Give Your Virtual Assistant Work They Can Own
By Aldridge Dagos, operations software engineer
What should you delegate to a virtual assistant first? Start with one recurring administrative responsibility you can demonstrate in real work. It should reach a visible finish without depending on your personal judgment.
Inbox triage or calendar coordination often fits. Pick one and prove the handoff before adding the next. A one-off request moves a task, while a recurring responsibility moves the need to remember, route, follow up, and close it. That difference matters more than the size of the task list.
A founder forwards an email to a VA and writes, “Please handle this.” The VA drafts a reply, then asks whether the customer should receive a refund. The founder answers. The VA asks where to record it. Later, the founder notices nobody scheduled the follow-up. The founder delegated every step, yet remained the memory and the route.
Demand for virtual assistance is not the hard part. In Upwork’s report published January 8, 2026, new virtual-assistance contracts from small and medium businesses rose 18 percent month over month in December 2025. The report covers Upwork’s United States marketplace, so it is a hiring signal, not proof that every handoff works. The harder question is whether the new person receives work they can actually own.
Illustrative handoff 01
A task returns for instructions. A responsibility returns only for judgment.
Task-by-task handoff
Repeated owner re-entry- 01 Owner assigns
- 02 VA starts
- 03 Owner clarifies
- 04 VA resumes
- 05 Owner checks
- 06 Work closes
Owned responsibility
One judgment point- 01 Brief once
- 02 VA routes work
- 03 Exception arrives
- 04 Owner decides
- 05 VA closes
Start with work that returns every week
The best first responsibility is not simply work you dislike. It should return often enough to teach and stay stable enough that you can describe a recognizable finish.
Inbox triage is useful because the work exposes its own boundaries. Known questions can receive approved answers, while leads move into the CRM and invoices reach accounting. Scheduling requests follow the calendar rules. Personal notes and unusual complaints move into a protected review path, along with any message that could change a contract.
Dan Martell’s Buy Back Your Time goes further and argues that an administrative assistant should own both inbox and calendar. That is a strong destination. For a first handoff, I prefer a narrower start: move one recurring responsibility and watch where it returns to the owner. Fix those gaps before adding another live surface.
This is not hesitation about delegation. It is a way to see whether the operating agreement works.
If the VA owns inbox triage, define a result both of you can inspect: the shared inbox receives attention twice each workday, and known requests get an approved answer. Every message that creates work reaches the right task or customer record. Follow-ups carry a due date, while sensitive items stay protected. At the end of the cycle, nothing remains merely because the owner forgot to look. The role now has a visible finish.
Try the same test on calendar work, where “Book this meeting” names only one task. Calendar coordination continues through availability rules, protected focus time, required attendees, confirmation, rescheduling, and a clean record. The scheduling and booking system can enforce some of those rules, while the VA handles the human exceptions around them.
A responsibility should also be worth learning. When work returns each week, the VA sees patterns and gains useful judgment. They notice stalled requests and missing customer details. Over time, they also learn which follow-up wording gets a clear answer. Random errands do not create that learning curve.
Start small enough to observe, but not so small that ownership disappears. One complete loop beats ten disconnected favors.
Hand over the finish, not one task
A clean handoff transfers context before it transfers volume. Show the VA three real passes through the inbox, not one perfect demonstration. Martell calls this the Camcorder Method. In the book, he recommends recording the real work while explaining why each choice happens and making three recordings because recurring work changes from one run to another.
The recordings should not become a pile of video nobody can search. Martell also has the person doing the work turn those recordings into a working playbook. That step tests their understanding and exposes ambiguous choices. It also catches the knowledge the owner relies on but never said aloud. In my Ten80Ten practice, we review that document together and the VA maintains it as the inbox changes.
The finish belongs in plain language. An email is not complete because somebody replied. It may need a CRM update, a promised document, a scheduled follow-up, or a task assigned to another owner. If the outcome creates a second obligation, the record must carry it.
This is where software can support ownership without pretending to replace it. I built an inbox triage pipeline that turns a full day of email into an organized briefing before work starts. The useful part is not sorting alone. It keeps each source link beside its category and next action, so the person reading the result can act without reopening the entire inbox.
When I inspect a handoff, I count how many times routine work had to climb back to the owner. One return may expose a missing rule. Repeated returns for the same choice mean the responsibility never fully moved.
Do not solve that problem by telling the VA to stop asking questions. Improve the boundary. Add the missing example or clarify the safe choice. If one fact changes the route, put that fork into the instruction and let the next instance test it.
The owner has work here too. Time recovered from email will disappear if it remains an empty patch in the day. Put it against sales, strategy, client judgment, product decisions, or another responsibility that only the owner can carry. Delegation buys very little when reactive work immediately fills the space.
Keep the exception path small
Ownership does not mean the VA decides everything. It means ordinary work moves without permission and unusual work reaches the right person in a useful form.
Take the refund request from the opening scene. The VA should not send “What should I do?” if the available record can support more. They can state the narrow problem, show the realistic options, and recommend one based on policy and customer history. Martell presents this as the 1:3:1 rule: one problem, three viable paths, and one recommendation.
The exact number of options matters less than the discipline. The VA investigates the exception before it reaches the owner, who can spend attention on the decision instead of rediscovering the case.
An exception queue helps when these cases recur. It gives the uncertain item a state, owner, evidence, and return path. More important, resolved exceptions can become new instructions. The VA should not have to escalate the same ordinary condition forever.
Some boundaries should stay firm. Final legal, clinical, tax, financial, and people decisions belong with the qualified owner. Keep irreversible external actions there too. Sensitive personal matters need deliberate review, and any promise that changes a contract follows the same boundary. A VA can gather the facts and prepare a draft without receiving authority the business should not delegate. When the evidence supports a direction, they can also recommend the next step.
Other decisions should move outward as trust and evidence grow. A scheduling conflict that once required the owner can become a calendar rule. A common customer request can receive an approved response. Each resolved case makes the responsibility more complete, as long as the playbook records the answer where the VA works.
This protects both sides. The owner receives fewer interruptions without losing control of consequential choices. The VA works inside explicit boundaries and gains decision room as the evidence grows. Their development comes from owning more of a result, not from guessing what the founder might have meant.
The next refund request arrives. The VA has already checked the purchase record, policy, prior correspondence, and customer value. The message to the owner contains one narrow exception and a recommended response. The owner makes the business decision. The VA records it, replies, schedules the follow-up, and closes the loop.
That is work a virtual assistant can own.
Frequently asked questions
What access should a virtual assistant receive at first?
Give the VA only the access required for the responsibility, through managed accounts and a password manager where possible. Keep personal, financial, clinical, and administrator-only surfaces outside the first handoff unless the role genuinely requires them and the security controls are ready.
How long should a delegation pilot run?
Run the pilot through several normal cycles and at least one real exception. The calendar matters less than whether the VA can reach the visible finish, use the escalation path well, and apply a corrected instruction on the next cycle.
Who owns the playbook if the VA changes?
Keep the playbook under business-controlled custody, and define ownership and reuse rights in the VA agreement. The person doing the work should maintain it in a shared system so another trained person can inherit the responsibility without reconstructing the rules from old messages.
How should I measure delegation beyond hours saved?
Count routine returns to the owner, missed finishes, reopened work, and repeated exceptions. Then inspect whether escalations arrive with enough evidence for a decision. Good delegation reduces avoidable owner re-entry without hiding errors or pushing consequential choices to the VA.