Which graphic design tasks AI takes from a studio, and which it cannot
SRF Capital Studio Research DeskFunding Intelligence, SRF Capital StudioAI is not taking design work in one sweep. It is taking the repeatable stages a studio used to bill for, one task at a time, and leaving the judgement calls more valuable than before.
Summary
- AI in graphic design works task by task: it strips out repeatable stages such as reference hunting, first-pass layouts and template assets, and leaves judgement work largely alone.
- Designers already use it widely, but by Figma's own survey only about a third trust what it produces, which is where a studio's value now sits.
- Studios should stop billing for stages a client can generate, and reprice around direction, research and systems before procurement teams force the change.
Most studio owners we speak to frame AI as one event that will either arrive or not. It has already arrived, and it did not arrive as one thing. It arrived as a dozen small tools, each of which removes a specific stage of work that used to show up on an invoice.
Figma's 2025 AI Report, a survey of 2,500 of its users run in January 2025, puts numbers on the mood. Seven in ten designers (72%) use generative AI tools, and 78% say it makes them more efficient. Yet only 47% say it makes them better at their job, and only 32% trust its output.
That gap between speed and trust is the most useful fact a studio owner has. AI makes the making cheaper. It does not make the deciding any cheaper, and clients are starting to notice the difference.
The task map
Where AI sits in a design studio's workflow, task by task
| Studio task | What AI does to it now | What that means for billing |
|---|---|---|
| Mood boards and reference search | Image models produce dozens of directions in minutes | Curating direction still has value; the hours of searching do not |
| Wireframes and first layouts | Tools such as Figma's AI features draft layouts from a brief | A starting point for design decisions, no longer a billable stage |
| Social templates and basic graphics | Canva's Magic Design lets a marketing team self-serve | Only campaign thinking justifies an agency fee |
| Icons, illustrations, variants | Variations at close to zero marginal cost | Per-asset pricing is falling and will keep falling |
| User research synthesis | Helps tag transcripts and spot patterns | Research design and the insight itself stay human |
| Design systems | Generates component variants quickly | More components means more need for someone to govern the system |
| Brand strategy and positioning | Little meaningful effect | Pricing power rises as production prices fall |
| UX strategy and information architecture | Little meaningful effect | Grows in value as tactical UI work is automated |
The pattern is simple enough to use as a rule. Where a task can be specified fully in a brief, a tool can increasingly do it. Where the task is deciding what the brief should say, a tool cannot, because the answer depends on this client, this market and this year.
AI makes the making cheaper. It does not make the deciding any cheaper.
Where the common view goes too far
The common line is that mood boarding and social templates are simply "eliminated". We think that overstates it. The searching is gone. The editorial eye that picks three directions out of fifty, and explains to a founder why the other forty-seven are wrong for the brand, is worth more when the options are endless.
The honest version for a studio is this: the stage is no longer billable as hours, but the judgement inside it can be billed as direction. Studios that fold that judgement into a fixed creative-direction fee keep the value. Studios that keep invoicing "mood board, 12 hours" lose it.
What a studio looks like after the shift
As an illustration, take a 20-person studio in 2020: four strategists, a dozen visual, UI and production designers, and four support staff, earning perhaps ₹15 lakh to ₹25 lakh per head. A plausible shape for the same revenue a few years out is about 12 people. Five strategists, four senior designers who direct AI output and own quality, two design engineers who build the workflows and systems, and one operations lead.
On that model, revenue per head roughly doubles. The people who leave are the ones hired mainly for execution speed. That is a scenario, not a forecast, but the direction matches what studios tell us about their hiring: fewer juniors, more seniors, and a new kind of hire who is half designer and half engineer.
What to do this quarter
- Audit last year's invoices by task. Tag every line as repeatable or judgement. The repeatable share is revenue at risk, and it is usually larger than owners expect.
- Stop quoting stages a client can generate. Replace them with a direction fee or fold them into a fixed-scope deliverable where the speed gain becomes your margin.
- Put the time saved into research and systems. Those are the tasks clients cannot self-serve, and they lead naturally to retainers.
- Write down your AI policy. Which tools, on which work, with what disclosure to clients. Procurement teams have started asking, and ownership of generated assets is still unsettled, a risk we cover in the risks that break design agencies.
Repricing is the hard part, because the instinct is to pass the savings on. Before you cut a fee, read your rate card is not your price. For where AI sits in the wider sector economics, see the design and creative industry report, and for how the four formats differ, what design consulting sells.
The tools will keep improving. The studios that do well will be the ones that stopped selling the parts of the job that were never the point.
How useful was this article?
One tap. It tells us what to write more of.
About the author
SRF Capital Studio Research Desk
Funding Intelligence, SRF Capital Studio
The next one
Get what we publish next, by email.
Working notes on raising, borrowing, protecting, growing and structuring capital in India. One email a week at most, and you can leave any time.
We use your address only to send this. See our privacy policy.
We store your address to send you these emails and nothing else. See our privacy policy.
Related reading