Artificial intelligence is changing how products are designed, built, tested and experienced.
The most visible change is speed.
Teams can now generate ideas, write code, create designs, draft content and analyse information far faster than before. Work that once required several people or several days may now be completed in a fraction of the time.
That increase in productivity is real.
But it also creates a more difficult question:
Does 10x the output produce 10x the value?
The answer is not automatic.
AI can increase quantity, but quality still depends on judgement, context, customer understanding and disciplined execution.
This is the central challenge for product teams entering the next phase of AI adoption.
1. The productivity gain is real
AI can significantly increase the amount of work a team produces.
A product manager can explore several problem statements at once. A designer can create multiple concepts in parallel. An engineer can accelerate implementation, debugging and documentation. A marketing team can test more messages and formats.
This changes the economics of product development.
The bottleneck is no longer always the ability to create the first version of something. Increasingly, the bottleneck is deciding which version deserves to move forward.
When production becomes easier, selection becomes harder.
Teams may generate ten options instead of two, but they still need to know which one best solves the customer’s problem.
That requires more than speed.
It requires judgement.
2. The quality-versus-quantity debate is real
AI makes it easy to produce a large volume of work.
The danger is that teams begin to confuse activity with progress.
More designs do not necessarily create a better experience. More features do not necessarily create a better product. More content does not necessarily create more trust.
The ability to generate quickly can create pressure to publish quickly.
This is where experienced people become essential.
They can recognise when an output is incomplete, generic or poorly aligned with the real customer need. They can identify when another iteration is required. They can challenge an answer that appears confident but lacks depth.
Most importantly, they do not take AI-generated output for granted.
AI should accelerate thinking, not replace it.
The strongest teams will use AI to expand the range of possibilities while maintaining a high standard for what reaches the customer.
3. Product onboarding must deliver value faster
Traditional onboarding often assumes that customers are willing to invest time before experiencing value.
They are asked to watch tutorials, complete guided tours, read documentation or configure several settings.
That approach is becoming less effective.
Customers increasingly expect to understand the benefit of a product within the first few minutes.
This changes the purpose of onboarding.
The goal is no longer to explain the entire product. The goal is to help the customer achieve one meaningful outcome as quickly as possible.
That may mean reducing the number of steps, removing unnecessary explanations and guiding the user directly toward the moment where the product becomes useful.
The first five minutes matter more than the first fifty features.
4. AI may become the interface
Another important shift is that users may not always interact directly with a product’s interface.
Instead, they may use an AI assistant to complete tasks on their behalf.
This possibility raises fundamental product questions.
What happens when the user no longer navigates through menus and screens?
What happens when the customer expresses an intention in natural language and expects the system to complete the workflow?
Does the visual interface become less important?
Not entirely.
The product experience and the AI interaction both matter.
A strong visual interface creates trust, clarity and control. A strong AI interaction creates speed, convenience and accessibility.
The best products will not treat these as competing priorities.
They will design a coherent experience across both.
The user may begin with a conversation, move into a visual workflow, review the result and return to the AI for refinement.
The interface will become more flexible, but the need for thoughtful design will remain.
5. Product teams may work more directly with customers
AI may also change how teams validate decisions.
Traditional A/B testing will still have a place, especially where teams have enough traffic and a clearly measurable outcome.
But product development may increasingly favour more direct collaboration with customers.
Teams can use AI to prototype faster, place early concepts in front of users sooner and iterate based on richer qualitative feedback.
This can shorten the distance between the customer problem and the product response.
Instead of spending weeks preparing a polished experiment, teams may create a working concept in days and review it with customers immediately.
This approach is not less rigorous.
It simply shifts the source of learning.
The strongest teams will combine data with direct customer understanding rather than relying on one method alone.
6. Speed is not a defensible moat
If every company can access similar models, speed alone will not create a sustainable advantage.
A competitor may use the same tools, generate similar features and move at a similar pace.
The real question is what makes the product difficult to replace.
A defensible moat might include:
- A strong community around the product
- A compelling value proposition
- Proprietary context or data
- Deep integrations with customer workflows
- Trust built through consistent delivery
- A product experience that becomes part of how customers work
These advantages take time to build.
They are often based on relationships, accumulated knowledge and deep understanding of a specific problem.
AI can accelerate execution, but it cannot instantly reproduce the trust, context and loyalty a company has earned.
7. Design systems can become either an accelerator or a bottleneck
As AI-generated design and code become more common, the design system becomes increasingly important.
A strong design system gives teams clear standards, reusable components and a shared language. It helps AI-generated output remain consistent with the product.
A weak or restrictive design system can become a bottleneck.
Teams may be able to generate ideas quickly but struggle to convert them into a coherent, scalable experience.
The solution is not to copy successful products feature by feature.
Teams can take inspiration from tools such as Lovable, Figma, Claude Code or Linear, but the final experience must reflect the needs of their own customers.
The best design systems will support faster experimentation without sacrificing consistency.
8. Human judgement becomes more valuable, not less
There is a common assumption that as AI improves, human expertise becomes less important.
In many areas, the opposite may be true.
When the cost of generating output falls, the value of good judgement rises.
Teams need people who can distinguish a strong idea from a plausible one. They need leaders who know when to move quickly and when to slow down. They need specialists who can recognise subtle quality issues that an inexperienced person may miss.
Experience provides context.
It helps people understand what customers really mean, what risks are hidden and what trade-offs matter.
AI can support these decisions, but it does not remove responsibility for them.
9. A practical operating model for AI-enabled product teams
Product teams can respond to these changes by adopting a few practical principles.
Start with the customer outcome
Do not begin with what the AI can generate.
Begin with the problem the customer needs to solve.
Deliver value early
Design onboarding around the first meaningful outcome, not a complete explanation of the product.
Use AI to increase options
Let AI help create alternatives, explore scenarios and accelerate first drafts.
Apply human judgement before delivery
Review the output critically. Check its accuracy, relevance, consistency and value.
Collaborate directly with customers
Use faster prototyping to shorten feedback cycles and learn sooner.
Invest in the moat
Build community, context, trust, integrations and workflows that competitors cannot easily reproduce.
Treat the interface and the AI interaction as one experience
Design how they work together rather than treating them as separate products.
Conclusion
AI can generate 10x the output.
That is an important capability, but it is not the final objective.
The goal is not to produce more for the sake of producing more.
The goal is to create better outcomes for customers.
That requires speed, but also judgement.
It requires automation, but also understanding.
It requires new tools, but also strong product fundamentals.
The teams that succeed will not be the ones that simply generate the most.
They will be the ones that know what is worth building, when another iteration is needed and how to turn greater productivity into meaningful customer value.
AI changes how the work gets done.
It does not change the responsibility to build something genuinely useful.
Aleks Vladimirov
Senior Manager at Oracle | Engineering Leader in AI, Cloud and Product Development
