The B2B Podcast Index
Index
All categories
MarketingSalesSaaSFinanceHROpsLeadershipCustomer SuccessAI & DataProductStartups & FoundersRevOpsEngineering & DevTools
MethodologySubmit
Best of:MarketingSalesSaaSFinanceHROpsLeadershipCustomer SuccessAI & DataProductStartups & FoundersRevOpsEngineering & DevTools
An independent project byFame
SearchBest episodesGuestsInsightsMethodologySubmit a podcast
Index/Product/All Things Product with Teresa and Petra
All Things Product with Teresa and Petra artwork

AI-Shaped Problems

All Things Product with Teresa and Petra · 2026-06-30 · 15 min

0:00--:--

Key moments - from our scoring

Substance score

32 / 100

Five dimensions, 20 points each

Insight Density7 / 20
Originality7 / 20
Guest Caliber6 / 20
Specificity & Evidence7 / 20
Conversational Craft5 / 20

Rather than hunting for big, transformative use cases, both hosts advocate starting with the mundane daily tasks already on your to-do list. Teresa's core practice - picking one thing from her daily work each day and testing whether AI can help - has become her primary learning mechanism after a year and a half of daily use. This approach sidesteps the chicken-and-egg problem of not knowing what to use AI for until you understand what it's good at. Petra complements this by highlighting how exposure through community (her monthly Claude Code Show & Tell sessions, blog posts like "AI at Home" and "AI at Work") helps people see the possible and sparks their own ideas. The hosts push back against AI FOMO and over-engineering: you don't need MCP servers, plugins, or agentic setups from day one. Simple copy-paste into Claude or ChatGPT is enough. They stress that bias for action and agency matter most - just trying something, failing, and iterating beats paralysis. Even a failed prompt can become a learning loop: "this was terrible, how can you help me do better?" Both admit they've ignored most trendy tools (Teresa has never used Cursor; she's only used one MCP server) because they stay problem-focused. The key insight: stop looking for inspiration in the AI tool ecosystem; look at your own task list and play.

Key takeaways

  • →Start with one task from your daily to-do list each day and experiment with AI, not by hunting for transformational use cases - this habit reveals what AI is good at and drives real learning.
  • →Exposure to what others build (through communities, show-and-tells, blogs, or YouTube) plants seeds of possibility without requiring you to adopt their exact solution; you'll adapt it to your own needs.
  • →Ignore the noise of new MCP servers, plugins, and tools unless they solve a specific problem you have right now; problem-first thinking collapses an infinite universe of options into three relevant ones.
  • →You don't need to know how to use AI upfront - just open Claude or ChatGPT, describe your task, and iterate on bad responses by telling the AI what was wrong and asking it to improve.
  • →Bias for action and agency mean trying something imperfect immediately rather than waiting to understand everything; failure is part of the learning loop, not a blocker.

In this episode

  1. 1The AI Learning Challenge: Identifying Problems Big Enough for AI
  2. 2Building Daily Habits to Experiment with AI on Real Tasks
  3. 3Recognizing What AI is Actually Good At Through Repeated Testing
  4. 4Finding Inspiration Through Community and Seeing Others' Use Cases
  5. 5Starting Simple: Avoiding Overcomplicated AI Setups and Integrations
  6. 6Problem-First Approach: Using Tasks to Cut Through Noise and Hype
  7. 7Bias for Action: Overcoming Fear by Just Trying Something

Mentioned

ClaudeChatGPTCursorMidjourneyTrelloTeresa TorresPetra Wille

Guests

Teresa Kors

Topics in this episode

TrelloClaudeChatGPTMidjourneyCursorMCP serversAI at Home (blog post)AI at Work (blog post)Claude Code Show & Tell sessionsbias for action

Questions this episode answers

How do I figure out what tasks AI can actually help me with?

Pick one thing from your daily to-do list each day and try it with AI, testing the boundaries of what works and what doesn't; this habit reveals patterns over time about where AI excels, and you'll keep discovering new capabilities even after a year and a half of daily use.

Should I set up MCP servers, plugins, and integrations to start learning AI?

No - start simple with copy-paste into Claude or ChatGPT, then only add integrations if they solve a specific problem you're facing; the vast majority of new tools and servers are noise unless they directly address your current task.

I feel behind on AI and don't know where to start - what's the first step?

Have bias for action: open Claude or ChatGPT, type "I have to do this task - how can you help?" and start iterating on the response; you don't need to know how to fix it, just tell the AI what's terrible about its answer and let it improve.

How can seeing other people's AI projects help me learn?

Exposure through communities, show-and-tells, blogs, and videos plants seeds of what's possible, which often sparks memories of forgotten ideas you can build for yourself - you adapt their approach to your own context rather than copying it outright.

What if my AI task seems too small or niche to be worth doing?

Size doesn't matter - even bespoke projects that have zero market value are worth building because they teach you what AI can do and give you pride in what you've made; the learning and momentum are the real benefit.

What our scoring noted

Our reviewer’s read on each dimension, with quotes from the episode.

Insight Density

7 / 20

The episode surfaces a few practically useful ideas - the daily one-task AI habit, forcing yourself past wrong assumptions about AI's limits - but these sit inside a lot of conversational filler, mutual agreement, and well-worn advice that any attentive B2B operator will have already encountered.

i still Forced Myself To do This because i Still have assumptions about what AI won't be good at. And those assumptions are becoming not true every day
The problem is we don't recognize those things until we start using AI

Originality

7 / 20

The 'start with a real task, ignore the tooling noise' framing is sensible and mildly pushes against AI-hype FOMO, but it's not a genuinely contrarian or first-principles argument - it has become the standard mainstream onboarding advice and is not developed further here.

I get To ignore all The noise. it's only relevant to me if It's helping Me With the task of doing right now
I am NOT downloading every new AI app that comes out! I ignore the vast majority of this stuff

Guest Caliber

6 / 20

There is no external guest; this is a two-host conversation between two product coaches/consultants. Both have practitioner experience in product management but they are discussing their own personal AI habits, not operating at scale in a B2B context that would be directly actionable for founders, sales leaders, or ops practitioners.

Hi, folks. This is all things product with Petra Wille and Theresa Kors
I wrote an AI at home blog post. I wrote in a I work blog post

Specificity & Evidence

7 / 20

A handful of concrete specifics appear - one MCP server used for pushing to Trello, a community member who built a concert review website, monthly Show & Tell sessions - but there are no business metrics, dollar figures, adoption data, or named company outcomes to anchor the claims.

I Have used ONE MCP server A grand total of one. And I used it for a very specific task- ...of pushing content to Trello
One guy Built this concert review website For himself. Its very elaborate. It probably has zero market other than him

Conversational Craft

5 / 20

This is a friendly co-host chat with near-constant mutual validation ('Yeah,' 'Exactly,' 'Me too,' 'Me neither'); there is no external guest to probe, no challenging follow-ups, and no moment where either host pushes back on or stress-tests the other's claims.

Yeah, that's how i like to go about it as well more or less yeah okay
Yeah! And i don't use a million mcp servers. you know The vast majority of what's out there is noise. Yes!

Conversation analysis

Computed from the transcript - who did the talking, and the words that came up most.

Most-used words

help10terrible7problems6enough6problem6started6build6task6skills6share5trying5learning4behind4built4teresa3episode3

Episode notes

What you'll hear: The daily habit Teresa used to go from feeling behind to using AI confidently every day Why starting with terrible results is actually the point How to filter out the noise (MCP servers, plugins, new apps) and stay focused on what matters Why finding community - whether that's show-and-tell sessions, YouTube, or local meetups - accelerates learning faster than any tool A simple first prompt for anyone who doesn't know where to start Key takeaway: If you have a to-do list, you have problems AI can help with. Start with one. Give it 15 minutes. It'll probably be bad - and that's exactly right. Timestamps 00:00 Intro - are people's problems too small for AI? 02:15 Teresa's daily habit: test one to-do item with AI every day 06:40 The "feeling behind" problem - and why everyone feels that way 09:00 Petra's time-boxing approach vs.

Full transcript

15 min

Transcribed and scored by The B2B Podcast Index.

Hi, folks. This is all things product with Petra Wille and Theresa Kors And we're so happy. you here Teresa. a few weeks back We recorded in an episode how you and I go about learning?

All Things. AI and i was talking About my time boxing approach and You were Talking about your. I just use it to solve Challenges and Problems and Tasks that I'm Having And ever since I'm carrying this around in conversations with my coaches and what I am realizing, that would like to hear your thoughts about it is people who are not Teresa Torres seem...to have too small problems maybe?

To really create meaningful change with AI or not enough inspiration What they could use AI for. so that is a lot of the conversation That I'm having there. They try to inspire them to think bigger when it comes to what things they could try AI support for. So any thoughts on that?

Yeah, I don't think The problem is that they don't have enough big enough problems learning curve here, right? So how do you don't know what to use AI for until You Know What AI is good at and like How To Use AI? so there's a little bit of Like A chicken in an egg problem. I can share with.

i did it And then I Can also Share how I'm trying to help other people. yeah this exact same Problem. so the first thing that I Did Was I just started to build a daily habit Where I tried to pick one Thing That I had to Do In my Job Every single day pick one thing that I had to do in my job and try To see if AI is good at it. So, I literally like okay?

I have ten things on my to-do list today. This is the one i'm gonna Have ai take a cut out And sometimes Ai was good at It and I Was pleasantly surprised. A lot of times Ai was awful At it But then it became this exercise Of Like Okay Well how can I help The Ai be better at it? and that's what drives the learning, right?

Like I'm just experimenting. I'm trying a lot of things. I am testing the boundaries of What is it good at And what does not get out and honestly after doing like five times i started to see oh this Is The type Of thing ai' s Good at an all share even on A year-and-a-half in of Doing This and i still Forced Myself To do This because i Still have assumptions about what AI won't be good at. And those assumptions are becoming not true every day, right?

Yeah because it's so rapidly changing what AI can do. and yeah I agree. So i think like everybody...like literally everybody-I don't even mean knowledge workers!

I mean literally everybody could be using AI to help them with something So like, I don't think it's true that people don't have things the AI could help with. However many billions of people we have on the planet they all have something AI can do to them. The problem is we don't recognize those things until we start using AI. so this habit of trying one new thing every day and doing a really great foundational habit That's what I see a lot of people being blocked by is this.

when I started this journey, i felt really behind. I think most people look at me and they don't think i'm behind anymore but i start-I started feeling really behind. so...i think this is that um all of us are behind.

yeah like what is the proverb? The best time to plant a tree was fifty years ago and the next best day right just get started. It can take five hours, but you don't have to let it. Take five hours so You could try one small thing a day and if it's not very good at it?

You can shoot move on want to spend A little more time to try to make it good or do I just wanna Move On And If it's particularly busy Day you might Just move on? That's what my Time Boxing Comes in. yeah if you got a Little bit extra Time you Might get Curious and Follow a Thread Right. Yeah, the other thing that I've learned and I read very broadly And one thing this helped me is just seeing what other people do?

So This Is something That i'm trying to curate for Other People. so I wrote an AI at home blog post. I wrote in a I work blog post. both of These Blog posts Just list lots and Lots and Lots Of ways you can use AI and I'm not I don't, i didn't write those saying this is how you should use AI.

I wrote them as This Is How I Use AI? Hopefully this inspires You to see how you might use it in your own way right. and then um...I also host monthly Claude Code Show & Tell sessions where people from my community just demo what they're doing And one.

It's a Way To Like Build Momentum and Inspire People into like. Let people show off what they're doing, which is really fun. People love the story. Yeah like it's just super fun and Just like when we were in elementary school We got to go to show and tell them share our favorite toy.

It's this same idea Like hey I built This cool thing. Here's What it does for me. And sometimes people Share things that are so bespoke. One guy Built this concert review website For himself.

Its very elaborate. It probably has zero market other than him, but he loved it. And he learned a ton and was so proud of that. the rest were blown away by how much you is able to do.

He's not an engineer never built anything before. So like doesn't matter if this thing isn't relevant for you? It shows what possible And then it might spark this seed of like, oh I've always had this idea that I forgot about. If you could build this than I can build that!

and i think thats really important is to find peers in community whether its through my community or by just watching youtube videos local meetup It doesn't matter right? Find a way To just be exposed to what's becoming possible. Yeah, and don't be too stressed by it We've recorded an episode on AI FOMO. I know.

go back listen to that one but just briefly The more complex these AI setups become the more people are afraid of Using it to some extent because they think do too much magic things which is then out of their control and stuff like that. I constantly find myself in these conversations where i tell people you don't have to connect it to everything on your machine. This is definitely not something that need to do, start learning what actually can do and is capable about. you can keep a draw the conversational end.

you can copy and paste text into your prompts and see how this turns out and then at some point maybe off. fine with it to create MD slash text files on your machine, but you don't have to go full agentic API MCP logins for everything from the beginning. I think that would be weird anyways. so making new friends.

Yeah people shouldn't do this all day every. No, me neither. Right? Like I let it read and write file like text files but It doesn't have access to my email.

Yeah And i don't use a million mcp servers. you know The vast majority of what's out there is noise. Yes! This why we talked about in our previous episode that I really frame as...

I Have problem need to solve..i have piece work im trying to do. Im not going out and try every new technology. I'm not looking for MCP servers, for the sake of MCP Servers.

I am NOT downloading every new AI app that comes out! I ignore the vast majority of this stuff. so like i said im a year and half in using it everyday. still have never tried Cursor which blows people's mind.

but I don't need it. Me neither! Im pretty happy with my tool...I have used ONE MCP server A grand total of one.

And I used it for a very specific task- Me too! of pushing content to Trello, right? Um...I don't really use plugins.

I actually find the plugin interface to be terrible. I steal skills from ideas from other people and look at their code so that i can learn how to build it. Yeah but I don't steal skills form others because thats another thing I can't, i have to create these skills myself. To get the best outcome possible and i get my inspiration from other people's skills but i always have to build them myself.

yeah i didn't mean i actually take their skills...i mean.. i steal their ideas from their skills. okay gotcha what they built never works for me out of the box.

no it's a value of this is you get to make it your own right? The hyper personalization is what we're looking for exactly Exactly. So I think, like the reason why i find starting with a task or starting with A problem to solve is so helpful Is that? I get To ignore all The noise.

it's only relevant to me if It's helping Me With the task of doing right now and then That Like takes this infinite universe And turns into like. okay There's like three things that might be able to help me Right Now. Let me have Claude explain those three things to me and then I'm going to pick one. Yeah, that's how i like to go about it as well more or less yeah okay.

so that's How People Could Find Big Enough Problems And Enough Inspiration To Learn More About AI On Their Own Terms? One thing on close with is we talk a lot about agency We Talk A Lot About Like Bias For Action and I think those are the right terms to be throwing around, but i'm not sure we fully understand what they mean. And I think uh always when we do something for the first time there's this fear of the unknown. yeah like I'm looking at a task...

I know I need to do the task! I'm gonna play Trice's game and say how can AI help with this? and then immediately you're going to feel stuck..you're going.

I don't even know where to begin. And this is where you want to remember those terms bias towards action. have some agency, You don't have to know how do it all? All you have to do is try something.

It might totally fail. that's okay. literally Try something and not try something could be open up chat GPT Open up Claude Literally type i Have To Do This Task. How Can You Help?

And then read the response and take it from there. Like that's it. like The magical thing about these tools is you don't have to know how to do anything. They will walk me through it, and It'll probably be terrible.

but let's be honest the first response Will Probably Be Terrible? You can just be like. you literally Can be like. this was terrible.

What can I tell you? To help you Do a better job Mm-hmm so you Don't even Have to Know How to fix it. you Could Just be like This was terrible. And then the AI will ask you, what is terrible about it?

What did you not like?" Okay well here's maybe what I could know. right and so... I had such a fun session with mid-journey yesterday which took us nowhere!

It was exactly the same conversation. i was like oh my god all these pictures are terrible!! So we really need to rewrite this prompt from the ground on because the results are not great!!! Then we did it and now.

So, I think you started this with. people might not have big enough problems for AI to help. If You had a to-do list you've problems that i can help With? I like the framing right very helpful!

I Think it's just play A little bit. Yeah go Play! I Like That. thanks Teresa.

Thanks Petra.

Related episodes across the Index

Other episodes covering the same guests and topics, from across The B2B Podcast Index.

  • How we built Grok Bot in a month | Roman Ugarte (SpaceXAI)Lenny's Podcast · on Cursor90 / 100
  • Wins Above Replacement: The New Way to Judge FoundersVenture Unlocked · on ChatGPT88 / 100
  • Paul Graham On Startups, Ambition, and Great FoundersY Combinator Startup Podcast · on ChatGPT88 / 100
  • The AI-Native Law Firm, with Ryan Walker of General LegalMeeting of the Minds · on Claude88 / 100
  • 657. Waziri Garuba, CEO of Harlem Labs, Introducing G.R.I.O.TUnleashed · on Claude80 / 100
  • Unresolved.cx - When Feedback Has To Matter - Paul TuckerUnresolved.cx · on Claude80 / 100

More from All Things Product with Teresa and Petra

All episodes →
  • Quality Of Evidence80 / 100
  • Creating Experiences65 / 100
  • Organizational Change Is Exhausting75 / 100
  • Learning Together75 / 100
  • Procurement61 / 100
Explore the best B2B Product podcasts →
All All Things Product with Teresa and Petra episodes →