B2B Growth for Scaleups. leahtharin.com

B2B Product-led Growth & Sales

Honest, sometimes funny, battle-tested cross-functional Growth & Product insights, board member, growth advisor.

I love working in the operational details with my clients and am an Innovation Product & Growth Nerd.

I run PLG and PLS cohorts at Maven:
https://maven.com/leah-tharin/productledgrowth

and wrote Leah's Product-led Growth Guide:
https://www.leahtharin.com/p/the-product-led-growth-guide

www.leahtharin.com -> (Blog, Podcast, Guides, Rates to work with me. contact)

Leah Tharin

I'm not saying I called it, but I totally called it.

https://lnkd.in/e2UUdZHY

Totally deserved. Rock on Anthropic

Open article: Anthropic turns the tables on OpenAI in critical revenue category by axios.com

A line chart that shows monthly shares of first-time enterprise customers choosing OpenAI or Anthropic from December 2025 to February 2026. OpenAI fell from 59.7% to 26.7%, while Anthropic rose from 40.3% to 73.3%.

Anthropic turns the tables on OpenAI in critical revenue category

There's a 90% chance that you're browsing linkedin right now because Claude is down. #iseeyou

There's a 10% chance that one of your colleagues is having a heart attack right now because of it

Yes, I'm still alive. In other news, I almost subscribed to Whisper today. Almost. My Obsidian vault says I went down from 45 Subscriptions for business-related services to 12 in the past month.

Will write about it soon and probably a bit more when I'm done being completely exhausted from working, AI, context managing (keeping 1'000 md files under control, probably the most elegant thing i ever built in my life), and sleeping too little.

Anthropic, you guys are rocking it.

Also, I created a knitting/stitching app for Taylor Swift fans with Claude, and the people who use it are loving it. Don't ask, still proud.

Ciao.

Using Lovable for the first time for fun, after playing adult in Claude code for 3 weeks straight.

Can you tell that I'm a prompt engineer?

Pretty good stuff there Elena Verna, you should try it out too. Claude is commenting my efforts with:

"Oh Leah, Lovable doesn't know you like I do, I know what drama really means for you."

Immediate promotion.

(Lovable is cool though, can see how it's good with a context machine behind it, that feeds great plans)

I caught the same wrong fact in three different reports last week.

Numbers were correct. Charts were clean. Conclusions were garbage.
The date range included bot traffic. Nobody flagged it. The knowledge that those weeks were polluted? Lived in my head.

My AI caught it. Because I taught it that fact three weeks ago in one line of text.

FFS, we're spending billions optimizing retrieval. Bigger context windows. Better RAG. Faster inference.

Nobody is asking whether the fact being retrieved is actually true.
Retrieval correctness is not factual correctness.

The fix isn't a better model. It's a text file where you write down what you already know. 30 seconds. One correction. Every future output gets better.
I wrote about how I actually do this, why I think about reports like code commits, and the Sisyphus problem of building context.

👉 https://lnkd.in/egtVj5Ku

Product Manager

I bring structure to companies growing through uncertainty. Product Manager who's spent 6+ years joining fragmented organizations and building operating models that work across AdTech, cloud/cybersecurity, logistics and automotive. Certified in LLM foundations, AI agents and AI for Product Management. This allows for building products where AI creates measurable impact, not just features.

How I work:
→ Operating Models and AI that scale:
My philosophy: frameworks aren't bureaucracy, they're velocity tools. When feature approval took weeks, I introduced OKRs and next-now-later roadmaps tied to measurable business outcomes. Decision time collapsed to days. Generally speaking, I don't "manage” stakeholders, I use to build frameworks that create alignment so teams can ship without endless debate. I integrate AI across discovery: Claude code for interview, feedback synthesis and assumption mapping, Perplexity for competitive analysis, Miro AI and Lovable for same-day prototypes. The result: I validate value, usability, feasibility, and viability in days not weeks. I also use Claude Code and Cursor for API architecture decisions validation.

→ Platform and Architecture:
I use AI tools to prototype technical concepts with working code before engineering commits. At Brand Metrics: designed JSON definition standards and measurement frameworks breaking through scalability limits. The approach works because debates about architecture resolve faster with working prototypes. Hours instead of weeks. Previously at Iver: rebuilt fragmented customer portal and commercial workflows (4+ portals → 1 unified platform). Environment patching went from days to hours.

→ Customer Research and Strategy:
I use Jobs-To-Be-Done, Impact mapping, ICP definition (among others) to achieve evidence-based prioritization. 
Recently: 20+ discovery interviews at Brand Metrics identifying 2 new customer segments, expanding TAM by 5%.

→ Cross-industry pattern:
I've built products in AdTech (Brand Metrics), cloud/cybersecurity (Iver), logistics (Greencarrier), automotive (Bilia). Same pattern every time: enter unfamiliar domain, systematic customer research, deeply understand business outcomes,use rapid experimentation to deliver measurable impact. I then establish frameworks so teams can repeat without me. Pattern recognition across contexts.

Languages spoken on a daily basis: English, French, Spanish, Swedish

Open to: Product Manager roles in SaaS/B2B platforms where platform architecture, discovery rigor, and measurable impact matter.

Pierre Ortega

Ship it. Fast. Iterate.

AI agents felt like a black box to me a year ago. I understood the concepts but couldn't build one. As a product manager, the best way to stay relevant is to get your hands dirty, build product sense through 𝑑𝑜𝑖𝑛𝑔 𝑟𝑎𝑡ℎ𝑒𝑟 𝑡ℎ𝑎𝑛 𝑟𝑒𝑎𝑑𝑖𝑛𝑔: experiment, prototype, validate, stay close to the customers, own the business case. 

That's exactly what the AI Agents Buildathon organized by The Product Compass (Thanks Paweł Huryn and Olia Herbelin 🙏) gave me the opportunity to do. 🛠️ Link here: https://lnkd.in/dAZXbgJj

I engaged on Track 3 (AI Agent Architect), and our team built NextStep: a multi-agent system where 4 specialized agents coordinate to discover jobs, match profiles, generate tailored CVs, track applications, and learn from user feedback over time. Built on n8n for orchestration, Supabase, Lovable for the frontend, LangChain for Observability and Evals and OpenAI and Gemini as the underlying LLMs.

What a journey! From ideation to tech stack decisions, PRD, prototype and all the way to a defensible business case for VC funding. All of this in three weeks, with real teammates and real constraints.

Something that really struck me is that 𝑤𝑜𝑟𝑘𝑖𝑛𝑔 𝑤𝑖𝑡ℎ 𝐴𝐼 𝑎𝑔𝑒𝑛𝑡𝑠 𝑖𝑠 𝑟𝑒𝑚𝑎𝑟𝑘𝑎𝑏𝑙𝑦 𝑠𝑖𝑚𝑖𝑙𝑎𝑟 𝑡𝑜 𝑤𝑜𝑟𝑘𝑖𝑛𝑔 𝑤𝑖𝑡ℎ 𝑝𝑒𝑜𝑝𝑙𝑒: Concepts like guardrails, context, memory, LLM-as-a-judge, human-in-the-loop, evals, observability, and fallbacks aren't just engineering patterns. They're the same principles that make teams function. 🤝 

➡️ Give an agent the right context and guardrails and it performs
➡️ Without memory, every interaction starts from scratch. 
➡️ Without observability, you're flying blind. 
➡️ Without fallbacks, one failure cascades.

One moment made this concrete: a test user edited a generated CV (swapped passive voice for active and cut a sentence in half). Our Adapter agent picked up the pattern. The next CV reflected that preference without being told again. Same as onboarding a great new hire: give them feedback once, and they run with it.

This experience democratized something that felt like a complete mystery to me. 🔓 And if you're a PM who hasn't gotten your hands dirty with agents yet, the next cohort is the right moment.

Finding sources that actually help you cut through the AI noise with frameworks, you can use immediately and with no fluff, is genuinely rare. That's what The Product Compass does, and this Buildathon was proof of it.

👉Take a look at the gallery of other projects and vote for your favorite! https://lnkd.in/g9B5NPPe

#AIProductManagement #AIAgents #BuildInPublic #LLM #ProductManagement

Product Manager | B2B SaaS Platforms | AI Product Development & Data Architecture

Open article: AI Agents Buildathon for PMs: AI Solution Gallery is Live! by productcompass.pm

AI Agents Buildathon for PMs: AI Solution Gallery is Live!

taplio

The Battle

Leah Tharin

Pierre Ortega

Followers

Posts

Engagements

Likes

Performance ScoreLeah Tharin's LinkedIn Performance

Performance ScorePierre Ortega's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Leah Tharin

Pierre Ortega

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Leah Tharin's LinkedIn Performance

Performance ScorePierre Ortega's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreLeah Tharin's LinkedIn Performance