SXSW 2027 · PanelPicker
Design track · 60-min panel

We Summarized the Humanity Right Out of It

AI doesn't break user research. It automates a flaw that was already there, and it can flatten people at a scale beyond what teams can see.

▲ Vote for this panel See it happen ↓ 30 seconds. one sticky note.

AI & research•Inclusive design•Synthetic personas•Not an anti-AI talk

Watch a person become a theme.

Hit synthesize to see a story morph into a common denominator. Designing for averages is the risk we run when we theme clusters in an affinity map, build synthetic personas, or summarize with Gen AI. Notice what survives.

One of these quotes is real. The rest are composites, drawn for illustration.

Story 1 / 5
Discarded
Kept

0 human details discarded across 0 stories.

Now imagine ten thousand of these, averaged overnight, shipped as “the insight.”

We didn't fix the flaw. We automated it.

MOVE 01

The flaw came first.

Long before AI, we watered human stories down to time, trust, peace of mind. The details we lost first belonged to the people we heard from least.

MOVE 02

AI learned from us.

Research tools trained on that watered-down practice. A summary captures what's common. Insight lives in tensions and surprises. The blind spot now repeats with speed and at scale.

MOVE 03

Augmentation becomes replacement.

It starts as help. It ends as a replacement. The budget and the real customer relationships quietly fade. Once they're gone, they're hard to rebuild.

The average is the one customer who doesn't exist.

Every point is a person. Convergence is trained to find the crowd in the middle. It lets go of everyone the crowd isn't. That's not a bug in the model. It's the definition of a summary.

the signal was out there ↗

Meet your panelists.

We met in 2019 and worked on the same innovation pod at Nationwide with a shared goal to find unmet customer needs through ethnographic field work. It was some of the most exciting and human work of our careers. We believe the future is both AI and human, not one or the other. We came to talk about how to keep automation from flattening people's stories, and how to still work humanly with its help.

Even our stories can get lost in synthesis. Give it a try or come up with your own.

Ally Hwang

Product & Design Leader

Ally is a design and research leader with 14 years across enterprise software, insurance, blockchain, education, and mental health.

She's currently UX Director at Nationwide, where she leads enterprise product design. She believes empowering people is the foundation of great teamwork and outcomes. Off the clock, she's usually tending her garden, playing Neopets, deep in a graphic novel, or watching pro-wrestling. She has 4 cats that drive her nuts but she wouldn't have it any other way.

Enterprise design leader

Shara Coutourier

Product & Design Leader

Shara is a product and design leader working at the intersection of AI, enterprise platforms, and the practical realities of how professionals get work done.

She currently leads product and design for a consulting and analytics organization inside a global payments company, where her focus is on using AI responsibly to augment expert work rather than replace it. Before that, she spent several years as a researcher and product strategist at Microsoft, studying AI adoption, productivity, and workflow. Her earlier career spans government consulting and zero-to-one startups. Outside of work, she is usually with her niece Charlotte or her dog Raekwon, traveling, or teaching herself to DJ.

AI product leader

Elizabeth Leab

Product Leader
Moderator

Elizabeth is a product leader with 14 years experience across insurance, fintech, retail, and global commerce, with a particular talent for discovery: finding the business opportunities that drive customer value and solve actual problems.

She's currently a Senior Product Manager at Amazon, where she owns a global transportation platform spanning 16 marketplaces. Her career has been built on customer-facing products, including leading CX development for Amazon Go stores through Just Walk Out technology and leading the tech team for Self Checkout at The Home Depot. Her favorite way to build is through balanced teams across business, data, engineering, and UX. She'll argue that research and design are crafts that deserve real investment, not tasks a product manager absorbs. Outside of work, she plays field hockey with the Seattle Women's and Co-Ed Field Hockey Club, spends time with her people and her cats, and is always chasing her next best meal and cocktail.

Commerce product leader

Ally & Shara last shared a stage in 2021 with “When Majority Shouldn't Rule.” This is the sequel, five years and one AI boom later.

Their 2021 talk showed how recruiting tools, affinity maps, and personas quietly erase the people on the margins. The takeaway was simple. When research chases the majority, the people who most need to be heard go unheard. Now AI runs every one of those steps, so the same exclusion happens automatically, at a scale no team can see.

What is your research pipeline built to miss?

This isn't an anti-AI talk. It's a talk about knowing what your tools are built to miss, and protecting the part of research no tool can do. Help us bring it to Austin.

▲ Vote on PanelPicker

↑ drop our PanelPicker link in here before Aug 5