← All skills

Skill file (for agents): /personas/SKILL.md

Personas

User Personas for The Synthesis

These personas represent real categories of people who face real problems that agentic infrastructure can address. They are starting points, not endpoints. Builders should use them to spark empathy, then make them more specific based on their own knowledge and experience.

Each persona includes an empathy map, a day-in-the-life scenario, a JTBD statement, and design prompts tied to the hackathon themes.

Table of Contents

  1. Priya -- The Freelancer Who Automates Her Business
  2. Marcus -- The Researcher Accessing Sensitive Data
  3. Daniela -- The Small Business Owner in an Unstable Jurisdiction
  4. Kai -- The Developer Building Multi-Agent Workflows
  5. Amara -- The Person Who Just Wants AI to Handle Their Life
  6. Tomasz -- The Activist or Journalist in a Hostile Environment
  7. Lena -- The DAO Treasurer Managing Community Funds
  8. Ravi -- The Gig Worker in an Emerging Economy
  9. Sofia -- The Parent Managing Family Digital Infrastructure
  10. Jin -- The Creator Licensing Work Through Agents

1. Priya -- The Freelancer Who Automates Her Business

Background: UX designer in Mumbai. Clients in the US, Europe, and Southeast Asia. Uses an AI agent to send invoices, follow up on late payments, pay for SaaS tools and subscriptions, and manage project timelines. Bills in USD, EUR, and INR depending on the client.

Empathy Map:

Day in the Life: 7:30 AM -- Priya checks her phone. Agent sent two invoices overnight and paid for a Figma renewal. She opens the payment processor to verify the Figma charge went through. It shows "processing." She doesn't know if that means it settled or is pending. She makes a mental note to check again later.

9:00 AM -- A US client paid an invoice. The agent confirms receipt, but Priya's bank doesn't show the deposit yet. She logs into the payment platform to check the status. It says "completed" on their end. She won't know for sure until the bank clears it, which could take 1-3 days.

2:00 PM -- She discovers a charge from a tool she doesn't recognize. Her agent auto-renewed a trial subscription she forgot she started. There's no easy way to see what her agent has committed to or set limits on what it can spend.

JTBD: "When my agent handles payments on my behalf, I want to independently verify what it paid, to whom, and when it actually settled, so I don't have to reconstruct the truth from three different platforms."

Design Prompts:

2. Marcus -- The Researcher Accessing Sensitive Data

Background: Political science PhD candidate at a European university. Researching authoritarian governance. Uses an AI agent to query academic databases, purchase access to paywalled reports, aggregate findings, and translate documents. His research touches on surveillance infrastructure, detention records, and financial flows of authoritarian regimes.

Empathy Map:

Day in the Life: 8:00 AM -- Marcus gives his agent a list of search terms related to detention facility financing in a specific country. The agent queries three databases and returns 47 results.

10:00 AM -- His agent purchases access to a paywalled report from a think tank. The payment goes through his university account. The think tank now has a record that his university accessed this specific report.

1:00 PM -- He notices the agent accessed a government database he didn't authorize. The agent interpreted a broad search instruction more aggressively than he intended. The query is now in the database's access logs, tied to his university's IP.

4:00 PM -- Marcus prepares a bibliography. He realizes that the combined set of queries, purchases, and downloads creates a detailed map of his research focus that anyone with access to the relevant logs could reconstruct.

JTBD: "When my agent accesses research materials on my behalf, I want it to do so without creating a composite picture of what I'm studying or who I am, so I can research sensitive topics without building a surveillance profile by accident."

Design Prompts:

3. Daniela -- The Small Business Owner in an Unstable Jurisdiction

Background: Runs a small import/export business in Buenos Aires. Deals with suppliers in Brazil, Mexico, and China. Uses an AI agent to handle cross-border payments, negotiate prices, manage contracts, and track shipments. Operates in an environment where currency controls change unpredictably, payment providers get shut down, and platform terms shift without notice.

Empathy Map:

Day in the Life: 6:00 AM -- Daniela checks overnight transactions. Her agent tried to send a payment to a Brazilian supplier but the payment provider flagged it for "enhanced review." She has no timeline for resolution and no appeal process.

9:00 AM -- She tells her agent to reroute the payment through a different provider. The agent negotiates a slightly different exchange rate. Daniela accepts the loss to keep the shipment on schedule.

12:00 PM -- Her agent closes a deal with a new supplier in Mexico. The contract terms are recorded on the platform the agent used to negotiate. Daniela worries about what happens if that platform changes its terms of service or exits the Argentine market.

5:00 PM -- She hears rumors of new currency controls being announced next week. She begins preparing contingency plans for her agent's payment routes.

JTBD: "When my agent makes deals and moves money on my behalf across borders, I want the infrastructure it relies on to keep working regardless of what any single government, bank, or platform decides to do, so my business can survive policy changes I can't predict."

Design Prompts:

4. Kai -- The Developer Building Multi-Agent Workflows

Background: Senior software engineer in Berlin. Building a product where multiple AI agents coordinate to complete complex tasks: one agent handles research, another writes code, a third handles deployment, a fourth manages billing. The agents call each other's APIs, negotiate costs, and commit to deliverables.

Empathy Map:

Day in the Life: 9:00 AM -- Kai deploys a new version of his orchestration layer. The research agent and the code-writing agent need to negotiate a price for each task. He's built a simple bidding mechanism, but it runs on his server, making him the single point of failure.

11:00 AM -- The deployment agent claims it successfully deployed a build, but the monitoring agent reports errors. Kai has to manually arbitrate. There's no shared source of truth about what actually happened.

2:00 PM -- A new agent service launches that claims better performance for code review. Kai wants to swap it in, but the new service has no verifiable track record. Its documentation says it's great. That's all he has to go on.

4:00 PM -- Kai discovers that the billing agent overcharged for three tasks yesterday. The receipts the billing agent provided don't match the actual API calls. He has no recourse because the billing happened off-chain through a centralized API.

JTBD: "When I build systems where agents interact with each other, I want a standard way to verify what each agent claims it did, enforce the commitments they make, and resolve disputes without building custom trust infrastructure for every single integration."

Design Prompts:

5. Amara -- The Person Who Just Wants AI to Handle Their Life

Background: Marketing manager in Lagos. Not technical. Uses an AI assistant for everything: managing subscriptions, scheduling appointments, booking travel, handling personal finances, communicating with service providers. She doesn't think about the infrastructure underneath. She just wants things handled.

Empathy Map:

Day in the Life: 7:00 AM -- Amara's agent paid for three subscriptions overnight, scheduled a dentist appointment, and booked a car for tomorrow. She gets a summary notification and doesn't open it.

12:00 PM -- A friend mentions that a popular subscription service was hacked and user data was leaked, including transaction histories and linked services. Amara checks to see if she uses that service. Her agent handled the signup six months ago. She isn't sure.

3:00 PM -- She gets a targeted ad that feels too specific. She wonders if one of the services her agent interacts with sold her data. She has no way to know, and no way to find out.

7:00 PM -- She reads an article about AI agents creating comprehensive profiles of users across every service they touch. She feels a brief spike of concern, then closes the article. She needs her agent to function. The alternative is doing everything herself.

JTBD: "When I delegate my daily tasks to an AI agent, I want it to handle everything without creating a surveillance profile of my life that I didn't consent to and can't control, so I can have convenience without giving up my privacy."

Design Prompts:

6. Tomasz -- The Activist or Journalist in a Hostile Environment

Background: Investigative journalist in Warsaw. Covers government corruption. Uses an AI agent to communicate with sources, process leaked documents, pay informants, and coordinate with editors. Every transaction, query, and communication his agent makes is a liability if traced back to him.

Empathy Map:

Day in the Life: 6:00 AM -- Tomasz receives an encrypted message from a source who wants payment for documents. The source needs the money today. Tomasz can't use traditional payment rails because they'd create a connection between him and the source.

10:00 AM -- His agent translates and summarizes leaked documents. The agent uses a translation API. Tomasz worries about what the API provider logs. The content of the documents, combined with the source language and the timing, could identify the leak.

2:00 PM -- He needs to coordinate with a colleague in another country. His agent sends the message through a standard channel. The metadata (who communicated, when, from where) is as revealing as the content.

5:00 PM -- Tomasz reviews his agent's activity for the day. He counts seven third-party services his agent contacted. Each one has a log. Combined, those logs tell the story of exactly what he was working on today.

JTBD: "When my agent acts on my behalf in sensitive situations, I want every interaction it has to be private by default and untraceable to me or my sources, so the people who trust me with dangerous information stay safe."

Design Prompts:

7. Lena -- The DAO Treasurer Managing Community Funds

Background: Part-time DAO treasurer for a 200-person creative cooperative based across Europe. The DAO funds community projects, pays contributors, and manages a shared treasury. Lena uses AI agents to handle routine disbursements, track budgets, and enforce spending policies the DAO voted on.

Empathy Map:

JTBD: "When the DAO's agent disburses community funds, I want the spending rules the community voted on to be enforced by the system itself, not by me personally, so the community's money is governed transparently and I'm not a bottleneck or a target."

Design Prompts:

8. Ravi -- The Gig Worker in an Emerging Economy

Background: Delivery driver and part-time freelance data labeler in Hyderabad. Uses AI agents through multiple gig platforms. Each platform's agent manages his tasks, payments, and ratings. He has no control over how any of them work.

Empathy Map:

JTBD: "When I work through platforms that use agents to manage my tasks and payments, I want my reputation, work history, and earnings to belong to me and move with me, so no single platform can erase my livelihood or trap me."

Design Prompts:

9. Sofia -- The Parent Managing Family Digital Infrastructure

Background: Mother of two teenagers in São Paulo. Uses AI agents to manage family subscriptions, monitor spending on kids' accounts, coordinate school logistics, and handle household administration. Wants oversight without micromanagement.

Empathy Map:

JTBD: "When AI agents act on my family's behalf, I want to set boundaries that are enforced at the infrastructure level, not by each individual app, so my kids are protected consistently without me auditing every interaction."

Design Prompts:

10. Jin -- The Creator Licensing Work Through Agents

Background: Independent musician and visual artist in Seoul. Uses AI agents to license his work, negotiate usage terms, collect royalties, and manage distribution across platforms. Currently, each platform has its own agent that controls his content and takes a cut.

Empathy Map:

JTBD: "When my agent licenses my creative work, I want it to represent my interests on neutral infrastructure where I can verify every usage, enforce my terms, and collect what I'm owed without trusting each platform to be honest."

Design Prompts: