Thinslices

19/10/2026

Some product risks can be settled before the build, and some only show up once real users or running code are involved.

The useful question before committing a build budget is which is which.

The slides map four checkpoints, from the design phase to running code, and close with five questions to answer before the money is spent.

Any question nobody can answer is where the remaining pre-build work belongs.

Article link in comments.

15/10/2026

Engaging a product design agency before the groundwork is done wastes money and extends timelines. https://hubs.li/Q04spm_k0

Validated problem. Defined scope. Someone who can approve a direction without broad consensus.
Without all three, a short discovery is the cheaper way to buy clarity than the full engagement.

14/10/2026

Three questions worth asking any product design agency before signing anything.
1. What happens when discovery shows the original direction is wrong?
2. How do you recruit real target users, not internal teams?
3. Who actually does the work, and are they dedicated or split across clients?

Link in comments.

13/10/2026

A clickable prototype can now be built in an afternoon. So what is it still proving when it reaches whoever approves the build budget?

AI tools have loosened the link between polish and the thinking behind it.

The slides cover four checks, plus three things a prototype can never tell you. Start with the failure path. A prototype that only works when everything goes right usually comes with an estimate that assumes the same. Link in comments.

12/10/2026

Fair question for software companies that spent the last few years waiting to see where AI would land. We spent them building with it.

That time taught us where AI shortens delivery, where it quietly adds rework and how to spot the difference before it reaches a client’s roadmap.

Hard to be afraid of something you already know how to use.



06/10/2026

19.3% is the net average productivity gain Gartner reports for enterprise AI coding agents. It's a realistic figure, and almost nobody planned around it.

The same Gartner guide says 90% of engineering leaders see improvements. McKinsey's 2026 survey found 80% of respondents reporting better individual productivity, while 37% attributed any EBIT impact to AI and only 6% attributed more than 5% of EBIT.

The feeling scales faster than the return.

DORA's 2025 research adds a caution worth taking seriously: teams adopting AI tend to ship more, and their releases tend to get less stable. Its read is that AI amplifies whatever a team already is. If review and release are the slow parts, they stay slow.

So when the renewal or the next pilot expansion comes up, the useful question is a simple one: what number are your agents returning? Link in comments.

Gartner reports 90% of engineering leaders see gains from AI coding agents, at a 19.3% net average. Why the gap appears, and how to measure it first.

25/09/2026

You can choose the right vendor and still get a weak product design phase. Part of the outcome is decided on your side of the table, before kick-off.

Six things the phase needs from you:
1. A decision-maker who attends, not one who signs off at the end
2. Experts with their calendars blocked for week one
3. Technical lead time that survives price negotiation
4. A shared definition of "complete"
5. Technical context on day one
6. A calendar that fits the timeline

None of them depend on the vendor. All of them are easier to fix before the contract than halfway through discovery.

The carousel covers the rest: who runs the phase, what AI changes and what to expect at the end. Save it for the next proposal that lands on your desk. Full article in the first comment.

Which of the six is hardest to get right on your side?

23/09/2026

The easiest line to cut in a product design proposal is the most expensive one to lose.
It goes like this. Three roles on the proposal: a product manager, a UX designer and a technical lead, the senior developer who owns architecture, feasibility and the estimate. The total comes in higher than expected.

Nothing is being built yet, so the technical role looks like the obvious line to trim. The technical lead drops to a few days. Everyone signs.

Four weeks later there is a prototype that looks right and an estimate nobody can defend. No one had the time to research the integrations, test the uncertain parts or check each flow for feasibility while it was still being designed. So the estimate rests on assumptions, and the first development sprints become the place where those assumptions get tested.
That is exactly what the product design phase was paid to prevent.

If the budget cannot carry a proper technical allocation, there is a better trade: narrow what the phase covers instead of thinning the role. A smaller scope estimated properly is worth more than a bigger one estimated on guesswork.

We wrote up what a product design phase needs from the buyer before it starts, including the questions worth asking any vendor about their technical lead. Link in comments.

Which line in a proposal have you seen cut that later cost the most?

Who runs a product design phase, how much of your time it takes and the six decisions on your side that shape the estimate you get back.