2.9%
of employed people in Pakistan do digital platform (gig) work as their main job.
Why it matters here Platform-arranged work is already a measured part of Pakistan’s labour market, and much of it is local and physical rather than online.
Thesis
Why a task can be the unit of work.
Task4Task is an AI-powered marketplace for getting things done — paid or by skill exchange, remote or in person. This page sets out what we believe about how work can be organized: what is being built, why it may matter, what remains a hypothesis, and what still needs to be proven. For the investor briefing, see Investors.
No market-size claims, revenue figures, fee schedules or growth projections are published here. Where something is a hypothesis, it is labelled as one.
01
The thesis
A task can be the smallest unit of work.
Much of what people need done is bounded: a sofa moved, a tap fixed, a logo delivered, a lesson taught. These are tasks — specific outcomes with a beginning and an end — not ongoing jobs.
If a task is the unit, a marketplace can organise demand and supply around that unit: form it clearly, match people to it, agree terms, guide execution, and leave a record. That is the core idea Task4Task is built on.
This is a product thesis, not a claim about market size. Whether tasks become a durable unit of marketplace activity is something the product still has to demonstrate.
02
The problem
Needs that do not fit neat employment or conventional service structures.
People have needs that do not always map to hiring an employee or buying a packaged service. Some need a one-time task. Others need recurring help, urgent help, a local physical task, remote work, a specific skill, or a skill exchange rather than cash.
Today those needs are often arranged through word of mouth, messaging groups, or informal asking around. That works when you already know someone. It fails when you do not, or when the need is urgent, unfamiliar, or hard to describe.
Task4Task starts from that gap: make the need legible as a task, then give both sides a structured path from request to completion.
03
The product
Need → AI → Task → People → Agreement → Contract → Execution → Completion
See how the marketplace worksThe system is a pipeline, not a directory of profiles. A person states a need. AI helps form it into a structured task. People discover it, propose, and agree. Terms become a contract. Execution is guided — including physical travel and remote submission where relevant — then completed and reviewed.
Connecting two profiles is only the middle of that chain. The product’s claim is that structure around the task — formation, agreement, guided work, evidence and reputation — is what makes strangers able to finish work together.
04
AI as infrastructure
AI assists. People decide.
See AI help form a taskAI is infrastructure for making work understandable, not a decision-maker. It can help with task formation from natural language, missing information, content review before publishing, and skill-based recommendations and matching.
People remain responsible for what is published, who is hired, which proposal is accepted, and how the contract is carried out. That boundary is a product rule, not a slogan.
05
Marketplace model
Demand and supply, across paid, barter, remote, physical and urgent work.
Walk the product journeysDemand is people who need something done. Supply is people who can do something. Both sides use the same marketplace through two workspaces: Hiring and Getting Work.
The marketplace is designed to support paid work and skill exchange; remote work and physical work; and urgent instant requests to nearby workers who are Available Now. Those modes share proposals, contracts and completion — so the unit stays the task, not a siloed product line.
06
Trust
Trust compounds from identity, work and evidence — not from a single score.
Read the trust & safety layerThe trust architecture includes identity verification, profiles, skills, certificates where relevant, reviews after a job, separate hiring reputation and working reputation, disputes with supporting evidence, and execution history inside the contract.
The hypothesis is that trust compounds when signals are earned through completed work and preserved separately for how someone hires and how they work. Whether those signals change behaviour enough to sustain a marketplace is still something to prove.
07
Network effects
Potential reinforcing loops — treated as hypotheses.
Network effects are possible in a two-sided task marketplace. They are not treated here as established results.
- ResearchPotential network effect
More useful tasks → more opportunities for workers. Hypothesis: richer demand draws and retains supply.
- ResearchPotential network effect
More capable workers → more useful choices for hirers. Hypothesis: supply quality increases demand conversion.
- ResearchPotential network effect
More completed work → more reputation and trust signals. Hypothesis: history lowers friction for the next hire.
- ResearchPotential network effect
More structured task data → potentially better recommendations and matching. Hypothesis: formation quality improves discovery.
- ResearchPotential network effect
More activity → potentially better marketplace liquidity. Hypothesis: density makes matching faster and fairer.
08
Business model
Business model under development.
No final business model has been established for public description. Transaction fees, subscription prices, commissions, revenue numbers and financial projections are not published here because they have not been locked as product facts.
Monetisation directions remain under consideration as the product and marketplace behaviour are researched. Investors should treat economics as an open workstream, not an implied outcome.
- ResearchWhat we need to prove
Which value participants will pay for — if any — and how that aligns with liquidity, trust and completion.
09
Defensibility
Potential sources of advantage — not proven moats.
Features are not moats. The following are potential sources of defensibility if they compound in practice. None is claimed as a proven moat merely because it exists in the product specification.
- ResearchPotential moat
Structured task data — if formation quality and history become hard to replicate from listings alone.
- ResearchPotential moat
Reputation graph — hiring and working reputations earned across completed contracts.
- ResearchPotential moat
Task-to-skill relationships, execution history, trust infrastructure, matching intelligence, marketplace liquidity and workflow depth — each only if usage makes them cumulative.
- ResearchProven moat
None claimed. Defensibility remains to be demonstrated through durable usage, trust and liquidity.
10
What remains unproven
What we still need to prove.
Credibility for investors depends as much on what is not yet known as on what is specified. The following are open until real marketplace evidence exists. No numbers are assigned because production metrics have not been published.
- ResearchMarketplace liquidity
Whether demand and supply meet often enough, in enough categories and places.
- ResearchRepeat usage
Whether hirers return with the next task and workers with the next proposal.
- ResearchTask completion rates
Whether structured contracts raise completion versus informal arrangements.
- ResearchSupply–demand balance
Whether the marketplace stays useful as it grows, without chronic shortages or oversupply.
- ResearchMatching quality
Whether skill-based recommendations put the right work in front of the right people.
- ResearchTrust formation
Whether verification, dual reputation and evidence change willingness to hire strangers.
- ResearchEconomics
Whether a sustainable model can fund the marketplace without harming liquidity.
- ResearchRetention
Whether both sides stay after early experiments.
- ResearchDispute rates
Whether guided execution and change requests keep serious disputes rare and fair.
- ResearchWillingness to use barter
Whether skill exchange is used beyond edge cases.
- ResearchAdoption of AI-assisted task creation
Whether people prefer natural-language formation over blank forms — and whether it improves outcomes.
What we are currently researching
Open questions, not findings.
These match the public research programme. Findings appear only when there are findings.
- 01How much of a task can be understood from one request (RQ-01).
- 02How two people agree an exchange is fair without money (RQ-02).
- 03What lets strangers trust each other enough to meet (RQ-03).
- 04How much location should be shared, with whom, and for how long (RQ-04).
- 05When a task is urgent, who should hear about it first (RQ-05).
- 06How hiring and working reputation should be shown together (RQ-06).
- 07How disagreements about work should be settled fairly (RQ-07).
- 08Whether skill-based recommendations put the right tasks in front of the right people (RQ-08).
- 09How both sides always see the same state of a task (RQ-09).
What is different
The task, not the profile, is the unit of work.
A difference in product model, not a claim about anyone else’s product. Whether it produces better outcomes is one of the things still to prove.
A familiar mental model
- Person
- Search
- Opportunity
- Chat
- Work
A familiar way to get something done starts with people: search for someone, find an opening, then work out the details in conversation. What was agreed, and whether it was done, lives in a chat thread and in two people’s memories.
The Task4Task model
- Needadds Someone’s own words
- Taskreads The wordsadds WhatSkillsWhereWhenValuePeopleExtras
- Peoplereads Skills · Where · Whenadds ProposalsConversations
- Agreementreads Value · The chosen proposaladds Contract terms
- Executionreads Where · Contractadds Travel and arrivalMessagesSubmitted work
- Completionreads Contract · Submitted workadds ApprovalReviewsReputation
In Task4Task the task itself is the thing everyone works on. It starts as a need and keeps its context, from what and where to what was agreed, through every step. Each step reads what came before and adds to it, so nothing has to be re-explained, and both sides look at the same record.
Market context
The world the thesis is about.
External research only, with its source. None of these figures is a Task4Task metric, and no market size is derived from them.
154–435M
people do online gig work worldwide: 4.4% to 12% of the global labour force.
Why it matters here Work arranged task by task, through platforms, is a large and growing way people earn, especially in developing countries.
5×
growth in the number of digital labour platforms worldwide between 2010 and 2020, to at least 777 active platforms.
Why it matters here Work is spread across many separate platforms, each organised around one kind of work.
$90M → $501M
reported losses to job and employment-agency scams in the United States, 2020 to 2024. Reports tripled.
Why it matters here When work is arranged between people who don’t know each other, trust is a real cost, not a detail.
These figures describe the market, not Task4Task. Task4Task’s own numbers are published on the Network, and there are none yet.
Proven · being tested · unproven
Where the evidence stands.
Real Proven
- Product architecture established
- AI task formation introduced as a core capability
- Trust architecture specified
- Research programme published
- Interactive product demonstration website
These are product and company milestones. No marketplace outcome is proven yet.
Research Being tested
- RQ-01 How much of a task can be understood from one request?
- RQ-02 How do two people agree that an exchange is fair, without money?
- RQ-03 What lets two strangers trust each other enough to meet?
- RQ-05 When a task is urgent, who should hear about it first?
- RQ-06 Should someone’s reputation as a hirer count separately from their reputation as a worker?
- RQ-08 Can recommendations based on skills put the right tasks in front of the right people?
- RQ-09 How do both sides always see the same state of a task?
Open questions under design and research. No findings have been published.
Research Unproven
- Marketplace liquidity
- Repeat usage
- Task completion rates
- Supply–demand balance
- Matching quality
- Trust formation
- Economics
- Retention
- Dispute rates
- Willingness to use barter
- Adoption of AI-assisted task creation
The larger opportunity
Every skill, made usable. Paid or traded, online or in the street.
If a sentence is enough to start a task, many more tasks get asked for. If a skill is enough to pay for one, many more people can hire. If the map is part of the contract, the marketplace reaches work that never touches a screen.
Task4Task is being built for pakistan first: a task, a proposal, a contract, a review. Whether that opportunity materialises depends on the proof points listed above — liquidity, completion, trust and economics among them.
Continue in Research, the Network, or walk the product experience.