Your First Freelance Hire in 2026: Subcontractor, VA, or AI Agent?
A US VA from Belay runs `$1,600 to $4,400/mo`. A Filipino VA via OnlineJobs.ph runs `$320 to $800/mo`. A senior AI specialist on Upwork runs `$100 to $300/hr`. The right first hire in 2026 depends on which constraint is binding.
The Delivvo team· May 30, 2026 8 min read
A US-based virtual assistant from Belay runs $1,600 to $4,400/month depending on hours, working out to roughly $20 to $55/hour (Belay, 2025). A Filipino VA hired through OnlineJobs.ph costs $320 to $800/month for 20 hours/week, plus a $69/month employer platform fee (OnlineJobs.ph, 2026). Upwork's average freelancer rate is $39/hour in 2026, with US-based freelancers averaging $47.71/hour, most rates falling between $24 and $62/hour (Upwork, 2026). At the top of the market, AI and ML freelance specialists command for senior work.
Keep reading
$100 to $300/hour
Then there is the fourth option that did not exist in this conversation three years ago: an AI agent. ChatGPT Pro runs $20/month for individual use, ChatGPT Business $25/seat/month, Claude Pro $20/month, Notion AI bundled. None of these replace a human hire across the board, but for specific recurring tasks they have made the "should I hire someone" math much messier.
This 2026 guide walks through how to make the decision when you are growing past a one-person operation, with real cost data, the IRS rules that still matter, and a decision framework that picks the right hire for the constraint you are actually trying to solve.
The three constraints driving the hire
Every "should I hire" conversation we have with freelancers in 2026 reduces to one of three constraints. The constraint determines the right answer.
The first is skill. The freelancer is turning down work because they lack expertise in a specific area. The right hire is a subcontractor with that skill, not a generalist VA and not an AI agent.
The second is time. The freelancer can do all the work but is spending too many hours on low-value tasks. The right hire is a VA (or an AI agent for the most-repetitive tasks).
The third is cash flow. The freelancer needs to expand capacity without taking on fixed costs. The right hire is either a project-based subcontractor (variable cost matches variable revenue) or an AI agent (low fixed cost).
Picking the wrong hire for the constraint is the most common reason first hires get unwound within six months.
Option 1: Subcontractor
A 1099 contractor (or international equivalent) you bring in on a specific engagement or as a recurring partner. The relationship is project-based or hourly, the contractor uses their own tools, sets their own hours, and works for other clients.
When it works. When the constraint is skill, the deliverable is bounded, and the work is high enough value to absorb the contractor's rate. The math is simple: if the project pays you $10,000 and the subcontractor takes $3,500, your gross margin is healthy and the engagement scales.
When it does not work. When the work is recurring low-value tasks. Paying a senior contractor $50/hour to manage your inbox burns money fast.
Operational reality. The IRS still uses the same multi-factor test for 1099 classification: behavioural control, financial control, and the relationship of the parties. The threshold for 1099-NEC reporting is $600 in a calendar year (IRS, 2025). If you treat a subcontractor like an employee (set their schedule, provide their tools, require exclusivity), the IRS reclassification risk is yours, not theirs.
Common 2026 pricing. Specialist subcontractors land in the $60 to $200/hour band for skilled work. The Upwork average of $39/hour is closer to the generalist tier. For senior AI/ML specialists, the band is $100 to $300/hour (Upwork, 2026).
Option 2: Virtual Assistant
A part-time or full-time assistant, usually working remotely, handling administrative and operational tasks. The two main cost tiers in 2026 are US-based and offshore.
US-based VA. Belay charges $1,600 to $4,400/month depending on hours, which works out to roughly $20 to $55/hour (Belay, 2025). Other US providers (Time Etc, Boldly, Worxbee) sit in similar bands. The premium buys time-zone alignment, native English, and lower coordination overhead.
Offshore VA, Philippines. OnlineJobs.ph runs $320 to $800/month for 20 hours/week, plus a $69/month employer fee (OnlineJobs.ph, 2026). Full-time Filipino VAs sit in the $600 to $1,500/month range for skilled roles. The trade-off is time-zone overlap, communication overhead in the first three months, and the operator's responsibility for managing the relationship more actively.
When it works. When the constraint is time, the work is recurring, the tasks are well-documented, and the freelancer is willing to spend the first two months in real management mode. The VA earns back their cost when the freelancer's hours move from low-value work to high-value client work or new business development.
When it does not work. When the freelancer cannot articulate the tasks they want done, or when the volume is too low to fill a part-time role. A part-time VA at 20 hours/week who only has eight hours of real work to do is going to drift.
Option 3: AI agent
A tool stack of AI products operated by the freelancer, doing what would otherwise be VA work. In 2026 this realistically means ChatGPT Pro or Business, Claude Pro, Notion AI, plus a workflow tool (n8n, Zapier, or Make) to chain calls.
Total monthly cost. A solo freelancer using ChatGPT Pro ($20), Claude Pro ($20), and a Notion or workflow subscription ($20 to $50) lands at $60 to $100/month. Compared to the cheapest VA option (Filipino VA at $320 to $800/month), the AI agent is 5x to 10x cheaper.
When it works. When the constraint is cash flow, the tasks are well-bounded, and the freelancer is comfortable specifying tasks in writing. Inbox triage, content drafts, meeting note summaries, simple research, and most "first pass" tasks in any creative discipline are well-served by AI agents in 2026.
When it does not work. When the work requires real judgement, sustained context across sessions, or a human in the loop for any sensitive interaction. The AI agent cannot call a client, cannot handle a payment dispute, and cannot run an account on behalf of the freelancer in any way that survives the first edge case.
The realistic shape. Most freelancers we see in 2026 who scaled past $200K/year solo did so with an AI agent stack handling roughly 30 to 60% of what would have been VA work, while still hiring a part-time VA for the remainder. AI as a complement, not a replacement, is the dominant pattern.
The 2026 decision framework
A short checklist that maps constraint to hire:
| Constraint | Volume | Recommended first hire | |---|---|---| | Skill gap | High value per task | Specialist subcontractor | | Skill gap | Low value per task | AI agent + freelancer review | | Time | Recurring, well-documented | Offshore VA | | Time | Recurring, sensitive | US VA | | Cash flow | Variable workload | Project-based subcontractor | | Cash flow | Steady workload | AI agent stack |
The freelancers who skip this framework and hire the first option that feels modern (AI agent) end up under-served on judgement-heavy work. The ones who hire the first option that feels safe (US VA) end up over-paying for capacity they do not use.
What changes about the math in 2026
Upwork shifted its platform fee structure in May 2025 from the previous tiered model to a per-contract variable fee of 0 to 15% (typically around 10%) (Upwork, 2025). For freelancers using Upwork to find subcontractors, this slightly lowered the all-in cost of a contractor relationship and made smaller, exploratory engagements economic again.
The IRS 1099-NEC threshold remains $600 per contractor per year (IRS, 2025). The 1099-K threshold for third-party payment processors was restored to $20,000 under the OBBBA, which simplified reporting for freelancers paying contractors through Stripe or PayPal Business at smaller scales (see also our guide on the restored 1099-K threshold).
The BLS Contingent Worker Supplement is the canonical primary source on the contractor share of the US workforce (BLS). For non-US freelancers, the equivalent classification rules vary by jurisdiction and have tightened materially in the UK (IR35) and the EU (Platform Work Directive) over the past 18 months.
The hidden cost everyone underestimates
Two costs that do not show up in the price comparison and almost always blow up first-time hires.
The first is the freelancer's own management time. A part-time VA needs roughly 5 to 10 hours/week of management for the first two months, dropping to 2 to 4 hours/week once the relationship stabilises. Subcontractors need less but in spikier patterns around handoff points. AI agents need almost zero ongoing management but require the freelancer to write good prompts and review every output.
The second is the cost of poor delegation. A VA given vague instructions produces vague work that the freelancer then has to redo. A subcontractor given a poorly-scoped project produces a deliverable that fails review. An AI agent given a sloppy prompt produces output that needs a full rewrite. In all three cases, the cost is the freelancer's time on rework, which usually exceeds the savings from the hire itself.
The fix is the same regardless of hire type. Write the task down. Show an example of the output. Define what "good" looks like. The freelancers who do this consistently get value from their first hire in the first month. The ones who skip it spend three months thinking they made the wrong choice.
Making the hire operational
A simple onboarding playbook that works for all three types of hire.
Signed agreement. A subcontractor agreement for 1099 work, a VA service agreement for managed providers, or a documented prompt and acceptance criteria for AI workflows.
Documented tasks. Three to five concrete tasks with expected output, due date, and review steps. Vague briefs are the single biggest first-hire failure mode.
A weekly check-in cadence. Even with AI agents, a structured weekly review of what worked and what needs to change catches problems before they compound.
An audit trail. Every project, every invoice, every deliverable in one place. If you run client work through a portal tool like Delivvo, use the same portal to track the work going out to subcontractors so the chain is visible end to end.
FAQ
Should my first hire be a VA or a subcontractor?
If the constraint is your time, a VA. If the constraint is a skill gap, a subcontractor. The mistake to avoid is hiring a VA when the actual problem is you cannot deliver client work alone, or hiring a subcontractor when the actual problem is you are spending too many hours on email.
Can an AI agent really replace a part-time VA?
For a specific slice of work, yes. For the full surface of a VA role, no. The realistic 2026 pattern is an AI agent handling 30 to 60% of what would otherwise be VA work, with the remainder still going to a human.
What does the IRS require to classify a worker as a contractor in 2026?
The same multi-factor common law test that has applied for years: behavioural control, financial control, and the relationship of the parties. The DOL has its own classification framework for wage-and-hour law that overlaps but is not identical. Both have tightened over the past two years for gig and platform-based work.
How much should I budget for a first hire in year one?
Realistic 2026 budgets: $5,000 to $15,000 total for the first year of an offshore VA, $25,000 to $55,000 for a US-based VA, $15,000 to $50,000 for a part-time subcontractor on retainer, and $1,000 to $3,000 for an AI agent stack with workflow tools. Plan for productivity to drop in months one and two regardless of which option you choose.
What is the most common first-hire mistake?
Hiring before the work is documented. The freelancer thinks they will figure it out as they go. They burn through three weeks of paid time training the hire, conclude the hire is not working, and dismiss them. The hire was almost never the problem.
The 2026 takeaway
The right first hire in 2026 depends on which constraint is binding. Skill gaps go to specialist subcontractors. Recurring time problems go to VAs (US-based for high-trust or time-zone-sensitive work, offshore for cost-sensitive volume). Cash-flow-constrained capacity expansion goes to AI agents, with a part-time VA layered on once the work is documented. The freelancers who hire well in 2026 spend more time on the brief than on the hire decision itself, because the brief is what makes any of the three options actually work.
