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

Buğra Gündüz

CEO @ HockeyStack (YC S23) 🖤🏒

We build AI agents that increase revenue by centralizing customer data and understanding what drives revenue. Backed by Bessemer, General Catalyst, Y Combinator, et al. We hire exceptional outsiders that want to win. Email me at bugra@hockeystack.com if you're one of them.

Co-Founder & CEO

OpenAI selected HockeyStack as a GPT-5.4 launch partner.

We've been running our revenue agents on GPT-5.2 since we launched. OpenAI gave us early access ahead of release, and the jump is not subtle.

The clearest example is in our meeting briefs. Before a rep walks into a call, our agent pulls 400,000+ tokens of CRM history, call transcripts, email threads, and deal activity then tells the rep exactly what to do in the meeting, who's in the room, and what risks to watch for.

With 5.2, the output was good. Directionally correct. Useful.

With 5.4, the risks are more precise. The participant roles are richer. The model actually triangulates across the full context instead of pattern-matching on the most recent signals.

That difference matters when you're running agents across an entire sales organization.

We're not building a copilot. We're building the operating system for enterprise GTM - one where agents run the pipeline and humans focus on judgment and relationships. Every model improvement compounds directly into revenue outcomes for our customers.

5.4 is a meaningful step toward that.

CEO @ HockeyStack -- Building AI agents for sales & marketing

45% of our team at HockeyStack are immigrants. Their resilience, work ethic, and determination drive our success every day. The American Dream is alive, and we’re proud to be a part of it.

Our immigrant team members bring diverse perspectives, innovative ideas, and an unmatched drive to excel. They remind us that hard work and ambition can overcome any obstacle, no matter where you start or what challenges you face.

Their stories inspire us all to dream bigger and push further!

At HockeyStack, we are committed to fostering an inclusive culture where everyone, regardless of background, can thrive.

I believe the American Dream will never fade because of people like them. Their success stories aren’t just personal victories; they’re a testament to what’s possible when talent and opportunity come together.

Here’s to building a future that lifts everyone up.

America (specifically San Francisco) is still the best place to build.

Last year, we closed a deal with one of the largest AI companies in the world. It died 11 times, org charts shuffled twice, and after 7 months we finally got to Closed Won. Here is the step-by-step breakdown of the sales process (and lessons learned):

DEAL BREAKDOWN

Cold Call → First Foot in the Door

Marketing had been warming the account for months — webinars, thought leadership, ads, but no one raised a hand. One well-timed cold call landed a 30-minute intro.

Meeting 1 – Champion Identified

The AE hit gold: the person who picked up became our internal champion. Pain points: disjointed attribution, siloed GTM data, manual reporting.

Meeting 2 – Stakeholder Demo (Tech Team)

45 minutes. 20+ technical questions. We left with a tentative yes, if Security signs off, you use our models only for our instance, and more.

Meeting 3 – Technical Scoping

We started scoping the project for the team, what we need to do on our end, and even more technical questions around AI.

Meeting 4 – Org Restructure

Our champion’s boss got promoted, two new VPs parachuted in, budget frozen.

Revival #1 – Value Re-Frame

We learned the new priorities for the business, started building our business cases around them, and started gaining momentum again. Multi-threaded with other champions and got them on separate meetings.

Meeting 5 – Executive Alignment

We started involving C-level and VPs from our end. Multi-threaded even more. Found introductions from our network, investors, and customers.

Mid-Cycle Curveball – Champion Quits

Champion moved to another AI company. Momentum nosedived.

Revival #7 – Multi-Threading Pays Off

Because we’d already looped in RevOps, Data Science, and Finance, we still had three internal sponsors. Security slowed the process again.

Final Mile – In-person event

We had a CMO/CRO event in San Francisco and invited our champions. Luckily, they were free and attended. Had the chance to connect in person, have a good time, and build a real relationship. Around 10.30 pm, we got the verbal. They met with 2 customers that week for reference calls.

Signature Day

Docusign at 6:14 p.m. PT. Slack went nuts. Team popped bubbly 2 minutes later. First deal that Engineering team was more pumped than our Sales team.

LESSONS LEARNED

1. Cold call is not always demand generation. If marketing’s warming the pond, a single call can hook the whale and capture the demand.

2. One champion is none. Multi-thread by the second meeting or prepare for a funeral.

3. Security is your secret sales stage. Treat it like a mini-deal with its own discovery, objections, and success metrics.

4. Deals don’t die; reps let them. Every “no” is just a new play to run. Reference call, pilot, new stakeholder, fresh business case.

taplio

The Battle

Leah Tharin

Buğra Gündüz

Followers

Posts

Engagements

Likes

Performance ScoreLeah Tharin's LinkedIn Performance

Performance ScoreBuğra Gündüz's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Leah Tharin

Buğra Gündüz

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 ScoreBuğra Gündüz's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreLeah Tharin's LinkedIn Performance