LLenny's Podcast
← All frameworks
StrategyJenny Wen (head of design at Claude)

Building Trust Through Speed

Ship an honest 'research preview' early, then visibly iterate on feedback so speed builds trust instead of eroding it.

Difficulty
Easy
Time to result
~weeks to results
Steps
3
Confidence
90%

When you can't perfect a product before launch, Jenny Wen's approach is to release it early, clearly labelled as a rough research preview, but bind that release to a visible, continuous iteration loop. Trust is preserved not by polish but by shipping fast, responding to feedback publicly, and demonstrably improving. The failure mode is releasing early and then going silent.

Origin

Articulated by Jenny Wen on Lenny's Podcast, drawn from how Anthropic launched Claude Co-work as a labelled 'research preview.'

Core principles

  • 01Ship when the benefits already outweigh the cons, even if quality is rough
  • 02Label rough releases honestly as a research preview and caveat the flaws
  • 03'This is the worst it's ever going to be' — set the expectation of improvement
  • 04The trust promise is the iteration, not the initial quality
  • 05Public responsiveness (fixing things people report, fast) is what maintains the brand

How to run it

  1. 1

    Decide there's already real value

    Only ship early when there is something genuinely powerful that some users will benefit from right now, even with known flaws.

    Pro tip Dogfood it internally first so you know the powerful core is real before exposing it.

    Watch out Shipping something with no real value early doesn't build trust through speed; it just ships something bad.

  2. 2

    Label it honestly and caveat the flaws

    Release it explicitly as an early research preview, telling users it may be hard to work with, may be low quality, and has issues.

    Pro tip Frame it as 'the worst it's ever going to be' to anchor expectations on the improvement trajectory.

  3. 3

    Make and keep the iteration promise

    Commit publicly to taking feedback and iterating, then actually respond to people, fix things quickly, and show continuous shipping.

    Pro tip Have the team publicly reply to comments and announce 'we fixed this yesterday' so people feel heard and used.

    Watch out You lose trust if you release early and then nothing ever happens; silence after a rough launch degrades the brand.

In the wild

Claude Co-work research preview

Anthropic labelled Co-work a research preview and shipped it knowing it had flaws, believing the benefits outweighed the cons, then iterated publicly on feedback and kept shipping fixes fast.

The launch 'blew up the internet' and trust held because the team visibly responded and improved rather than going quiet.

Common mistakes

Releasing early then going silent

The single fastest way to lose trust around an early release is to ship it and then let nothing happen; that degrades the brand more than never shipping.

Treating the label as the whole promise

Calling something a preview isn't enough; users must actually see their feedback acted on and the product continuously improving, or the honesty rings hollow.

Is it for you?

Best for

Product and design leaders launching AI or 0-to-1 products where usage can't be fully predicted before release

Not ideal for

High-stakes domains (health, finance, safety) where an early flawed release causes real harm rather than just friction

From the transcript

this is the worst it's ever going to be

25:00

we believe the benefits outweigh the cons

25:30

building trust through speed

26:30

if you release it early and then nothing ever happens

26:00

From the episode

The design process is dead. Here’s what’s replacing it.

Jenny Wen (head of design at Claude)