Fahad Manzur
AI strategy and consulting work on Upwork grew 51% year over year in the second quarter of 2026, in a quarter where the marketplace as a whole shrank 4% and active clients fell to 763,000. An AI strategy consultant, in Upwork’s sense of the category, is the person a small business hires before any build starts: to decide which of its workflows should be automated at all, which tools to use, what it will cost, and how to roll it out without breaking the team. Upwork’s own hiring page prices that role at $35 to $60 an hour, with the most experienced consultants above $100.
What Upwork actually reported in August
The numbers come from Upwork’s second quarter 2026 results, published on August 10. Gross services volume was $966.4 million, down 4%. Revenue was $191.7 million, down 2%. Against that backdrop, GSV from AI-related work rose more than 22%, reaching roughly $330 million on an annualised basis, and the AI Strategy and Consulting sub-category grew over 50%.
The more useful detail is in the prepared remarks from the earnings call. CEO Hayden Brown described the constraint SMB clients are running into: they have access to the technology but not to people who know how to apply it to their specific industry, workflow, or business problem. Her words were that SMBs are hiring AI experts to guide them in making their businesses more AI native. That is advisory work, and it is being bought by the client segment Upwork is investing in hardest, the Business Plus tier, where GSV rose 174% year over year.
The demand that never shows up in the job post
One statistic in the remarks deserves its own paragraph. In a recent Upwork survey, nearly half of talent said their most recent job was AI-related, yet only 16% of those jobs mentioned AI in the posting. Clients now assume AI will be used and no longer bother to say so.
For a freelancer this changes how you read the job feed. A post asking for “a content calendar for a Shopify store” or “cleaning up our customer support macros” is, in 2026, very often an AI implementation job that the client has not labelled. The person who wins it is the one whose proposal names the tool, the workflow, and the measurable result, rather than the one who waits for a post with “AI” in the title. The remarks also list the other growth area: clients arriving with a half-finished AI project and asking someone to turn a vibe-coded concept into a production website or humanise translation output. Both are consulting problems before they are build problems.
Six things an AI strategy consultant does for a small business
Strip away the title and the work breaks into a small number of repeatable deliverables. A workflow audit: sit with the owner, map the ten most repetitive tasks in the business, and rank them by hours saved and risk. A tool selection memo: for each candidate task, recommend one tool, its monthly cost, and what data it will touch. A pilot plan: pick one workflow, define what “working” means in numbers, and run it for two to four weeks. A rollout and training session for the two or three staff who will actually use the system. A data and privacy note, since most SMBs have no idea what they are uploading to which vendor. And a 90-day review, which is where the retainer comes from.
None of these require a machine learning background. They require knowing the current tool landscape well, being able to write clearly, and being honest when a task should not be automated. If you have already delivered automation builds, outcome-based pricing maps neatly onto the pilot and review stages, because both have a number attached.
Is it realistic for a Pakistani freelancer to charge consulting rates?
Upwork’s AI consultant hiring page gives clients a range of $35 to $60 per hour and notes the most experienced consultants charge $100 or more. That is the buyer-facing number, so it is the ceiling a client has been told to expect before they read your profile. It also sits well above what routine execution work in the same niche pays, a gap explored in the site’s look at where AI freelance rates are heading.
The difference is the buyer. An execution task is compared on price against every other bidder. A consulting engagement is compared on trust, and trust is built with evidence: a short written audit of a real business (your own, a relative’s shop, a former client), a one-page pilot plan with real before-and-after figures, and a profile that speaks the client’s industry language rather than tool jargon. Two or three of those artefacts, done properly, are the whole portfolio.
How to position for the work this month
Start by picking one industry you already understand, for example ecommerce operations, property management, a dental clinic, a logistics broker. Rewrite your Upwork title and overview around that industry’s problems, not around the tools you know. Build the audit template once and offer it as a fixed-price entry product; a modest fixed fee for a first audit filters for clients who are serious and gives you a paid reason to be inside their business.
Then read job posts for the unlabelled AI demand described above. When a post asks for a manual task, propose the automated version with a specific tool and a specific saving, and offer to run it as a paid pilot. Because Business Plus clients are the ones buying this work, it is worth understanding what Business Plus changes for freelancers and how those clients search.
Finally, be aware that Upwork is pushing hard on where clients arrive from. The company embedded its marketplace inside ChatGPT and Claude during the quarter and launched its MCP server, which means a growing share of consulting briefs will be scoped by an AI assistant before a human ever posts them. A profile written in plain, specific language about business outcomes is what those systems match on.
The caution in the same report
Upwork cut its full-year revenue guidance to $730 million to $750 million and said plainly that lower-complexity work continues to burn off as it is automated. The 51% growth in AI strategy is real, but it is growth in a category that is still a fraction of the marketplace, and it is being bought by fewer, larger clients. The practical reading is that the execution-only freelancer is being squeezed from both sides, by automation below and by consulting above. Moving one step up the chain, from doing the task to deciding which tasks get done, is the move the data supports.
