone plus the expected timeline for the delivery.

1. MVP development:
2. Integration with Discord and OpenAI (weeks 0-4)
3. Better conversational UX and dev ops
4. Pilot study:
5. Experiments with prompt engineering / facilitation logic
6. Bot installation flow and settings (web UI)
7. Generating proposals / prompts for specific DAO needs

[CAPABILITY/ FEASIBILITY] Please describe the deliverables, outputs and intended outcomes of each milestone.

Milestone 1: MVP Development (Alpha baseline) — 31 July 2023

The goal of the first milestone will be to get an MVP prototype of the bot. It shall be able to run a meeting session with several participants, and shall have sufficient functionality to support a basic series of experiments in the pilot study.

Milestone 2: Pilot study — 31 September 2023

The goal of the second milestone is to conduct a pilot study to explore different user interaction patterns with Harmonica. We will try to learn from this study to adapt and improve the alpha baseline into a beta version. If the results of the study are favourable, we will have a beta baseline that we can bring to a wider audience. If they are not favourable, we will have specific recommendations for how to move forwards.

We will use a continuous testing approach to develop the beta.

• Two populations will be used for testing
• Paid product testing services (such as Amazon Turk)
• Communities interested in the alpha
• Data will be gathered via querying users and recording sessions.
• Both populations will receive basic questions
• ”did you like this?”
• “was the problem solved?”
• We will reach out to at least 20 Cardano communities for feedback
• In the alpha stage, all sessions will be recorded (provided the consent of the communities) so that a session can be analysed more thoroughly if it merits it

[RESOURCES & VALUE FOR MONEY] Please provide a detailed budget breakdown of the proposed work and resources.

→ Product development:

• Discord bot / frontend (180hrs (@170 ada/hr) = 30,600)
• Prompt engineering (60hrs (@170 ada/hr) = 10,200)
• UX copywriting (30hrs (@170 ada/hr) = 5,100)
• Web UI (bot installation + settings) (60hrs (@170 ada/hr) = 10,200)

→ Product Distribution, Sales and Marketing (10,300)

• Organize an event: “Decision-Making Prompt Jam”
• Create the necessary communication materials
• Engage with Cardano communities

→ Tools, Infrastructure & Operations (10,300)

• Cost of running: OpenAI license to use of GPT-4
• Amazon Turk or some other paid testing

Total: 76,700 ada

[RESOURCES & VALUE FOR MONEY] How does the cost of the project represent value for money for the Cardano ecosystem?

The project builds upon previous R&D and non-remunerated labour invested by the contributors, adding over $100k of value to the project delivered to the Cardano community.

[IMPORTANT NOTE] The Applicant agreed to Fund10 rules and also that data in the Submission Form and other data provided by the project team during the course of the project will be publicly available.

I Accept