Research & validation — Field Notes

NIELSEN NORMAN GROUP

LDLR DESIGN FIELD NOTES

Curation beats prompting

NN/g’s power-user study lands one shift: context curation, not prompt craft, is where the leverage moved.

BY TANNER KOHLER / SEP 18, 2026

The 3 Roles of Context for AI Agents

Curating context now matters more than writing the perfect prompt.

Why it matters

— Every power user studied had built a context library before agents got good

— Three roles: global rarely changes, local tracks the task, ambient streams in raw

— Handling follows role: curate global, let the agent write local, let it sift ambient

— A deliberately lazy prompt worked fine because the MD file did the heavy lifting

Read on Nielsen Norman Group →

Research & validation Systems & primitives

UX COLLECTIVE

LDLR DESIGN FIELD NOTES

The second collision

Neeman reads Nader into AI: the model's error is the first crash, the interface's handling of it is yours.

SEP 18, 2026

Altman and Amodei's 'unsafe at any prompt' moment for AI

Capability has a leaderboard. Harm has a mailing list.

Why it matters

— 'Prompt better' is Detroit's nut behind the wheel defense, reframed for AI

— Stanford's 2026 Index: hallucination rates from 22 to 94 percent across 26 models

— The metric to steal: wrong-answer survival rate, measured with seeded failures

— Pad the interior: citations, visible uncertainty, cheap undo before Article 86 forces it

Read on UX Collective →

Critique & decision-making Research & validation

UX COLLECTIVE

LDLR DESIGN FIELD NOTES

The confidant, logged

Brêtas follows surveillance into the thought before the thought: drafts confided to AI, logged, scored, discoverable.

SEP 10, 2026

Your first draft now has a log

We invented a confidant the law never recognized as a confidant.

Why it matters

— Trust No Bot, 2024: disclosures surface in translation requests and code edits

— ActivTrak splits one dataset: self-care screen for you, named ranking for your manager

— Ravid’s 94-sample review finds no performance gain from monitoring, transparency helps attitudes

— February’s Heppner ruling: Claude conversations went to a third party, unprivileged

— Judge collection by its worst plausible use, and who walks data back

Read on UX Collective →

Research & validation Critique & decision-making

NIELSEN NORMAN GROUP

LDLR DESIGN FIELD NOTES

Three keys, one lock

Fessenden names the challenges and five moves, simplicity through co-creation, that raise the floor for every user.

BY THERESE FESSENDEN / SEP 7, 2026

Neurodivergence: What is it and Why Does it Matter

Neuroinclusive design helps all users: simplicity, clarity, control, support, and co-creation.

Why it matters

— Executive function, communication, and sensory processing name the challenge areas

— Five moves: simplicity, clarity, control, support, and co-creation with users

— Co-creation puts neurodivergent users in the process, not just the tests

— Six minutes, sized for the next team accessibility review

Watch on Nielsen Norman Group →

Research & validation Craft & detail

NIELSEN NORMAN GROUP

LDLR DESIGN FIELD NOTES

Plausible is no one

Krause draws the AI line in empathy mapping: a sticky note you can’t trace to a real person is an assumption in costume.

BY RACHEL KRAUSE / AUG 28, 2026

AI Can't Replace Real Research in Empathy Mapping

What it produces could describe a frustrated user of almost any product, which means it describes no one in particular.

Why it matters

— The pairs tell it: “I wish the app would ask before swapping items” vs the oat milk swapped for whole milk and charged before the notification landed

— Real quote, banking: “I always transfer $1 first to make sure I typed the account number right”

— Legitimate: clustering, cleanup, and summarizing data you collected. Compromising: generating any of it

— The trace test: can this sticky note be walked back to actual research evidence?

— Uncomfortable answers mean the map needs more research, not more prompting

Read on Nielsen Norman Group →

Research & validation Critique & decision-making

NIELSEN NORMAN GROUP

LDLR DESIGN FIELD NOTES

Sweep once, teach twice

Kaley and Budiu name the custodial era: production got cheaper than evaluation, so UX triages what already exists.

BY ANNA KALEY AND RALUCA BUDIU / AUG 28, 2026

The Custodial Era of UX: Cleaning Up After AI

It can now take less time to create an experience than to decide whether that experience is useful, usable, trustworthy, or coherent.

Why it matters

— The custodial pattern, five steps: new idea, rapid production, lagging evaluation, user problems, UX cleanup

— Triage before polishing: evidence of need, mental-model fit, integration cost, opportunity cost

— The outcome isn’t always “redesign it” — keep, simplify, integrate, postpone, or remove

— Match evidence to risk: quick tests for low-stakes features, rigor reserved for high-stakes workflows

— Feed cleanup lessons back into generation: UX-context files, design systems, content standards

Read on Nielsen Norman Group →

Critique & decision-making Cross-functional & process Research & validation

LINEAR

LDLR DESIGN FIELD NOTES

Hours up, not away

Qi’s report from 47,900 workspaces: AI writes half of everything in Linear, PRs are up 111%, and no time came back.

BY TIM QI / AUG 17, 2026

AI usage patterns in software teams

As far as we can observe, teams are working more, not less, suggesting AI has a Jevons paradox quality beyond token consumption.

Why it matters

— AI now authors just under half of everything created in Linear

— Product adoption climbed fastest, 12% to 34% in six months

— CEOs at 201+ companies went 9% to 36% — the report's largest jump

— Coding-agent teams tripled weekly PRs; teams without went 8 to 10

— Nothing shrank to make room — AI landed on top of existing work

Read on Linear →

Cross-functional & process Research & validation

NIELSEN NORMAN GROUP

LDLR DESIGN FIELD NOTES

One output isn’t proof

Budiu retires the one-good-output proof: representative inputs, repeated runs, and confidence intervals instead.

BY RALUCA BUDIU / AUG 14, 2026

One AI Output Is an Example, Not an Evaluation

A single good output is like a participant who completes a task: encouraging, but not an evaluation.

Why it matters

— Ten representative inputs, five runs each: fifty answers before judging

— Two systems can share an 80% score and fail in opposite ways

— Test-input and run-to-run variability answer different questions; neither substitutes

— For nondeterministic systems, a passing QA test is a sample, not proof

— If runs are too expensive to repeat, the evidence is insufficient

Read on Nielsen Norman Group →

Research & validation Critique & decision-making

MICROSOFT DESIGN

LDLR DESIGN FIELD NOTES

Trust sets during setup

Dhaka and Sikdar watched sales reps meet an agent, then distilled five principles: aha before configuration.

BY POOJA DHAKA AND GARIMA SIKDAR / AUG 13, 2026

Designing the first five minutes

The fix isn't eliminating the setup. It's reordering it.

Why it matters

— Generate a baseline draft from minimal input, then show sources improving it

— Design for partial completion and save-and-resume; show what's done and what's next

— Absent plain-language progress cues, participants just issued repeated prompts

— Ask for context only when it materially changes the result, then retain it

— Users project ChatGPT priors onto narrow agents; show example tasks early

Read on Microsoft Design →

Research & validation Craft & detail

NIELSEN NORMAN GROUP

LDLR DESIGN FIELD NOTES

Sense is a closed loop

Kohler rebuts empathy-plus-creativity: sense is pattern recognition built by shipping and staying to see the results.

JUL 24, 2026

A Concrete Definition of "Product Sense" (and How to Build It)

AI may rob you of the chance to develop product sense if you aren't careful.

Why it matters

— Against Walter's empathy-plus-creativity: sense is pattern recognition from completed loops

— Fire chiefs, NICU nurses, chess grandmasters: intuition built by seeing outcomes

— Wicked situations feel like pattern matches and reinforce the wrong lesson

— Stay long enough to see results; document your logic before looking at them

— Half-loop builders feel experienced but never learn whether the building collapsed

Read on Nielsen Norman Group →

Critique & decision-making Research & validation

NIELSEN NORMAN GROUP

LDLR DESIGN FIELD NOTES

Context, not documents

Every prompt is now a design decision. Alicea’s answer: stop persuading humans and curate context that steers models.

JUL 24, 2026

UX-Context Design: Using UX Knowledge to Inform AI-Generated Design

Its success is measured by whether AI output improves, not by whether stakeholders are convinced by it.

Why it matters

— When PMs and engineers prompt, every generation is a design decision

— Personas persuade humans; models need the underlying reasoning instead

— Proposes UX.md beside DESIGN.md: research synthesis, glossary, user and world models

— Research findings become generation constraints, not insights in a report

— Continuously curated beside the code; there is no handoff moment

Read on Nielsen Norman Group →

Systems & primitives Research & validation Cross-functional & process

THE MERHOLZ AGENDA

LDLR DESIGN FIELD NOTES

Sentiment is an outcome

Twenty exec interviews and a 458-response pulse check converge: how your org behaves predicts how design feels about AI.

JUL 23, 2026

The sentiment around AI + Design is... complicated

The bifurcation isn't out there, it's within us.

Why it matters

— Twenty VP-plus interviews, Config crowd-work, and a 458-response pulse check behind this

— Roller Coaster is the plurality sentiment at 26%; the split runs within people

— The less AI-mature the org, the more negative its people, security included

— Directors sit with managers at 33% negative; VPs step-change to 34% positive

— His prescription is grace at scale, created by whoever holds the agency

Read on The Merholz Agenda →

Research & validation Hiring & team Cross-functional & process

LENNY'S NEWSLETTER

LDLR DESIGN FIELD NOTES

The fault line: identity

The second tech-worker sentiment survey finds an industry splitting on AI identity, researchers at the anxious edge.

JUL 7, 2026

How tech workers are feeling in 2026: a workforce splitting in two

Every gain becomes the new baseline, and the people expected to hit it are running out of room to breathe.

Why it matters

— AI identity beats role, level, and company size combined as an optimism predictor

— Significant burnout jumped from 44.7% to 55.7% in a single year

— 53% would steer a newcomer away; the field's NPS lands at minus 39

— Designers report the most overwhelm, the worst-rated managers, and the biggest comp squeeze

— Top fear isn't replacement at 22%; it's more output for the same pay at 51%

Read on Lenny's Newsletter →

Research & validation Hiring & team

BUILDING NUBANK

LDLR DESIGN FIELD NOTES

Simplicity is strategy

Five lessons from 127 million customers, where hiding complexity is a deliberate architectural and cultural decision.

BY MARISTELA CALAZANS / FEB 16, 2026

What changes when your product reaches 127 million people

Simplicity at scale is not a design preference. It's a strategic decision.

Why it matters

— Quarterly Customer Connection visits put teams in unstable-internet, older-device reality

— Accumulated friction, not catastrophic failure, is the real loyalty killer

— Caixinhas cut a dozen fund options to a two-step habit-building flow

— Global core, local nuance shipped Colombia's Bre-B six months faster

— Intelligent friction: slowing users down protects life savings on unusual transactions

Read on Building Nubank →

Cross-functional & process Research & validation Craft & detail

UX COLLECTIVE

LDLR DESIGN FIELD NOTES

A skeptic’s AI playbook

A research leader of eight argues for measured adoption: pressure-tested tools, tracked outputs, guardrails first.

APR 30, 2026

Thoughtful AI implementation for UXR leaders

I'm arguing for a measured approach to their usage.

Why it matters

— Pressure-test NotebookLM with likely stakeholder questions before opening access

— Capture before and after results when tuning prompts with ops

— Track who queries research repositories and how outputs shape decisions

— Half of Americans are more concerned than excited about AI

Read on UX Collective →

Research & validation Critique & decision-making

HARVARD BUSINESS REVIEW

LDLR DESIGN FIELD NOTES

The new work confidant

Employees turn to AI for advice, support, and friendship. The researchers say leaders must protect human connection.

BY CONSTANCE NOONAN HADLEY AND SARAH L. WRIGHT / MAY 1, 2026

Employees Are Relying on AI for Personal Support. That's Risky.

Research shows how AI can weaken human connection at work.

Why it matters

— More than half of 1,545 surveyed knowledge workers feel lonely at work

— Workers seek career advice, emotional support, and friendship from AI

— AI conversation can quietly substitute for, then erode, human connection

— The authors call on leaders to intervene before reliance hardens

Read on Harvard Business Review →

Hiring & team Research & validation

MEDIUM

LDLR DESIGN FIELD NOTES

AI ate the PM talk

A multi-agent read of 638 Lenny pieces, June 2019 to March 2026, charting what production loses and judgment keeps.

MAR 18, 2026

What 638 Practitioner Voices Reveal About Product Manager's AI Transformation

Speed raises the bar for where rigour is applied, rather than lowering it.

Why it matters

— AI-tagged content jumped from 4% of output to 67% in three years.

— Built transparently: five Claude agents reading the corpus, the method shown in full.

— Jenny Wen argues the design process designers were taught is effectively over.

— Products become organisms that learn, pushing pricing toward outcomes, not seats.

— What endures: conviction, taste, influence, and a focus on outcomes.

Read on Medium →

Research & validation Prototyping & exploration Cross-functional & process

DESIGN SYSTEMS COLLECTIVE

LDLR DESIGN FIELD NOTES

Readable by machines

Indeed’s build story: structured metadata over prose, and JSON five times cheaper than Markdown for the same query.

MAY 14, 2026

Fully Machine-Readable Design Systems

Why it matters

— AI reads your design system differently; structured metadata beats prose documentation.

— Foundations should be always-on; components served on demand through an MCP.

— At Indeed, JSON cost five times less than Markdown for the same query.

— Progressive disclosure maps what AI needs without flooding its context window.

— The system you wrote for designers won't work for agents without a rewrite.

Read on Design Systems Collective →

Systems & primitives Craft & detail Research & validation

UX COLLECTIVE

LDLR DESIGN FIELD NOTES

AI holds the context

Shakir’s workflow: the AI holds context across Linear, Slack, and research while the designer does the thinking.

MAY 15, 2026

How I use AI to partner on design problems

Why it matters

— The AI holds context across Linear, Slack, and research; you do the thinking.

— Code prototypes move feedback from visual details to actual UX problems.

— Bad ideas surface faster when they're running, not sitting in Figma.

— Decision tracking lives in a folder per round, not buried in comment threads.

— The longer the project runs, the more useful the accumulated context becomes.

Read on UX Collective →

Prototyping & exploration Cross-functional & process Research & validation

MICROSOFT DESIGN

LDLR DESIGN FIELD NOTES

The missing seat

Frontier AI teams optimize what’s measurable, so Microsoft’s researchers want a seat where the metrics get chosen.

BY MARTIN LINDE DANTY, RABEEZA A., ANNE KLÆRKE-OLESEN, MADHUMAY SINHA, MICHAEL O'SULLIVAN / APR 28, 2026

The missing seat at the frontier team table

When evaluation is driven primarily by available data, teams risk optimizing what is measurable rather than what actually matters.

Why it matters

— User research belongs in evaluation, not adjacent to it.

— Session-length looked strong. Task-completion cut repeat queries by a third.

— Sales-order workflow assumed isolation. Research surfaced cascading approvals.

— Concept tests caught a workflow misread that would have cost weeks.

Read on Microsoft Design →

Research & validation Cross-functional & process Critique & decision-making

MICROSOFT WORKLAB

LDLR DESIGN FIELD NOTES

Workers ready, orgs not

Microsoft’s 2026 Work Trend Index, 20,000 workers across ten countries: only 26 percent see leadership aligned on AI.

BY MICROSOFT WORK TREND INDEX TEAM / MAY 5, 2026

Agents, human agency, and the opportunity for every organization

Workers are ready. Their organizations aren't.

Why it matters

— One in five workers sit in the Frontier zone.

— One in ten are ready individuals blocked by the organization.

— Only twenty-six percent say leadership is consistently aligned on AI.

— Sixty-five percent fear falling behind. Forty-five percent stay in old work.

Read on Microsoft WorkLab →

Cross-functional & process Research & validation Storytelling & narrative

NUBANK DESIGN

LDLR DESIGN FIELD NOTES

Darlings, killed by data

Nubank A/B tested whether to scale or shut an internal HR product. The data was unambiguous; the team’s reaction wasn’t.

BY NUBANK EDITORIAL / AUG 27, 2025

A/B Testing for Decision of scaling or decommissioning an Human Resources product

Increased usage didn't always translate to feeling recognized — a critical insight missed without experimental

Why it matters

— A/B testing applied to decommissioning, not just feature launches.

— Internal tools deserve the same rigor as customer-facing products.

— Usage below threshold: kill it. Above: scale it.

— Hardest part isn't the data, it's letting the team accept it.

Read on Nubank Design →

Research & validation Cross-functional & process Critique & decision-making