Christophe Jammet

The design principle is to understand users’ mental models and design for how people actually work, not how teams think they should. They co-authored a Gather white paper with Rudy Chang, a long-time collaborator on the argument, and thanked Chiara Bartlett for design and Nick Douglas for editorial guidance.

Privacy automation scaled because it preserved human expertise at critical decision points. Platform migrations succeeded by mapping mental models before changing workflows.

Jammet frames the argument as a discipline teams already know from agile work but set aside during the AI gold rush: restore user research in AI projects rather than invent a new method.

Read original

In a Gather Signal episode, they describe a practical workshop: get everyone in a room, name the real pain points, and give people a low-cost way to solve them through vibe coding. The episode also features Jason Oberholtzer, Andrea Bezark Stein, and Rudy Chang.

The low-cost solution is vibe coding, after the group identifies real pain points and before it commits budget to a direction.

The argument sharpens the usual case for research: it gives teams evidence to challenge a contested executive idea.

Read original

Their Eames chair analogy: a beautiful chair left on the sidewalk is not a strategy.

• They contrast money spent on Super Bowl spots and airport posters—which, they say, do little for most B2B advertisers—with reluctance to spend a fraction of that to put the right case study in front of forty people who could actually buy.
• Targeted paid amplification exists, works, and is underutilized; the post itself was deliberately targeted.

A billboard with vague platitudes gets six figures, while a much smaller spend to reach forty possible buyers with the right case study gets refused.

Jammet says paid amplification takes skill and that the post was deliberately targeted. The claim is that targeted amplification is underused, not that it is a mechanical fix.

Read original

The stack handles transcripts with speaker attribution, titles and descriptions, chapters, designed and captioned Shorts, thumbnails, scheduling and uploads, launch posts, paid media plans, and performance reporting.

• A human editor still handles the full episode edit.
• They offer the approach to anyone with a podcast, webinar series, or pile of recorded conversations who wants to atomize and scale them, and thank host Jason Oberholtzer, AI and design systems collaborator Hilary Davis, and team members Matthew K., Kirsten C., and Michael Terwindt.

The stack produces clips, captions, thumbnails, launch posts, paid media plans, and performance reports, while a human editor handles the full episode edit.

The main asset stays under human editorial judgment; AI handles repeatable production and distribution outputs.

Read original

Start here for the distribution argument and a tested alternative.

Jammet writes that B2B teams often create strong, persona-specific content and then will not spend a fraction of the production budget to make sure that persona sees it.

• The post uses an Eames chair on a sidewalk as an analogy: a beautiful object left where its likely buyer might happen to walk by is not a strategy.
• Jammet contrasts that with money that does appear for a Super Bowl spot or airport poster, formats Jammet says do little for most B2B advertisers.
• The specific example is spending six figures on a billboard with vague platitudes while balking at a much smaller sum to put the right case study in front of forty possible buyers.
• Jammet says targeted paid amplification already works and is criminally underutilized, and notes that if the reader saw the post, it was probably because Jammet targeted them on purpose, not because of the algorithm.

For brands planning a campaign

For a client with strong content but weak distribution, budget paid amplification to reach a defined buying committee rather than relying on broad brand placements. Test whether the message reaches the people who can buy.

Start here for Jammet’s clearest case for human-centered design.

This post announces a Gather white paper on human-centered design for AI and automation.

• Jammet argues that teams on agile projects already know to listen to users before building and to design for how people actually work rather than how teams think they should.
• During the AI gold rush, Jammet says that discipline was tossed aside and enterprise pilots collapsed not from technical limits but because nobody did the work to understand how humans would interact with them.
• The post names two implementation patterns: privacy automation that scaled because it preserved human expertise at critical decision points, and platform migrations that succeeded by mapping mental models before migrating workflows.
• Jammet credits Rudy Chang as a long-time collaborator on the argument, Chiara Bartlett for design, and Nick Douglas for editorial guidance.

For brands planning a campaign

For an AI-enablement or automation engagement, budget for user research and mental-model mapping before the build. Position human-centered design as the mechanism that determines whether AI initiatives deliver ROI or become write-offs.

Start here for Jammet’s practical case to budget holders.

Jammet argues that UX research is usually the first thing cut when corporate marketing budgets tighten—and that this is a mistake for a less obvious reason: research gives teams a way to challenge a bad idea from the top by saying “the research says otherwise” instead of “I disagree.”

• A practical move from a Gather Signal episode: get everyone in a room, name the real pain points, and give people a low-cost way to solve them, with vibe coding as the low-cost option.
• Jammet asks what executive idea UX research has saved people from and points to takes from Jason Oberholtzer, Andrea Bezark Stein, and Rudy Chang in the full episode.

For brands planning a campaign

For a client whose stakeholders are divided on an AI or content direction, propose a short research workshop before committing budget. Position research as decision protection and evidence, not just persona development.

Start here for the actual automated outputs.

Jammet describes building an automated post-production stack for The Gather Signal, which crossed 1,600 YouTube subscribers after three episodes.

• The stack handles transcripts with speaker attribution; titles and descriptions; chapters; designed and captioned Shorts; thumbnails; scheduling and uploads; launch posts; paid media plans; and performance reporting.
• The only part it doesn’t touch is the full-episode edit, which still needs a human editor, Ash.
• Jammet offers the approach to anyone with a podcast, webinar series, or pile of recorded conversations they want to atomize and scale.
• They thank host Jason Oberholtzer, AI and design systems collaborator Hilary Davis, and team members Matthew K., Kirsten C., and Michael Terwindt.

For brands planning a campaign

If a client has an archive of webinars or podcast episodes, propose a repeatable AI-assisted repurposing pipeline for distribution assets and paid-media plans, with a human editor handling the main episode.

In a white paper, they point to privacy automation that scaled by preserving human expertise at critical decision points and platform migrations that succeeded by mapping mental models before moving workflows.

White paper with implementation examples; mental-model mapping before migration; user research before solution design.

For brands planning a campaign

For an enterprise AI or automation client, use research to identify adoption barriers and decision points before building campaign messaging. The practical angle is not simply that AI is used, but that it is designed around how teams already make decisions.

Examples: Source 1

A team can say, “the research says otherwise,” instead of “I disagree.” On a Gather Signal episode, they make the case for getting everyone in a room, naming real pain points, and giving people a low-cost way to solve them through vibe coding.

Group workshop; naming real pain points; low-cost prototyping or vibe coding to validate a direction.

For brands planning a campaign

When a client has a contested idea, propose a small research sprint or workshop to surface real pain points and produce actionable evidence. Low-cost prototypes can test a direction before a large campaign commitment.

Examples: Source 1

The Eames chair problem is their analogy: a beautiful chair on the sidewalk is not a strategy.

• They contrast broad spends such as Super Bowl spots and airport posters with the reluctance to put a case study in front of forty people who could actually buy.
• Targeted paid amplification exists, works, and is underutilized; the post itself was targeted on purpose.

Targeted paid social and content amplification; narrow audience targeting; cost comparison against broad awareness advertising.

Christophe says paid amplification takes both art and science, and that the post was deliberately targeted. The claim is about underuse, not a purely mechanical fix.

For brands planning a campaign

Build paid distribution into content plans rather than treating it as an afterthought. For a B2B client, shift budget from broad awareness formats to precise amplification of case studies and other high-value assets aimed at named buyer segments.

Examples: Source 1

Christophe directed their AI enablement work at their own content and built an automated post-production stack: transcripts with speaker attribution, titles and descriptions, chapters, Shorts designed, cut, and captioned, thumbnails, scheduling and uploads, launch posts, paid media plans, and performance reporting. The full episode edit still requires a human editor.

Automated post-production stack; repurposing long recordings into clips, captions, thumbnails, launch posts, paid media plans, and performance reports.

The full episode edit is not automated; Christophe says it still needs an editor’s human touch.

For brands planning a campaign

For clients with podcasts, webinars, or recorded conversations, plan to atomize one recording into multiple assets and distribution routes. Use AI for repeatable production work, but keep human editorial judgment in the full edit and final quality check.

Examples: Source 1

They also say they built an automated post-production stack for episodes and directed their AI enablement work at “our own content.” For a partnership, ask whether their relationship to Gather is employment, ownership, or client work.

View example

Both use short, first-person LinkedIn posts to unpack a point from their working lives — a question a client asked, or a role that turned out differently from the advert. If Jammet's case for putting research before AI development interests you, Mas González shows what that looks like from the researcher's desk.

Mas González writes from the practitioner's side: whether a job actually uses qualitative research skills, how to interview for fit, and which parts of drafting discussion guides an AI assistant may touch. Jammet writes from the strategy side, exploring how research shapes what a company builds and where the money goes.

Read original · Original 2

Jammet describes the missed research that sinks an AI pilot; Van Cuyk, the rushed hire that costs a small business time, money and morale. Readers who like Jammet's practical checklists will find a similar format here.

Van Cuyk covers small-business hiring and team structure: when to bring in a specialist, how to define the role, and what onboarding needs. Follow Van Cuyk for staffing decisions, not user research, media planning or campaign design.

Read original · Original 2

Both write from direct client work rather than general commentary on marketing.

Keaney-Bolland focuses on positioning and differentiation for cybersecurity startups, using comparison exercises to find where a company can lead.

• Jammet focuses on research-led decisions and paid distribution once the content exists.
• Choose Keaney-Bolland when a brand sounds like everyone else; choose Jammet when no one sees the work.

Read original · Original 2

The posts describe a hands-on practice: drafting, editing and re-editing discussion guides and survey questions from scratch, then using AI as a second editor to test whether questions are more concise, better ordered or missing anything.

• The researcher reviews each suggested question to ensure it stays open-ended and non-leading.
• The researcher handles the data digging, synthesis, final report and delivery.
• This makes Diego a better pick when a client has a decision to de-risk before anyone writes a campaign message—for example, when a product team cannot explain why buyers behave as they do.
• A second post addresses the agency’s own hiring challenge: a role advertised as a UX researcher turned out to involve spreadsheets, survey data and web analytics.
• The post lists interview questions that would have surfaced the mismatch, useful for an agency staffing a research function and looking to avoid the same trap.

The posts show no paid-media planning, budget-setting or distribution work.

• If the agency needs to put an existing white paper or case study in front of a named buying committee, Jammet’s work—not these posts—supports the argument for targeted amplification that is already funded.
• Diego Mas González is the stronger choice to produce and defend the evidence; Jammet is the stronger choice to get it seen.

Examples: Source 1 · Source 2 · Source 3

The posts cover when a small-business owner should stop doing everything and start delegating; how to define the role actually needed and what success in that seat looks like; why hiring out of desperation or from a vague job description produces costly misfits; the difference between micromanaging and setting clear expectations; what a strong onboarding plan includes; and where AI can support hiring without replacing a manager’s judgment.

• A podcast episode also presents a case study of a business owner who turned around two struggling franchise locations with one strategic manager hire.
• This is a better fit when the agency is building content for a hiring, staffing or small-business audience, or when the campaign angle depends on the operational pressures of a growing owner rather than technology buyers.

The posts show no UX research method, AI content-production process or paid distribution.

• The campaign decision described in this page’s fit section—who can improve a B2B technology client’s content and paid-media strategy—is only partly covered: the people-and-role content could transfer to a hiring platform or franchise audience, but nothing here addresses reaching product, security or data leaders.
• For a reproducibility problem, look at Jammet; for content for a staffing campaign, look here.

Examples: Source 1 · Source 2 · Source 3

The posts describe using Blue Ocean Strategy to map where a company overlaps with the rest of the market and where it can lead, then building the whole story around those differences so the product becomes harder to compare.

• The RSAC Innovation Sandbox post adds firsthand observations about how security startups now present themselves: the recurring question of how a company differs from the big AI firms, increasingly bold branding in a crowded market, and slides and demos that have shifted from information-dense to clear storytelling and refined design.
• This is a better fit when the deliverable is positioning and messaging for a security startup launching or rebranding, particularly around industry events where the agency also wants a read on how peers are presenting.

This is brand and messaging work, not distribution.

• The posts offer no paid-media plan, budget argument or human-centered research process.
• To show how a defined buying committee would actually see the work, ask for a distribution plan.
• The approach is also specific to cybersecurity; it would need to be translated for an AI or privacy product sold to non-security buyers.

Examples: Source 1 · Source 2 · Source 3