With 250+ consumer engagements throughout each company and embedded in-house roles, Vitalii Khomenko has seen product groups from the within in a approach few designers have. Right here he makes the case for a brand new form of design workflow constructed for the AI period, and explains why the window to adapt is narrower than most founders assume.
For the final decade, each startup dialog about pace finally turned a dialog about engineering. Hiring extra builders, reducing scope, selecting a less complicated stack, it was virtually all the time a dev-side downside. Vitalii Khomenko, Principal Designer at Orizon Design and founding father of Fluente LLC, argues that is now not true. And he has a selected view on what groups ought to do as an alternative.
What makes Khomenko’s perspective price taking severely is not simply seniority. It is the precise form of his expertise. Over a profession spanning 250+ consumer engagements, he has labored each as an exterior specialist introduced in to ship and as an embedded designer who turned the only connective thread between product, model, and execution. That twin vantage level is uncommon. Most designers dwell on one aspect of that line. Khomenko has operated extensively on each, which provides him an unusually clear learn on the place friction truly originates inside product groups, and the place it would not.
His work spans early-stage startups discovering their first visible language and established manufacturers managing complicated, multi-disciplinary design techniques. Throughout that vary, sure patterns repeat. And the sample that began changing into inconceivable to disregard this 12 months pointed him towards a conclusion that’s reshaping how he thinks a contemporary design workflow ought to truly be structured.
The bottleneck did not disappear. It moved. And the workflow wants to maneuver with it.
The Drawback With How Design Groups Are Presently Working
When transport code was sluggish, the prevailing design workflow made sense. A number of rounds of revision, consumer testing, iteration, all of that match naturally into the window whereas engineering caught up. That window is now gone, and most design workflows had been by no means up to date to mirror it.
“Founders are asking design groups to maneuver on the identical pace as an AI-assisted dev pipeline, and design would not compress the identical approach code does,” Khomenko says. The result’s a workflow that was constructed round a constraint that now not exists, utilized to a pipeline that has essentially modified round it.
The pressured selection between polish and pace that product groups are actually navigating is not new as an idea. What’s new is how stark it has grow to be, and the way straight it traces again to a workflow downside moderately than a individuals downside. “As a result of growth is affordable and quick now, there’s fixed strain to ship the following factor instantly,” he explains. “However the merchandise that really retain customers are nonetheless those the place the small print had been thought-about fastidiously. Fewer revision rounds is not a course of enchancment if it means transport avoidable friction.”
Khomenko has watched this play out throughout sufficient groups, at sufficient levels, to be particular about the place the prevailing workflow breaks down. The difficulty is not design judgment. It is the place that judgment sits within the course of, and what format it is anticipated to supply on the finish.
Why Figma Handoffs Are Now the Bottleneck
To know Khomenko’s proposed repair, it helps to grasp precisely the place the present workflow stalls. For many product groups, the top of the design course of appears like this: a designer produces polished, dev-ready screens inside Figma, and fingers that file to engineering. Engineering then interprets these screens, interprets the visible selections into code, and builds from there.
That translation step, Figma file to working code, is the place the pipeline now breaks down. AI-assisted engineering has made the constructing half dramatically sooner. However the translation step hasn’t modified. An engineer nonetheless has to learn a Figma file, interpret spacing, extract belongings, make judgment calls about part construction, and write the implementation from scratch. That course of takes the identical period of time it all the time did.
“The handbook workflow inside Figma merely is not maintaining with AI-boosted engineering,” Khomenko says. “The design handoff is quietly changing into the factor everyone seems to be ready on.”
The result’s a pipeline the place the whole lot across the handoff has accelerated, and the handoff itself has grow to be the only level of drag. Figma, in different phrases, was constructed for a world the place engineering was the sluggish half. It wasn’t designed to feed a pipeline the place engineering strikes at AI pace.
The Repair: Closing Supply in Code, Not in Figma
That is the core of what Khomenko is proposing, and it is a extra vital structural shift than it’d first seem.
In his workflow, Figma would not disappear. It strikes. Quite than serving as each the exploration surroundings and the ultimate supply artifact, Figma turns into the visible exploration layer solely: the place the place model id is established, format logic is labored by, and design course is validated. It stays the appropriate software for determining what one thing ought to feel and look like.
However the last supply, the artifact that really goes to engineering, is code. Particularly, parts constructed into ready libraries that engineering can work from and scale straight, with none translation step in between.
The distinction in implementation pace is the purpose. When a designer delivers a coded part library moderately than a Figma file, engineering would not have to interpret something. The choices are already expressed within the language they’re constructing in. What beforehand required an engineer to learn, interpret, and re-create from a visible reference now arrives prepared to make use of. In a pipeline operating on AI-assisted growth, that distinction compounds quick.
“Figma remains to be the appropriate place to discover what one thing ought to feel and look like,” Khomenko says. “However handing over a Figma file and asking engineering to implement it’s now the sluggish path. Delivering in code is the quick path. That is the shift.”
The Designer This Requires
Executing this workflow requires a special form of designer, and Khomenko is direct about that. The UI/UX designer changing into a Design Engineer is, in his evaluation, not a distant risk. It’s a matter of time, and the groups treating it as a future consideration moderately than a gift actuality are those probably to seek out themselves caught on the handoff.
However Khomenko is equally clear about what this shift doesn’t suggest. Delivering in code is one a part of the function, not the entire of it. The complete design cycle is not going wherever. Making use of branding to imagery, producing animations, constructing advertising and marketing supplies, these stay important inventive duties that require a broad and different toolkit nicely past code. “The designer of the close to future is not simply producing code,” he says. “They’re shifting throughout the total cycle, model, movement, advertising and marketing, product, with design engineering as an vital and rising a part of that stack, not a alternative for it.”
What modifications is that code-ready supply turns into a core expectation inside that stack. The designer who can transfer fluidly from visible exploration in Figma to part supply in code, whereas additionally carrying the model considering, movement, and advertising and marketing creativity that differentiates merchandise on the floor degree customers truly expertise, is the profile that the brand new workflow calls for.
An Sincere Be aware on The place Issues Stand
Khomenko is candid that no single model of this workflow has totally standardized but. Each workforce is figuring it out otherwise. The instruments are evolving, the handoff expectations are shifting, and the processes groups are constructing round all of it differ significantly. There isn’t any off-the-shelf playbook.
However that is exactly the place his 250+ engagements grow to be greater than a credential. They are a dataset. Having watched sufficient groups, at sufficient levels, by sufficient cycles of tooling change, Khomenko can see by the noise of particular person variation and establish what’s directionally true even when the specifics are nonetheless in flux. “Every workforce’s workflow is de facto customized at this stage,” he says. “However some normal patterns of how that is going to look within the close to future are already seen.”
These patterns constantly level in the identical course: design judgment shifting earlier, last supply shifting into code, and the designer’s function increasing in technical vary whereas staying broad in inventive scope. The precise instruments and sequences will maintain shifting. The underlying logic of the workflow Khomenko is describing is already taking form within the groups paying closest consideration.
What Founders Ought to Do Now
For founders and product leads, the workflow Khomenko is describing has direct implications for the way groups are staffed, budgeted, and structured.
Entrance-load design judgment. The largest leverage within the new workflow is not sooner execution, it is higher downside definition earlier than technology begins. Carry senior design considering in earlier, when it could form what will get constructed moderately than polish what’s already accomplished.
Separate exploration from supply. Figma and the visible exploration part serve a definite function from the ultimate supply layer. Deal with them as separate levels with separate expectations, moderately than collapsing them right into a single step anticipated to supply the whole lot.
Shift the supply expectation to code. Whether or not meaning hiring for Design Engineer vary, growing it internally, or restructuring handoffs, the top of the design pipeline must arrive in a type engineering can work from straight. The interpretation step is the place time goes to die, and it is now probably the most avoidable a part of the method.
Defend the total inventive stack. Design engineering issues greater than it ever has. It isn’t, nonetheless, the entire job. Groups that cut back design to code technology will lose the model considering, movement, and advertising and marketing creativity that differentiates merchandise on the floor degree customers truly expertise.
Keep versatile. The workflow remains to be forming. Groups that construct rigidity into their course of now will discover themselves redoing that work inside a 12 months. The groups getting this proper are those constructing sufficient adaptability into their buildings to soak up what’s nonetheless coming.
“The Figma handoff made sense when engineering was the sluggish half,” Khomenko says. “Engineering is not the sluggish half anymore. Delivering in code is not the way forward for design workflow. For the groups paying consideration, it is already the current.”
