Emily Goligoski, qualitative researcher and audience strategist

Emily Goligoski has spent over a decade studying how people actually use products, services, and institutions.

Guest: Emily Goligoski, qualitative researcher and audience strategist Hosts: Donna Scarola and Stela Lupushor

Emily has spent over a decade studying how people actually use products, services, and institutions. She led audience research at The Atlantic, co-founded the Membership Puzzle Project at NYU, was the first head of research at Charter, and has done research work for The New York Times, Mozilla, and the BBC. She teaches the business of media at Columbia Journalism School. For Charter, she co-authored two of the more rigorous pieces of work on AI in the workplace: the 2023 playbook (with IRC4HR) on using AI in ways that enhance worker dignity and inclusion, and a 2024 follow-up on what companies and workers are getting right.

We get into why most employee personas should probably be retired, how to get honest answers out of research participants (hint: chocolate croissants), and why the assumption that AI wisdom lives at the senior levels is the thing leaders will regret most.

What we cover

  • (02:55) From newsrooms to the Charter AI playbook. Nearly everyone works at some point, which makes the workplace the place to study people across contexts and at scale.
  • (05:26) How Charter "ruined work" for Emily. Time blocking, stacked meetings, no internal meetings on Fridays. If the only time your employees can do their hardest thinking is after the kids are in bed, that is a design failure, not a productivity problem.
  • (07:39) "We ran a survey and built three employee personas." Emily's first reaction: relief that you asked employees anything at all. Then the tweaks: is a survey even the right method, your insights have a shorter shelf life than you think, and what are you showing people about what you did with their input.
  • (11:50) The no surprises policy. Never let someone be surprised by the outcome of something they had a hand in.
  • (12:14) Kick out your personas. Is HR ten years behind product design, again? Personas as shorthand: useful when grounded in real research and kept fresh, risky when they bucket people in ways that turn dehumanizing or quietly discriminatory. Emily's ask: "personas plus."
  • (15:47) The lightest lift: user needs and jobs to be done. Unlike personas, user needs by necessity change over an employee's lifespan. The bigger shift: centering the person, not the organization.
  • (17:37) The chocolate croissant script. How Emily opens every focus group so people tell her what actually happened this morning, not what sounds impressive. Skepticism as the most useful service a research participant can provide.
  • (19:43) Whose voice is missing. Every self-reported study is bounded by who has the time, inclination, and technology to participate. Also: most of human culture is unwritten, and AI privileges English.
  • (21:26) The worker dignity research, revisited. Authorship, and helping people create work they are happy to have their name on.
  • (24:21) Why women, workers of color, and workers 55+ adopt more slowly. AI optimism ("get on board or get left behind") deprioritizes real concerns: privacy, the burden of caregiving, environmental impact. The colleague who is furthest ahead on AI probably has the fewest obligations on their time.
  • (27:01) Cloudflare's AI-washed layoffs and Bolt deleting HR. Stela's worry about what comes next when the people making these calls carry the lowest caregiving burden.
  • (28:29) "AI is not the answer. It is one answer." One truth versus many truths, tone that tells you what you want to hear, and how those two things compound into real risk.
  • (31:27) Measuring AI impact beyond time saved. Diary studies, longitudinal research, a week in the life instead of 15 minutes. The question Emily always asks: what feels like unnecessary effort that you would want transformed in the next one to three years?
  • (34:54) Questions worth stealing. Stela's favorite icebreaker (what is your most embarrassing prompt?), Donna's manager questions applied to AI (how can I tell when you are frustrated, in flow, overwhelmed?), and Emily's addition: when did AI last do a great job by you?
  • (38:18) Executives failing out loud. The polished demo has its place, but nothing teaches like a leader hitting an error in public and crowdsourcing what to do next.
  • (40:02) Vibe coding at Columbia. Emily's graduate students finally stopped obsessing over visual fidelity and spent their time on the business case. Her classroom rule: use AI, disclose everything, and if you opt out, tell me why.
  • (43:12) Token monitoring and quota games. When usage becomes a metric, people play the metric.
  • (46:01) What leaders will regret three years from now. Believing that AI workplace intelligence comes from the senior levels. The people who will teach your workforce are scattered across the org chart.
  • (48:26) Stela's summary. Start with user needs, build personas plus, find who is excluded, design for the caregiving burden, talk to humans, and have the chocolate croissant. Disclose the croissant.

Mentioned in this episode

Get the latest episodes directly in your inbox