Amy Mitchell
banner
amycmitchell.bsky.social
Amy Mitchell
@amycmitchell.bsky.social
Product manager and author of Product Management IRL

https://amycmitchell.substack.com/
https://www.linkedin.com/in/amycmitchell/
Here's what AI-first product managers do:

🔹 Start with AI before you have the answer
🔹 Test an idea against potential scenarios
🔹 Come to discussions with options, not conclusions
🔹 Talk about outcomes, not features

"Starting without AI feels incomplete."
Mustafa Kapadia
January 10, 2026 at 3:30 PM
Progress isn’t one big “yes.” It’s dozens of tiny signs.
Sales asking questions.
Marketing previewing messaging.

None of these look like success on their own.
Together, they’re momentum.

Great product managers don’t wait for validation.
They build it. One signal at a time.
January 9, 2026 at 2:05 PM
If you feel stuck, it’s usually not because you lack answers.
It’s because you haven’t written down the right questions yet.

Writing questions is underrated product work.
It turns uncertainty into something you can actually move.

Clarity isn’t found. It’s drafted.
January 8, 2026 at 2:05 PM
AI explains your product before customers talk to you.
If your product can’t explain itself, AI fills the gaps with partial context.
Explainability is where context engineering becomes real.
Why Explainability Is the First Real Test of Context Engineering
AI evaluates your product before people do. Builder PMs design systems that hold up without them in the room
buff.ly
January 7, 2026 at 2:05 PM
Product work rarely gives quick signals.
No instant metrics. No immediate validation.

So product managers do the dangerous thing:
They wait for confidence to appear.

Momentum doesn’t come from certainty.
It comes from creating your own signals of progress.
January 6, 2026 at 2:05 PM
The silence after a win is the real test of product leadership.
You get exec buy-in. The next steps look obvious.
Then your team goes quiet.

That silence isn’t resistance—it’s a lack of direction.
People need paths, not just conviction.

Product leadership starts after the yes.
January 5, 2026 at 2:05 PM
Is your AI feature suffering from "silent failure"?
Product managers often treat AI like traditional code—ship it and move on. But AI requires "gardening," not just building.
More on handling post-launch decay for product managers: buff.ly/Hv8Q1yN
Why Most AI Features Fail After Launch (And How PMs Can Prevent It) | HackerNoon
Read up on why most AI features and products fail after launch and learn how product managers can prevent it.
hackernoon.com
January 3, 2026 at 3:25 PM
When product managers raise concerns without artifacts, it sounds like intuition vs velocity.

When you share eval notes, screenshots, and recordings, the conversation changes:

- concerns become visible
- tradeoffs become explicit
- alignment happens faster

Evals don’t slow teams down.
January 2, 2026 at 2:05 PM
You don’t need perfect coverage to ship confidently.

Evals trade exhaustive testing for prioritized validation:
a small set of high-impact, high-risk scenarios that reflect real customer use.

The goal is “we understand what could go wrong...and we’re aligned on the tradeoffs.”
January 1, 2026 at 2:05 PM
Each year, I write about what feels different in product work.
Lately, that difference shows up in flatter organizations, faster AI-enabled product work, and more ambiguous initiatives.

buff.ly/PCz6887
2026 Trends in Product Management
Where product manager influence comes from as structure disappears
amycmitchell.substack.com
December 31, 2025 at 2:05 PM
Evals are about testing what actually matters.

Product managers understand why:

- real customer journeys
- business context
- the failure modes that cause churn or escalation

Evals capture product judgment and make it runnable — so confidence doesn’t live only in one person’s head.
December 30, 2025 at 2:05 PM
Ever notice how everything can be “working as expected” and you still hesitate to ship?

It’s about customer workflows crossing team boundaries in your product

Evals give you a way to turn that intuition into something visible and actionable.
December 29, 2025 at 4:37 PM
Product managers who want AI search accuracy do this:

1. Keep your product knowledge base current

2. Make it easy for humans and AI to gain information

3. Help buyers know about you

This post from Eli Schwartz has more on AI visibility for product people.
SEO in Spotlight: C-Suite Prioritizes Visibility and Credibility | Eli Schwartz posted on the topic | LinkedIn
SEO teams are in the spotlight now more than ever. Why? 𝗕𝗲𝗰𝗮𝘂𝘀𝗲 𝘁𝗵𝗲 𝗱𝗲𝗰𝗶𝘀𝗶𝗼𝗻 𝗹𝗮𝘆𝗲𝗿 𝗺𝗼𝘃𝗲𝗱. The C-suite just started paying attention because their board, investors and…
buff.ly
December 27, 2025 at 6:01 PM
If performance and security deserve roadmap space,
so does operational maturity.

Without it, growth slows.
PM focus erodes.
Success becomes friction.

Scaling is not side work.
It is product work.
December 26, 2025 at 2:05 PM
“Who owns this?” rarely fixes anything.

A better question is:
What capability is missing for this product to scale?

That shift turns ownership fights
into growth conversations.
December 25, 2025 at 2:05 PM
You can ship a lot of launches and still feel like nothing’s moving.

I don’t think that means product managers need to “do GTM.”

I think it means some launches aren’t finished yet.
When Your Launches Ship but the Product Doesn’t Move
How finishing launches differently reframes the dilemma of GTM vs delivery for product managers
buff.ly
December 24, 2025 at 2:05 PM
When processes only work with you involved,
handoff is impossible.

Not due to resistance.
Due to dependency.

You are not hoarding work.
The system is not ready to let go.
December 23, 2025 at 2:05 PM
Your product is working.
That is why everyone keeps looping you in.

Sales questions. Custom requests. Onboarding edge cases.

Not because you should own them.
Because you always have.

Early success creates hidden drag.
December 22, 2025 at 2:05 PM
If you find yourself re-explaining your context and product to AI, then Rian van der Merwe has some time-saving ways to get more value from working with AI.

"Iterate on the prompts. If a prompt isn't helping, change it."

bit.ly/4oUNvnU
How I Use AI for Product Work
How I use AI for product management: PRD and OKR reviews, debate frameworks, technical explainers, and a Windsurf workflow with @ mentions and MCP
elezea.com
December 20, 2025 at 1:53 PM
You don’t grow beyond execution by fixing everything.

You grow by choosing:
• which friction is systemic
• which confusion is blocking others
• which work is worth walking away from

Strategy isn’t cleanup.
It’s building paths others can follow.
December 19, 2025 at 2:05 PM
Strategic product managers don’t ask for recognition.

They make chaos understood:
• patterns instead of noise
• narratives instead of updates
• decisions before escalations

That’s why leadership keeps pulling them “away.”
They’re not being pulled out — they’re being pulled up.
December 18, 2025 at 2:05 PM
“Follow the money” is good advice — just not when you’re pre-revenue.

Early product work needs a different kind of thinking.

I wrote about the work that comes before the money shows up.
The Work That Comes Before “Following the Money”
Product decision-making in the earliest stage
buff.ly
December 17, 2025 at 2:05 PM
If the hardest parts of your product job aren’t in JIRA anymore, that’s not scope creep.

It’s a signal.

You’re starting to work in the space between teams —
where strategy, translation, and influence actually live.
December 16, 2025 at 2:05 PM
Ever notice how some product managers never “own” features — but everything moves faster when they’re around?

They’re not avoiding product work.
They’re connecting it.

That role is real.
And more product managers are moving into it than they realize.
December 15, 2025 at 2:05 PM
Launch your product fast or experiment before launching? Itamar Gilad provides a case study of these approaches for adding an AI-powered feature.
December 13, 2025 at 2:02 PM