AI in my work
Using AI wisely
How I use AI to work faster and more efficiently, without relying on it blindly.
Topic
AI in my work
Reading time
3 minutes
Covers
Research, ideation, validation
Visuals
Original illustrations
Where AI fits in my process
My workflow
I lead
Frame the problem
Who is struggling, and why.
→
AI assists
Research
Summaries and themes, checked against sources.
→
AI assists
Ideate
Many options, fast. I choose.
→
I lead
Design and decide
Trade-offs and craft are mine.
→
AI assists
Validate
Pressure-test ideas and find gaps.
→
I lead
Ship and review
Work with engineering and own the result.
AI helps with the first pass. The framing, the decisions, and the accountability stay with me.
Where AI assists in my process, and where I lead.
A fast assistant, not a decision maker
I use AI every day to make my process faster and more efficient, and I treat it like a fast, well-read assistant: useful for a first pass, never the final word. I do not rely on it blindly. I check what it gives me, question it, and decide. The thinking that makes a design good, understanding the problem, weighing trade-offs, and taking responsibility for the result, stays with me.
Research, faster and checked
I use AI to summarize interviews, group feedback into themes, and scan large amounts of material. Then I check what it gives me against the sources. Anything that shapes a decision gets traced back to what people actually said.
Research, faster and checked
Sample research
Interview notes
P1
"I check three tools every morning."
P2
"I never know which item is urgent."
P3
"I just want one list."
→
Themes, drafted by AI
Too many places to check
Evidence
P1
P3
Verified
Priority is hard to see
Evidence
P2
Verified
People want a daily summary
Evidence
P3
Needs verification
I trace every theme back to what people actually said before it shapes a decision.
AI drafts the themes. I verify each one against the notes.
Ideate quickly, then choose
When I am exploring, AI helps me generate many options fast: layouts, wording, edge cases. I treat the output as raw material. I pick what is worth pursuing, combine ideas, and throw most of it away.
Many options, then a choice
Ideation
Prompt
Show six layouts for a daily summary screen for team leads.
Option 1
Dropped
Option 2
Kept
Option 3
Dropped
Option 4
Combined
Option 5
Dropped
Option 6
Combined
I treat the output as raw material. Most of it gets thrown away, and the choices are mine.
Most options get dropped. The choices are mine.
Pressure-test ideas before others do
Before I show an idea to a team, I ask AI to poke holes in it: what could confuse a first-time user, what happens when data is missing, where accessibility could break. It surfaces questions I might have missed. I decide which ones matter.
Poke holes before others do
Sample review
What I asked
What AI flagged
My call
What could confuse a first-time user?
The word "flagged" is unclear.
Changed
What happens when there is no data?
No empty state is defined.
Added
Where could accessibility break?
Status relies on color alone.
Fixed
What is missing from this flow?
There is no way to undo.
Not now
It raises questions I might have missed. I decide which ones matter.
AI raises the questions. I make the call.
Speed up the routine work
Documentation, sample content, renaming layers, and repetitive edits take time and need little judgment. AI does the first pass and I review it. The time I get back goes to talking with users, working with engineers, and refining the details.
How much AI helps, task by task
My own view
Task
How much AI helps
Summarizing interviews and feedback
Drafting copy and wording options
Creating realistic sample content
Writing documentation
Renaming and organizing layers
Making design decisions
Always me
Talking with users and engineers
Always me
How much AI helps varies by task, and some tasks stay entirely mine.
The thinking stays human
Design thinking is still mine. I frame the problem, make the decisions, and stand behind them. AI does not know the customer, the constraints, or the context the way the team does. I carry the same belief into the AI features I design: they appear where people need them, and a person approves the result.
What stays with me
Human judgment
AI helps with
First pass
Summaries and first drafts
Alternative layouts and wording
Spotting gaps and edge cases
Repetitive edits and documentation
I own
Final say
Framing the problem
Design decisions and trade-offs
Talking with users and engineers
Quality, taste, and accountability
Same belief
The AI features I design follow it too: placed where people need them, with a person approving the result.
What AI helps with, and what I own.
AI speeds up the work. The thinking stays human.
Next article
Built to scale →
Product design leader. Design functions,
systems, and products for complex SaaS.
Case Studies
Web design system
Native design system
Unified Inbox
AI capability
About
About
How I Work
Résumé
Contact
aln.dohotaru@gmail.com
© 2026 Alina Dohotaru
Visuals are original illustrations and do not depict proprietary product screens.
AI in my work
Using AI wisely
How I use AI to work faster and more efficiently, without relying on it blindly.
Topic
AI in my work
Reading time
3 minutes
Covers
Research, ideation, validation
Visuals
Original illustrations
Where AI fits in my process
My workflow
I lead
Frame the problem
Who is struggling, and why.
→
AI assists
Research
Summaries and themes, checked against sources.
→
AI assists
Ideate
Many options, fast. I choose.
→
I lead
Design and decide
Trade-offs and craft are mine.
→
AI assists
Validate
Pressure-test ideas and find gaps.
→
I lead
Ship and review
Work with engineering and own the result.
AI helps with the first pass. The framing, the decisions, and the accountability stay with me.
Where AI assists in my process, and where I lead.
A fast assistant, not a decision maker
I use AI every day to make my process faster and more efficient, and I treat it like a fast, well-read assistant: useful for a first pass, never the final word. I do not rely on it blindly. I check what it gives me, question it, and decide. The thinking that makes a design good, understanding the problem, weighing trade-offs, and taking responsibility for the result, stays with me.
Research, faster and checked
I use AI to summarize interviews, group feedback into themes, and scan large amounts of material. Then I check what it gives me against the sources. Anything that shapes a decision gets traced back to what people actually said.
Research, faster and checked
Sample research
Interview notes
P1
"I check three tools every morning."
P2
"I never know which item is urgent."
P3
"I just want one list."
→
Themes, drafted by AI
Too many places to check
Evidence
P1
P3
Verified
Priority is hard to see
Evidence
P2
Verified
People want a daily summary
Evidence
P3
Needs verification
I trace every theme back to what people actually said before it shapes a decision.
AI drafts the themes. I verify each one against the notes.
Ideate quickly, then choose
When I am exploring, AI helps me generate many options fast: layouts, wording, edge cases. I treat the output as raw material. I pick what is worth pursuing, combine ideas, and throw most of it away.
Many options, then a choice
Ideation
Prompt
Show six layouts for a daily summary screen for team leads.
Option 1
Dropped
Option 2
Kept
Option 3
Dropped
Option 4
Combined
Option 5
Dropped
Option 6
Combined
I treat the output as raw material. Most of it gets thrown away, and the choices are mine.
Most options get dropped. The choices are mine.
Pressure-test ideas before others do
Before I show an idea to a team, I ask AI to poke holes in it: what could confuse a first-time user, what happens when data is missing, where accessibility could break. It surfaces questions I might have missed. I decide which ones matter.
Poke holes before others do
Sample review
What I asked
What AI flagged
My call
What could confuse a first-time user?
The word "flagged" is unclear.
Changed
What happens when there is no data?
No empty state is defined.
Added
Where could accessibility break?
Status relies on color alone.
Fixed
What is missing from this flow?
There is no way to undo.
Not now
It raises questions I might have missed. I decide which ones matter.
AI raises the questions. I make the call.
Speed up the routine work
Documentation, sample content, renaming layers, and repetitive edits take time and need little judgment. AI does the first pass and I review it. The time I get back goes to talking with users, working with engineers, and refining the details.
How much AI helps, task by task
My own view
Task
How much AI helps
Summarizing interviews and feedback
Drafting copy and wording options
Creating realistic sample content
Writing documentation
Renaming and organizing layers
Making design decisions
Always me
Talking with users and engineers
Always me
How much AI helps varies by task, and some tasks stay entirely mine.
The thinking stays human
Design thinking is still mine. I frame the problem, make the decisions, and stand behind them. AI does not know the customer, the constraints, or the context the way the team does. I carry the same belief into the AI features I design: they appear where people need them, and a person approves the result.
What stays with me
Human judgment
AI helps with
First pass
Summaries and first drafts
Alternative layouts and wording
Spotting gaps and edge cases
Repetitive edits and documentation
I own
Final say
Framing the problem
Design decisions and trade-offs
Talking with users and engineers
Quality, taste, and accountability
Same belief
The AI features I design follow it too: placed where people need them, with a person approving the result.
What AI helps with, and what I own.
AI speeds up the work. The thinking stays human.
Next article
Built to scale →
Product design leader. Design functions,
systems, and products for complex SaaS.
Case Studies
Web design system
Native design system
Unified Inbox
AI capability
About
About
How I Work
Résumé
Contact
aln.dohotaru@gmail.com
© 2026 Alina Dohotaru
Visuals are original illustrations and do not depict proprietary product screens.
AI in my work
Using AI wisely
How I use AI to work faster and more efficiently, without relying on it blindly.
Topic
AI in my work
Reading time
3 minutes
Covers
Research, ideation, validation
Visuals
Original illustrations
Where AI fits in my process
My workflow
I lead
Frame the problem
Who is struggling, and why.
→
AI assists
Research
Summaries and themes, checked against sources.
→
AI assists
Ideate
Many options, fast. I choose.
→
I lead
Design and decide
Trade-offs and craft are mine.
→
AI assists
Validate
Pressure-test ideas and find gaps.
→
I lead
Ship and review
Work with engineering and own the result.
AI helps with the first pass. The framing, the decisions, and the accountability stay with me.
Where AI assists in my process, and where I lead.
A fast assistant, not a decision maker
I use AI every day to make my process faster and more efficient, and I treat it like a fast, well-read assistant: useful for a first pass, never the final word. I do not rely on it blindly. I check what it gives me, question it, and decide. The thinking that makes a design good, understanding the problem, weighing trade-offs, and taking responsibility for the result, stays with me.
Research, faster and checked
I use AI to summarize interviews, group feedback into themes, and scan large amounts of material. Then I check what it gives me against the sources. Anything that shapes a decision gets traced back to what people actually said.
Research, faster and checked
Sample research
Interview notes
P1
"I check three tools every morning."
P2
"I never know which item is urgent."
P3
"I just want one list."
→
Themes, drafted by AI
Too many places to check
Evidence
P1
P3
Verified
Priority is hard to see
Evidence
P2
Verified
People want a daily summary
Evidence
P3
Needs verification
I trace every theme back to what people actually said before it shapes a decision.
AI drafts the themes. I verify each one against the notes.
Ideate quickly, then choose
When I am exploring, AI helps me generate many options fast: layouts, wording, edge cases. I treat the output as raw material. I pick what is worth pursuing, combine ideas, and throw most of it away.
Many options, then a choice
Ideation
Prompt
Show six layouts for a daily summary screen for team leads.
Option 1
Dropped
Option 2
Kept
Option 3
Dropped
Option 4
Combined
Option 5
Dropped
Option 6
Combined
I treat the output as raw material. Most of it gets thrown away, and the choices are mine.
Most options get dropped. The choices are mine.
Pressure-test ideas before others do
Before I show an idea to a team, I ask AI to poke holes in it: what could confuse a first-time user, what happens when data is missing, where accessibility could break. It surfaces questions I might have missed. I decide which ones matter.
Poke holes before others do
Sample review
What I asked
What AI flagged
My call
What could confuse a first-time user?
The word "flagged" is unclear.
Changed
What happens when there is no data?
No empty state is defined.
Added
Where could accessibility break?
Status relies on color alone.
Fixed
What is missing from this flow?
There is no way to undo.
Not now
It raises questions I might have missed. I decide which ones matter.
AI raises the questions. I make the call.
Speed up the routine work
Documentation, sample content, renaming layers, and repetitive edits take time and need little judgment. AI does the first pass and I review it. The time I get back goes to talking with users, working with engineers, and refining the details.
How much AI helps, task by task
My own view
Task
How much AI helps
Summarizing interviews and feedback
Drafting copy and wording options
Creating realistic sample content
Writing documentation
Renaming and organizing layers
Making design decisions
Always me
Talking with users and engineers
Always me
How much AI helps varies by task, and some tasks stay entirely mine.
The thinking stays human
Design thinking is still mine. I frame the problem, make the decisions, and stand behind them. AI does not know the customer, the constraints, or the context the way the team does. I carry the same belief into the AI features I design: they appear where people need them, and a person approves the result.
What stays with me
Human judgment
AI helps with
First pass
Summaries and first drafts
Alternative layouts and wording
Spotting gaps and edge cases
Repetitive edits and documentation
I own
Final say
Framing the problem
Design decisions and trade-offs
Talking with users and engineers
Quality, taste, and accountability
Same belief
The AI features I design follow it too: placed where people need them, with a person approving the result.
What AI helps with, and what I own.
AI speeds up the work. The thinking stays human.
Next article
Built to scale →
Product design leader. Design functions,
systems, and products for complex SaaS.
Case Studies
Web design system
Native design system
Unified Inbox
AI capability
About
About
How I Work
Résumé
Contact
aln.dohotaru@gmail.com
© 2026 Alina Dohotaru
Visuals are original illustrations and do not depict proprietary product screens.