About a Quarter of Virtual Assistant Roles Now Require AI
An analysis of 322 job descriptions shows AI is being written into the roles that already demand judgment, not the routine ones. Technical and analytics roles require it in 44% of cases. Finance roles require it in 5%.
Key findings
- 23.9% of roles name AI in their requirements. One in ten names a specific AI tool rather than referring to AI in general.
- Among the smallest clients the share is higher, not lower. For clients employing ten people or fewer, 26.7% of roles require AI, above the 23.9% across the whole sample.
- The roles that require AI are the complex ones, not the routine ones. Technical and analytics roles require it in 44.0% of cases. Finance and accounting roles require it in 5.3%. The gap between the top and bottom category is about eight to one.
- Administrative work sits below the average. The largest category by volume requires AI in 21.8% of roles, below the overall 23.9%.
- Workflow automation is asked for as often as chat AI, and it is a different thing. Zapier and n8n together appear in more roles than ChatGPT does alone. They are integration tools rather than AI tools and are not counted as AI anywhere in this report.
- The overall share is flat across the period. Setting aside March, a particularly small month, every month sits between 22.4% and 29.3%. What rises is specificity, on a base too thin to call a trend.
How Common Is It
| Requirement | Share of roles |
|---|---|
| Mentions AI in any form | 23.9% |
| Names a specific AI tool | 10.2% |
About a quarter of the roles a business opens today carry an expectation that the person will work with AI. This is not a forecast about where hiring is going. It is a description of what is already in the job description.
It is also new to this profession. A requirement of this kind was not a normal feature of virtual assistant roles two years ago. We do not publish a figure for what the share used to be, because our records do not support one, and a comparison we cannot evidence is not worth making.
Which Clients Ask For It
The businesses opening these roles are small, and the smallest ones are not asking for AI any less.
The client base behind these roles is almost entirely small businesses. 62.7% of clients employ ten people or fewer and 91.0% employ fewer than fifty, for the period February to August 2026.
| Client size | Share of roles requiring AI |
|---|---|
| Ten employees or fewer | 26.7% |
| All roles in the sample | 23.9% |
This matters because national surveys consistently find that the smallest firms report the lowest rates of AI adoption. In this data those same small businesses are writing AI into the roles they open at a slightly higher rate than the sample as a whole. Adoption measured at the level of the company and adoption measured at the level of the job being advertised are not the same thing.
Coverage limit, which should be read alongside this finding. An employee count is recorded for 97 of the 322 roles, which is 30.1% coverage, and we cannot rule out that the roles without the field behave differently. Larger size bands fall below our 15-role reporting threshold and are not reported, so this table says nothing about whether the requirement rises or falls as clients grow.
Which Work Requires It
This is the finding that runs against the common assumption.
| Work category | Share requiring AI |
|---|---|
| Tech and development | 44.0% |
| Analytics and software engineering | 44.0% |
| Writing and content creation | 43.3% |
| Design and creative | 36.7% |
| Social media management | 35.6% |
| Marketing and advertising | 33.0% |
| Administrative and operations support | 21.8% |
| Real estate | 20.8% |
| E-commerce management | 20.5% |
| Sales and business development | 13.0% |
| Finance and accounting | 5.3% |
The public conversation about AI and work generally starts from the premise that routine, repeatable jobs are affected first. In this data the pattern is close to the opposite. The categories where AI appears most often are the ones that already require judgment, and the categories built on defined procedure sit at the bottom.
Administrative work, the largest category by volume, sits below the overall average at 21.8%.
We are describing where employers ask for AI skills, not where AI is capable of doing the work. Those are different questions and this data answers only the first.
“AI wins on speed and simple, repeatable tasks. The moment the work needs judgment, context, or catching an edge case the AI misses, a human pulls ahead fast. In our experience the tipping point isn't task volume, it's task ambiguity. The messier the decision, the more you need a human.”
The Tools Employers Name
| AI tool | Share of all roles |
|---|---|
| ChatGPT | 7.5% |
| Claude | 4.7% |
| Gemini | 1.2% |
| Midjourney | 1.2% |
| HeyGen | 1.2% |
| Runway | 0.9% |
| Synthesia | 0.6% |
Jasper and Perplexity each appear in under 0.5% of roles and are listed here only so the table is complete. The generation tools at the bottom, Midjourney, HeyGen, Runway and Synthesia, are concentrated almost entirely in creative and social media roles.
Workflow automation is counted separately
| Automation platform | Share of all roles |
|---|---|
| Zapier | 5.6% |
| n8n | 2.5% |
Zapier and n8n together appear in more roles than ChatGPT does alone, which is a finding in its own right about what small businesses expect a new hire to be able to operate. Employers are not only asking for someone who can write a prompt. They are asking for someone who can wire a process together.
But these are integration tools, not AI tools, and we do not count them as AI. A posting that asks for n8n with no AI referenced anywhere in it is not a posting that requires AI, and counting it as one would inflate the headline in a way we could not defend to anyone who read the underlying job descriptions. An earlier version of this report did count them. Eight roles entered the AI figures on that basis and have been removed.
A note on one entry above: VA Masters uses Claude in its own operations. It appears in the table because it appears in the underlying job descriptions, and removing it would misrepresent the data. It is reported at the share the data shows, with the connection stated rather than left for a reader to find.
What Is Changing, and What Is Not
| Month | Mention AI | Name a specific AI tool |
|---|---|---|
| February 2026 | 25.0% | 4.2% |
| March 2026 | 4.8% | 4.8% |
| April 2026 | 29.3% | 9.8% |
| May 2026 | 22.4% | 10.2% |
| June 2026 | 25.4% | 8.5% |
| July 2026 | 23.0% | 13.1% |
| August 2026 | 26.9% | 13.4% |
The share is flat
Setting aside March, which is a particularly small month and which we treat as noise rather than a data point, every month sits between 22.4% and 29.3%. Over this window there is no upward trend in how often AI is mentioned, and we are not going to present one.
Specificity rises, on a base too thin to call a trend
The share of roles naming a particular AI tool moves from 4.2% in February to 13.4% in August, and the direction is consistent across the period. That is consistent with employers moving past the stage of asking for general familiarity and toward knowing which tools their workflow actually runs on.
We present it as an observation rather than an established trend, and we do not build a headline on it. February rests on a single role, and the number of roles opened in a month ranges from the low twenties to the high sixties. An earlier version of this report described this movement in raw counts, which overstated how firm it is.
What we looked for and did not find
No growth trend in AI mentions. We tested this month by month and the line is flat. Seven months is a short window and the monthly base is small, so a modest change could be hidden inside the noise. But we found no trend, and we are not going to describe one that the data does not show.
No relationship between AI requirements and engagement size. Roles that require AI are opened at part-time and full-time hours in about the same proportion as roles that do not. AI in the requirements is not a marker of a bigger engagement.
No relationship between AI requirements and time to fill. Roles requiring AI do not take measurably longer or shorter to fill.
No answer on the larger client size bands. Only 30.1% of roles carry a recorded employee count, and once the smallest band is set aside the remainder falls below our reporting threshold. We can say the requirement is present in full among the smallest clients. We cannot say whether it rises or falls as clients get bigger.
We report these because a relationship that turns out not to exist is still a result, and because it shows which questions were tested rather than only which ones produced an answer.
For journalists
How to cite this data
Attribute figures to “VA Masters, AI in Virtual Assistant Job Requirements 2026” and link to this page. Always carry the data period with the figure, for example “23.9% of roles require AI, February to August 2026”. Charts and tables may be reproduced with attribution.
If you cited the earlier figures
Figures published before 15 September 2026 counted workflow automation platforms as AI tools. The revised figures are 23.9% and 10.2%, replacing 26.4% and 14.6%. We would rather flag this ourselves than have you find it. Details in the Methodology section.
Media contact
Interviews, additional data cuts, and methodology questions:
[email protected]
+1 (312) 766-0301
Press room & full dataset
Press coverage & media kit · State of the Filipino VA Industry 2026
Boilerplate
VA Masters is a global recruitment agency founded in 2021 that places rigorously vetted Filipino virtual assistants with businesses across the US, UK, Australia, Canada and Europe. Over 1,000 placements to date.
Company facts
Founded: 2021 · Founder & CEO: Alon Pearl · Placements: 1,000+ · Vetting: 6 stages
Questions & Answers
Common questions about AI requirements in virtual assistant hiring.
What share of virtual assistant jobs require AI skills?
23.9% of virtual assistant roles opened between February and August 2026 name AI in their requirements, and 10.2% name a specific AI tool rather than referring to AI in general. That is about one role in four.
Do small businesses ask for AI skills less often?
Not in this data. Among clients employing ten people or fewer, 26.7% of roles require AI, slightly above the 23.9% across the whole sample. National surveys consistently find that the smallest firms report the lowest rates of AI adoption, so it is worth being precise: adoption measured at the level of the company and adoption measured at the level of the job being advertised are different things. An employee count is recorded for 30.1% of roles in this sample.
Which virtual assistant roles require AI the most?
The technical and analytical ones. Tech and development roles and analytics and software engineering roles both require AI in 44.0% of cases. At the other end, finance and accounting roles require it in 5.3%, a gap of about eight to one.
Is AI replacing routine virtual assistant work?
This data does not show that. It shows where employers ask for AI skills, which is a different question from where AI can do the work. The pattern runs opposite to the common assumption: AI is written into the roles that already demand judgment, while categories built on defined procedure sit at the bottom of the table. Administrative work, the largest category by volume, sits below the overall average.
Which AI tools do employers ask virtual assistants to know?
ChatGPT leads at 7.5% of all roles, followed by Claude at 4.7% and Gemini at 1.2%. Zapier at 5.6% and n8n at 2.5% appear frequently but are workflow automation platforms rather than AI tools, and this report does not count them as AI.
Are AI requirements in job descriptions increasing?
Not in this window. The share of roles mentioning AI is flat between February and August 2026, sitting between 22.4% and 29.3% every month once March is set aside. The share naming a particular tool moves from 4.2% in February to 13.4% in August, but February rests on a single role, so we report that as an observation rather than a trend. We do not publish a figure for what the share used to be before this window, because our records do not support one.
Do AI requirements make a role harder to fill?
No. Roles requiring AI do not take measurably longer or shorter to fill than roles that do not, and they are opened at part-time and full-time hours in about the same proportion.
What is this analysis based on?
322 virtual assistant roles opened between February and August 2026 that carry a full role description, taken from the VA Masters positions database. A role counts as requiring AI if its description or skills field mentions AI in general terms or names a specific AI tool. Workflow automation platforms such as Zapier and n8n are not counted as AI tools.
Were these figures revised?
Yes. On 15 September 2026 we corrected an error that counted workflow automation platforms as named AI tools. Eight roles had been counted as requiring AI purely because they named a platform such as Zapier or n8n, with no AI referenced in the description. The headline moved from 26.4% to 23.9% and the named-tool figure from 14.6% to 10.2%. Anyone who cited the earlier figures should use the revised ones.
Can I cite these figures?
Yes. Attribute figures to “VA Masters, AI in Virtual Assistant Job Requirements 2026”, link back to this page, and carry the data period with the figure. For additional cuts or an interview, contact [email protected].
Methodology & Definitions
Source
The VA Masters positions database. 322 roles opened between February 2026 and August 2026 that carry a full role description.
Window
February 2026 is the first month in which role descriptions reach complete coverage in our system. Records before that month are largely missing descriptions, so any figure extending backwards would show a rise created by our own record keeping rather than by the market. For that reason no comparison to 2025 appears anywhere in this report.
How a role was counted
A role counts as requiring AI if its description or its skills field mentions AI in general terms (artificial intelligence, LLM, prompting, generative, machine learning) or names a specific AI tool. Matching used whole-word boundaries so that ordinary words containing the letters “ai” were not captured.
Workflow automation platforms are not counted as AI
Zapier, n8n and Make.com are integration tools. A role naming one of them with no AI referenced anywhere in the description is not counted as requiring AI. This rule removed eight roles from the AI count and is applied to every figure in this report. Separately, 25.5% of roles mention automation in some general form. That figure is also kept out of the AI number, because “automation” covers a much broader set of things including basic spreadsheet and email automation.
Revision of 15 September 2026
The version of this report published in September 2026 counted workflow automation platforms as named AI tools. The revised figures and the figures they replace are set out below so that anyone who cited the earlier version can see exactly what changed.
Overall share mentioning AI: 23.9%, previously 26.4%.
Share naming a specific AI tool: 10.2%, previously 14.6%.
Analytics and software engineering: 44.0%, previously 60.0%.
Tech and development: 44.0%, previously 58.8%.
Marketing and advertising: 33.0%, previously 38.1%.
Administrative and operations support: 21.8%, previously 22.3%.
Sales and business development: 13.0%, previously 14.5%.
Unchanged: writing and content creation 43.3%, design and creative 36.7%, social media management 35.6%, real estate 20.8%, e-commerce management 20.5%, finance and accounting 5.3%.
The movement in specificity was previously described in raw counts as rising from one role in February to thirteen in August. Under the corrected rule it is one role in February against nine in August, or 4.2% against 13.4%, and it is now presented as an observation rather than a trend.
Client size
Roles were joined to the client record and to the recorded employee count. The field is populated for 97 of the 322 roles, which is 30.1% coverage. Size bands other than ten employees or fewer fall below the reporting threshold and are not published.
Who writes the descriptions
Role descriptions are produced by a VA Masters recruiter working from a briefing call with the client, and are approved by the client before the role opens. They therefore represent a requirement the client has agreed to, not necessarily one the client raised unprompted. This distinction is the reason we describe the finding as what roles require rather than what clients are asking for.
Reporting threshold
Category figures are shown only for categories with at least 15 roles. Other, legal and compliance, and HR and recruiting fall below the threshold and are excluded rather than reported with a small sample.
Taxonomy
Bookkeeping Assistant is reported within finance and accounting, Video Editor within design and creative, PPC Manager within marketing and advertising, and Data Science within analytics and software engineering. Roles carrying more than one category appear in each category they carry, so the category shares do not sum to the total.
March 2026
March reads 4.8% against 22–29% in every other month, on a small monthly base. It is treated as noise rather than a data point. It appears in the table because removing it would be selective, and the text says explicitly that it is being set aside.
Exclusions
Roles opened by VA Masters for its own internal team. Roles without a description. Absolute counts of roles by category or by month. Client names, individual role details, and any figure relating to pay.
About this data
This report describes what passes through VA Masters, not the offshore hiring market as a whole. Every dataset of this kind reflects the business that produced it. Our client base skews toward the companies that find us, and most arrive on their own through search, through AI assistants, and through referrals from clients we already work with. That shapes what appears here.
We think the sample still says something real. It covers hundreds of roles across eleven reported categories and clients on four continents, and the patterns are consistent rather than driven by a handful of accounts. But a reader should treat it as a detailed view of one large pipeline, not a measurement of national demand. One bias is worth marking explicitly: clients who reach us through search and through AI assistants are likely to be more comfortable with technology than the average small business, which may lift the figures here above what a random sample would produce. Where a finding is unusually specific to our client mix, we say so in the section itself rather than leaving it to this note.
Get the full report & data pack
Enter your details below and the complete PDF is yours — every breakdown, the full methodology, and all the figures behind this report, ready to cite.
- The complete State of the Filipino VA Industry 2026 report as a PDF
- All the data behind the charts, cleared for citation with attribution
- Optional: request a custom data cut or an interview with our leadership team
Journalists and researchers welcome. Prefer email? Reach us at [email protected].