Qualtrics' Aug 19 Cut: The Provo XM ML Pool WARN Won't Show

Every sourcer chasing the Qualtrics cut is refreshing Washington's WARN portal and searching "Seattle, WA" on LinkedIn. That is the wrong move: the specialist pool worth mapping - XM Discover ML engineers and survey-platform engineers trained on Qualtrics' human-sentiment dataset - sits in Provo, where no WARN notice will ever land.

What actually happened on August 19, 2026

Qualtrics cut jobs across its dual Seattle and Provo headquarters and international offices on August 19, 2026, three months after closing the $6.75B Press Ganey Forsta acquisition. GeekWire confirmed WARN notices at Qualtrics Tower, 1201 Second Ave., Seattle. The Salt Lake Tribune covered the Utah side on August 22.

The company frames the reduction as duplication between two independently built organizations. That framing hides what the cut actually targets:

• XM Discover, the ML product line launched from Provo and Seattle in February 2022, which analyzes unstructured feedback across contact-center calls, chat, voice transcripts, emails, surveys, and social.
• Qualtrics AI's Socrates platform, launched May 2024 with over 100 AI models running on Amazon SageMaker and Bedrock.
• The survey-platform engineering org that directly overlaps Press Ganey Forsta's stack.

This is the fourth workforce action in under four years. Prior cuts: roughly 270 positions in January 2023 (about 5%) and roughly 780 in October 2023 (about 14%). Under new CEO Jason Maynard (ex-Oracle, took over February 2026), five senior executives were removed in April before this reorganization.

$6.75B

Why WARN filings are a decoy for this one

Washington's WARN law triggers at 50+ cuts per site, which is why Seattle showed up in public filings and Provo did not. Utah's threshold was not tripped on August 19, so Provo cuts of similar order are legally invisible. If you're only chasing WARN lists, you're systematically missing the deeper Provo bench.

Two mechanics make this worse than a normal underreporting problem:

1. Qualtrics has been Provo-native since 2002. It added Seattle later, but the historical center of gravity for XM Discover and survey-platform engineering stayed in Utah County.
2. Provo and Orem engineers routinely self-tag as "Salt Lake City," "Greater Salt Lake," "Utah," or leave region blank. A recruiter filtering by "Seattle, WA" or even "Provo, UT" will under-recall the specialist pool by a wide margin.

The visible signal points at Seattle. The actual releasable specialist bench is in Utah County.

The Refolk index: 283 total, 16 still inside, 267 already out

In Refolk's index, 283 engineers, data scientists, applied scientists, and ML engineers carry Qualtrics as an employer keyword. Only 16 currently list Qualtrics as employer, which sets a floor of 267 people who have already exited before the August 19 cut. That is the reachable ex-Qualtrics engineering pool sourcers should be working right now.

Here is the slice by geography and status:

The 6:1 visible ratio is the trap. Qualtrics has always been more Provo than Seattle for XM Discover and survey-platform work. The profile data skews Seattle because Utah engineers under-tag, not because the roles were there.

What the offshore number tells you

Five profiles across Kraków (2), Manila (1), Andhra Pradesh (1), and Monterrey (1) match the "international offices" language in GeekWire's coverage. If you source globally, Kraków is the most interesting: two profiles, English-fluent, EU time zone overlap with Seattle mornings, and lower relocation friction to Berlin, London, or Dublin than the other three.

The keyword combo that finds the real pool

The precise Boolean signature for an ex-Qualtrics XM ML engineer is "Socrates" OR "Text iQ" OR "XM Discover" combined with SageMaker or Bedrock. Skip "machine learning engineer Provo" as a starting query. It returns thousands of false positives and drowns the signal.

Here is why that combo works:

• Socrates is Qualtrics' internal ML platform name, launched May 2024. It appears almost nowhere else and is a near-guaranteed marker of a current or recent Qualtrics ML engineer.
• Text iQ is the sentiment and topic-model surface layer. Product managers and applied scientists list it on profiles.
• XM Discover is the umbrella product line. Engineers who worked on the ingestion side (calls, chat, voice transcripts, emails, surveys, social) tend to list it explicitly.
• SageMaker + Bedrock together is the AWS-native ML production stack Socrates runs on. This is the transferable skill set Meta, Google, Harmonic, and any voice-of-customer NLP shop will pay for.

This is the exact gap Refolk closes. You describe the person in plain English and get a ranked shortlist that ignores whatever the profile says about region.

The Press Ganey Forsta overlap tells you who actually got cut

When an acquirer says "duplication," the engineers who lose their jobs are almost always on the acquirer's older stack of the redundant product, not the acquired team's. That inverts the default sourcer assumption after a big deal.

Press Ganey Forsta brought a duplicate survey-platform and text-analytics ML stack into the combined company. XM Discover ML (older, Provo-anchored, February 2022 launch) is more exposed than Forsta's ML team, not less. If you're hiring voice-of-customer NLP talent, the August 19 cut is your window into the deeper of the two benches.

Engineers who survived four cuts in four years and three ownership changes in seven are, by revealed preference, the ones who chose to stay. This cut releases high-conviction specialists, not passive quitters.

Where the ex-Qualtrics engineers are already landing

The visible destinations in profile data cluster around three buckets: big-tech ML orgs, adjacent SaaS, and near-namesake companies that create Boolean confusion.

• Big tech ML absorbers: Meta, Google. The most common landings for the applied-scientist tier.
• Adjacent SaaS and enterprise: FloQast (Utah-adjacent, easy Provo transition), Harmonic, HSBC, Fair Trade USA.
• Direct XM competitors to watch: Medallia and SurveyMonkey. Neither has shown up heavily yet in profile data, but both have obvious motive to absorb displaced survey-platform engineers.
• Boolean traps: Qualytics and Quantrics Enterprises. Both surface constantly in "Qualtrics" keyword searches and burn hours of sourcer time. Explicitly exclude them from your Boolean, or accept that a chunk of your list is wrong.

267

If you spend a morning building a list off a company-filter search, a meaningful share of your matches will be Qualytics or Quantrics Enterprises employees who never worked at Qualtrics. When you ask Refolk for "engineers who worked at Qualtrics the experience-management company, not Qualytics or Quantrics," the disambiguation happens before you see the shortlist.

What high-conviction survivors actually look like

The engineers cut on August 19 survived four workforce actions in under four years (January 2023, October 2023, April 2026 exec purge, August 2026) and three ownership changes in seven years - bought, floated, and taken private inside seven years, culminating in Silver Lake's take-private under prior CEO Zig Serafin.

That matters for outreach. Passive quitters left in 2023. Engineers still on the payroll in August 2026 chose to stay through visible instability, which means:

• They believe in the XM Discover / Socrates technical bet and can articulate why. Ask them.
• They have strong internal networks. First-degree referrals from this cohort will surface the Provo names WARN filings never will.
• They will not accept a generic "we saw the news" cold email. Reference Text iQ, Socrates, or a specific XM Discover ingestion surface (voice transcripts, chat, social) in the first line.

The sourcing move for the next 30 days

Three concrete actions, in priority order:

1. Build the Socrates + Text iQ + XM Discover keyword list first. This is the precision layer. Skip title filters entirely; "Software Engineer" and "Applied Scientist" both hit, and title-filtering strips out the survey-platform engineers who don't self-describe as ML.
2. Second pass on geography ignoring region tags. Utah County engineers under-tag. Run the list through university (BYU, Utah, Utah State) and past-employer signals (Ancestry, Domo, Podium, Instructure) rather than trusting the region field.
3. Third pass for landing intent. The ex-Qualtrics cohort that already left (267 profiles) is a leading indicator of where the August 19 cohort will go. FloQast, Harmonic, and Meta are the fast lanes. If you can't out-compensate Meta, position on scope: XM engineers at Qualtrics owned end-to-end product surfaces that big-tech ML orgs rarely give juniors.

Recruiters who work this cut through WARN filings and Seattle filters will land a handful of conversations. The ones who run the Provo-inclusive query on the full 267-person floor plus the August 19 additions will land many more. That is the gap.

FAQ

How many ex-Qualtrics engineers are actually reachable right now?

Refolk's index shows 283 engineers, data scientists, applied scientists, and ML engineers with Qualtrics as an employer keyword, of whom 16 currently list Qualtrics as employer. That leaves a floor of 267 already out and reachable, before the August 19 cut adds more. The visible Seattle-tagged slice is only 6 people, which is the artifact most sourcers mistake for the whole pool.

Why is the Provo pool invisible in WARN filings?

Washington state's WARN law triggers at 50+ cuts per site, which is why the Seattle Qualtrics Tower cut at 1201 Second Ave. is public. Utah's threshold was not tripped on August 19, so Provo cuts of similar magnitude never enter a public database. Sourcers who work exclusively from WARN lists systematically miss the Utah County specialist bench.

What keyword combo identifies an ex-Qualtrics XM ML engineer with high precision?

"Socrates" OR "Text iQ" OR "XM Discover" combined with SageMaker or Bedrock. Socrates is Qualtrics' internal ML platform (launched May 2024, over 100 AI models), Text iQ is the sentiment layer, and XM Discover is the umbrella product for unstructured feedback analysis. This combo returns dozens of true positives instead of the thousands of false positives you get from "machine learning engineer Provo."

Which competitors are most likely to absorb the displaced engineers?

Meta and Google are the top visible landings in profile data for the ex-Qualtrics cohort that left before August 19. Direct XM competitors Medallia and SurveyMonkey have clear motive to absorb displaced survey-platform engineers next. Watch FloQast for Utah-County-anchored talent that wants to stay local, and exclude Qualytics and Quantrics Enterprises from your Boolean before they eat a third of your review time.