ReachUp

Join the waitlist 📧

Code For Africa (CfA)

AI Ethics Fellow

Africa Center for Digital Transformation  [ACDT]

Deputy Coordinator

TechForward Policy Fellow

Solutions for Youth Employment (S4YE)

The S4YE YAG is composed of 100+ enterprising global leaders who:

Provide input on the design of S4YE's and World Bank’s youth employment programs, through participation in board meetings and events.

Highlight youth innovation and enable the S4YE coalition to tap into young people’s perspectives.

Gain exposure and visibility in international networks.

The principles are Diversity, Respect, Innovation, Capacity-building, and Knowledge-sharing.

Member of the Youth Advisory Group 

Wilton Park

Research And Policy Analyst

CANAL+ Group

Product and Marketing Manager, myCANAL TV application for Africa and French overseas departments and regions.

Product Manager - Africa & DROM

Retention Marketing Officer

INSIGHTEURS

Co-Founder & Writer

Neopress LLC

Co-Founder- Head of Growth

Pulse

International Project Developer

Catholic Relief Services

Business Development Specialist

Writing to learn, learning to write. Sharing insights, ideas and stories about product, marketing and tech. Product Marketing Manager.

*The content and free resources I share are my personal views and do not reflect my employer's stance.*

Connect[....]ing the dots.

Adamou Boubacar Amadou

Emilia Korczynska

Passionate about SaaS product growth, and both the pre-sign up and post-sign up marketing. Talk to me about improving your acquisition, activation and retention strategy. 


Head Of Marketing

Doing all it takes to get more Product People over the code barrier with our code-free Product Adoption, Marketing & Experimentation Platform:  
Designing, developing and implementing a marketing strategy that ensures higher visibility 
Leading our stellar content team 
Implementing and driving a paid ads strategy to diversify customer acquisition. 
Driving funnel growth
Implementing an email marketing strategy 
Identifing and testing new acquisition channels
Overseeing, training, and coaching team members

It’s been exactly 3 months since we launched our first ever ABM campaign, and the results are in: over $650,000 in pipeline in 90 days, with a nice $12 in pipeline per $ spend. If you‘re considering getting into ABM next year - I have a few learnings to share: 

1) It’s brutal. There are no playbooks. Most ABM resources are very high level (‘strategic’) - there’s a painful lack of tactical resources on how to set the campaigns up, from whom to target and which channels to choose, how to set up the campaigns stages and account scoring, how to pick the right goals and metrics (so you can evaluate how the campaigns are doing and if you are on the right track), set the right budgets, and how to monitor the results. There are to benchmarks to measure against (ad CTR/CPM benchmarks are a vague leading metrics, and are so industry-specific you’ll be comparing apples to oranges.) We had to come up with nearly everything ourselves. One thing that helped us at the beginning was Kyle Poyar’s article “Your guide to GTM metrics 2.0” featuring “ABX benchmarks” (link in the comments). We used them (and tweaked) to set up our stages, benchmarks of how many accounts we’re expecting to move from one stage to another, and then - by combining that with our historical qualification rates, close rates and expected ACV - we got our revenue goals per 1000 accounts. Then working out was from the top-down revenue goals - that gave us an idea how many accounts we need to target. And then - based on the CPMs for the target personas in those accounts - we are able to come up with a budget. That’s theory, of course - and we’re already seeing how ‘real data’ diverges from the benchmarks, esp. for some stages/conversion points. We may need to adjust our forecasts based on real data to make them more realistic. But you need start somewhere. 

2) Don’t start if you don’t have extremely strong MarketingOps (ideally - a dedicated MOps person). So far we’ve spent 80% on the campaign ops, and 20% on the creatives. SADLY. Even with ABM tools - there’s a ton of setup you need to do.  

3) You need some tech stack to score your accounts, but you don’t need these expensive “ABM tools”. After evaluating tons of ABM tools (with platform fees of $30k - $160k per annum 🤯) we realised we don’t want to use display as our main channel for reaching target accounts. We used only ZenABM (from $59 - for de-anonymizing LinkedIn ad engagements, getting intent based on those engagments, account scoring --> and pushing it all to the CRM + providing revenue attribution) + Hubspot Marketing (~$1700 in our case - for audience building and management).

TBC - I hit the character limit 😉

VP of Marketing @Userpilot. Author @ Product Rantz.

Last week everyone was publishing or trying to get mentioned in listicles. Even I signed up for a “listicle management” tool. This week we’re deleting listicles. Google volatility is through the roof and some SEOs like Lily Ray (link in comments) have pointed out that it seems to be affecting especially sites that have engaged in excessive “AEO tactics” - using self-promotional listicles (where you mention yourself as the “No 1 tool for {Use Case}” in “Best tools for {Use Case}) both on their own site and on “guest blogs” - often using reciprocal (linked and unlinked) mentions. This gained popularity because of they affected “AI mentions” - RAG systems are fallible to listicle & recency bias.

I know it’s hard to resist tactics that are “working” for everyone else. I’ve also engaged in my own fair share of this. This is a stark reminder that producing content designed just for “gaming” AIO/AEO/GEO is a short-sighted tactic. 

Here are some other tactics that could also get you in the doghouse: 

⚫ Artificial “refinishing” - editing articles only to game “recency” factor (also a popular RAG-gaming tactic) - e.g. changing “Best user onboarding tactics 2025” into “…2026” - and changing the publication date without actually changing the article;
⚫ Thin programmatic content - lacking the plausibility that what is mentioned has ever been used/done (esp if you publish 800 of those with just different KW variations & some “spintax”…)  
⚫ Not cracked down on yet but I see it coming: Fake EEAT building by using first person singular and authors that obviously didn’t publish the posts (because it’s humanly impossible to crank out 70 pieces of “thought leadership” in a moth…)

Now - I’d hate the SEO community to do just a “blanket ban” on listicles right now. Helpful listicles *are* helpful. But we need to go back to what it means: 

⚫ Audit your site to list all listicle posts using search operators - site:example.com/guide/ best "1. ExampleBrand" or site:example.com/guide or - if you had some variation: (best OR top OR leading) ("1. ExampleBrand")
⚫ Delete such articles written for 0 SV Kws that aren’t getting traffic and don’t have any outbound links pointing to them, redirecting to “pillar” listicles
⚫ Save only the few ones that are really relevant and have been getting traffic and links, and focus on rewriting those; 
⚫ “Objectify” your listicle by splitting it into categories - let’s face it, there’s no “one size fits all” objectively “best” tool for everyone's needs. Rewrite your content to include nuance - and say objectively, that your tool is the best for e.g.   mid-market but not startups on a budget.
⚫ Rewrite each tool’s description based on actual use experience - you can tap into user review sites for that
⚫ For AI mentions, incorporate listicle elements naturally into other content types - strategy guides etc. - showcasing best tools for specific use cases in the right context.

Nothing gives me more imposter syndrome than being in this "group photo" 🙈 - but since I somehow* (*it's organized by my uglier half's company...does this make me a nepo baby? 🤔) make my way there - I will do my best to share the experience of running our 7-figure ABM program on the LinkedIn ABM Bootcamp 2026 - ( https://lnkd.in/eetWvsXX) put together by ZenABM alongside the *legends* of LinkedIn ads, ABM & GTM engineering: 

 - Tim Davidson - speaking about Thought Leader Ads for ABM

 - Patrick Spychalski - speaking about ABM with Clay

 - Gabriel Ehrlich - doing a Live Q&A: LinkedIn Ads for ABM

- Ali Yildirim🌲  - speaking about Multi-Channel ABM: How to Combine Paid Media + GTM Engineering
 
- Philip Ilic - covering LinkedIn Ads + Signal-Based Outbound Strategy

- Tas Bober - Why Your ABM Campaigns Deserve Better Landing Pages

- Maximilian Herczeg - Campaign Setup & Avoiding Mistakes That Kill ABM Performance

- Kamil Rextin - Enterprise GTM Systems

- Casper Emil Sciuto Rouchmann - ABM in 2026: The Ups and Downs of the Old Dog

- Katya Tarapovskaia - From Segmentation to Signal: AI + ABM Workflows to Win Target Accounts Faster

Excited (humbled...nervous 😅) to be part of this amazing lineup!  

The cohort consists of a hands-on core program of classes & workshops that will let you launch in 4 weeks, supplemented with guest lectures on adjacent ABM topics.

Cohort start: 26 January 2026

Seats are limited - 40% are already gone so hurry up!
 
 Don’t wait to sign up!

My husband has this conspiracy theory around enterprise software products that he calls “Salesforcification” - most of the API stuff flies right over my head, but TL;DR - that some companies make their products unnecessarily complicated so they can justify high prices. I feel the same way about a lot of the content published by consultants here - that sometimes they wrap things up in a way that makes the task at hand look so complicated and difficult, that is leaves the prospect with no doubt that they wouldn’t be able to handle it alone. I don’t know if it’s an intentional tactic, or a result of going ‘too deep’ into the rabbit hole. But a word of caution here: I’ve seen an inverse correlation between ‘cognitive load’ of the message and user engagement with paid social. If something sounds too difficult to comprehend at a glance (and a glance is all you get with paid social) - people don’t spend their precious time trying to figure out what you mean when mindlessly scrolling LinkedIn. Most of them aren’t in the market, and even those that are - aren’t in the ‘problem solving’ mode (like users landing on your website from cold search ads = actively searching for specific solutions). 

So to all agencies, consultants, SaaS companies launching Thought Leader Ads (as well as image ad) - keep it simple. Think of the words you’re using - can they be ‘dumbed down’ to reduce cognitive load? E.g. can you replace “without engineering dependencies” with “no code”?

Even Sam Altman found the Anthropic’s ads for Claude funny - but *why* are they so effective? A few years ago, I read Jonah Berger’s book “Contagious” - TL;DR - it described 6 principles of vitality - and imho, Anthropic ticked 4 out of 6 boxes with their ad:

So here’s why the Anthropic ads did so well, imho: 

1) Virality factor 1: Social Currency - People like things that make them look good, smart, or “in the know.” —> The Anthropic ad poked fun at ChatGPT, positioning it as this super cringe-y, embarrassing, intrusive bad character, that infiltrates your life and gives you misleading, deceptive or downright harmful advice for commercial gain. Hence - it made viewer feel subconsciously bad about using chatGPT - and good about using Claude (as the “good cop” in the AI story - this reminds me of Apple’s “1984” ads). The “coolness” factor of using Claude was reinforced by the “Keep thinking” slogan (as opposed to having your brain eroded with commercial messages, deceptively woven into harmless queries) - and with Dr. Dre’s “What's The Difference” soundtrack -  and that’s cooler than being cool. 

2) Emotion - not emotions are created equal: there are ones that drive action (sharing), and those that don’t. A rule of thumb - intensity beats positivity: high-arousal emotions (anger, awe, excitement, anxiety) drive sharing; low-arousal emotions (sadness, contentment) don’t.

—> Anthropic tapped into the arousing emotions of anger & OUTRAGE here - outrage at the intrusiveness & harmfulness of implicit ads in ai-generated content (reinforced by the words for the key emotions flashing on the screen), presented in this very graphic, relatable way - in the form of a relatable story. The characters in the ads could be us - or worse - our children. Which brings me to the next virality principle Anthropic ticked:

3) Stories. People don’t want ads - they want *relatable* stories. No comments  needed here - while they were a stretch for comic relief - we could feel the goosebumps of cringe thinking how something similar could happen to us, or worse still - a vulnerable loved one - if they used AI with ads…

4) Triggers - Linking the ad message to everyday cues,
e.g. associating KitKat with coffee → every coffee break becomes an ad for KitKat. —> you will remember Claude when listening to Dr Dre; or next time your chatGPT tells you “you’re absolutely right!” 

Other virality principles that the anthropic ad didn’t tap into: 

4) Public- If people can see it, they copy it. E.g.: White iPod headphones → free walking billboards (this wasn’t the case with this ad) 

5) Practical Value - basically “helpful content” Tips, hacks, checklists, benchmarks, that promise better performance or time/money/effort savings.

This is a reminder that vitality (unlike many of us think) - isn’t luck - it can totally be achieved by design. And looking at most B2B ads on Linkedin - I think we really need that reminder 🙃 This is also a note to self.

"Almost everything I worked on a year ago is now irrelevant."  

I met up with an old friend yesterday, who happens to be working on one of the largest tech companies' proprietary Large Language Model - and we had an interesting conversation on where LLMs are going and if/when agents will replace certain roles.

The question is: what jobs? And maybe in the nearest future employers will not be hiring for specific jobs at all anymore - but will be hire the person that can bring most leverage with AI (as per Alina Vandenberghe 🌶️'s post I read a couple of days ago - featured below - thanks for the food for thought!) 

It will be still hard to replace that systems thinking (some are trying, though) - and a whole lot of other intrinsically "human" skills: 

Taste & judgment - Knowing what shouldn’t be built or said.
Emotional intelligence - Reading the room, sensing tension, building trust.
Moral reasoning - Deciding what’s responsible, not just what’s possible.
Creative synthesis - Connecting unrelated dots into something new.
Leadership under uncertainty - Making decisions without perfect data.
Conflict resolution - Navigating ego, politics, and misalignment.
Cultural intuition - Understanding context, nuance, and subtext.
Relationship-building - Long-term trust over transactional output.

What about Product Management? Will it turn into "Agent Management"? 🤔 👇

Hey folks, I met up with an old friend yesterday, who happens to be working on one of the largest tech companies' proprietary Large Language Model - a

Open article: From Product Manager to Agent Manager? 🤔 by Emilia Korczynska • 3 min read

From Product Manager to Agent Manager? 🤔

Ironically, we're too busy to stop being so busy. “If we stopped the clock and didn’t have to generate pipeline, and didn’t have any time pressure, and you had all the engineering resources you need - how would your team change after we restarted the clock?” - I was asked today - and it’s a subversive question that I’ve been also thinking about. My team is really lean, but it’s been on everyone’s mind of course - that the only reason why a lot of tasks, and a lot of jobs, haven’t been automated out by e.g. Claude code yet is that we…are too busy to implement it properly. And we’re too busy because we can’t afford (or so it seems - perhaps very wrongly) to “stop the clock”. 

Then there is the security pushback, especially in enterprises. Pushback that companies like OpenAI and Anthropic are surely going to capitalise on - and likewise, SOC2 certified agencies. 

And then there’s the completely natural resistance to change. Things are tedious, but they are familiar and comfortable. Revamping the process completely requires learning completely new ways of working, redefining processes and roles. 

And then there are the job security concerns. Some people may be (consciously or unconsciously) pushing back to preserve their jobs. 

If it involves “dumpster-diving” content production or mindless number crunching or basic research or building marketing collateral with obsolete solutions and technologies - your concerns about your job are very, very real. 

The reason why we haven’t automated a lot of these jobs yet is not that that the technology isn’t ready - we aren’t ready.

Aazar Ali Shad shad recently wrote something here that is so *against* conventional wisdom that it stopped me in my tracks: 

What if being “on brand” is actually hurting you? 

This goes against everything we’ve been told in “marketing best practices” and by “advertising gurus” - I remember being told off by a consultant that our ads weren’t consistent enough = weren't contributing to “brand recall”. 

But think about most Enterprise brands you know. What is their “brand identity”? Vague, tall claims written in Open Sans on a dark gradient background? 

“Disrupt the future of AI with Acme Inc”. 

Stock photos of smiley AI-generated ladies dressed in casual smart in an office environment.

We scroll past those on social media without even registering their existence. 

So when my team did a “Competitor check” on LinkedIn’s Ads Library… I was like…“let’s not”. 

Remember: the fact that your competitor is 10x bigger doesn’t mean that whatever they are doing works 10x better. 

You don’t know their results. 

In fact - larger companies have more marketing dollars to waste - so they tend to be more forgiving of underperforming campaigns. 

AKA: publishing boring, bland, corporate pulp that can’t possibly move anyone’s finger, let alone the needle. 

Because “nobody gets fired for buying IBM” (Btw. It never fails to amaze me how far one can be carried by the sheer power of the established brand.) 

What Gartner said about you in their latest report? 

Great. I couldn’t care less. 

Remember that Linkedin is paid *social*. You pay to play, but you play on social media - and it only recently dawned on me…

…or actually it only dawned on me when we published this vide, and it got 1.1% CTR - the median CTR benchmark for SaaS* for Linkedin video ads is 0.24% CTR 

Why did it work? 

It was real. It was personal. It tapped into the TikTok trend “Come to work with me as XYZ”. It was pattern-interrupting. 

What if we just..used content that looks like it was posted by some? A real photo, a real selfie? 

Meanwhile: most of “on-brand” ads can’t do that. 

We may be winning brand recall. 

But what are we losing?

“Not on *my* wifi. Not on any device that connects to my wifi”. 

It was made very clear to me that I'm not getting the wifi in the event of divorce, so guess what - I’m still posting this manually, while all my friends’ OpenClaw agents are doing their work while they sip pinacoladas on a beach. Or getting their passwords changed and API keys stolen. Oh well. TL;DR - if you stepped away from the Internet for a couple of weeks, you probably missed the news that autonomous AI agents have launched their own social network, language, and religion (well, this was a prompt-prank, but sent strong cues of a looming AI apocalypse). 

The crux (or perhaps Claw) of the issue: OpenClaw’s architecture poses serious security threats. It's open-source, relies on community-contributed “skills” (plugins) for functionality. And it requires broad permissions on the host system.
 
Recent analyses by cybersecurity experts have highlighted multiple risk factors - from unvetted plugin code and prompt injection attacks, to misconfigured deployments exposing sensitive data.
  
Security researchers have discovered hundreds of malicious skills on ClawHub (OpenClaw's skill repository). Out of the 2,857 skills audited, they found 341 malicious skills (≈12%) across multiple campaigns. One large campaign (“ClawHavoc”) involved 335 skills that tricked users into installing a MacOS infostealer (Atomic Stealer) via fake prerequisites.
 
Attackers cloned popular tool names (e.g. “solana-wallet-tracker” or YouTube utilities), then embedded malicious install scripts in the skill instructions. Upon installation, these trojanized skills would fetch second-stage malware to steal API keys, credentials, crypto wallet secrets, and other sensitive data.
 
Notably, one of the top-downloaded skills on ClawHub was found to be part of such a malware campaign. A “Twitter integration” skill  instructed the agent to install a dependency via links that actually delivered a staged payload: an obfuscated shell command that downloaded a binary infostealer (a macOS trojan). Once executed, it raided the host for valuable data - browser sessions, saved logins, API tokens, SSH keys, cloud credentials - essentially anything the agent or user could access. 

These risks would naturally not be compatible with your SOC2 - particularly for enterprise and B2B environments (e.g. if OpenClaw is used for marketing or analytics on platforms handling customer data), which demand strict data governance, auditability, and compliance.

How can your organization have the (AI agent) cake and eat it? 

Read more to find out 👇

Hey folks, Two weeks is an eternity in tech - and since we last touched base, apparently AI agents have spawned their own social network, language and

Open article: ⛔ Strictly NOT for humans 🦞 - 🤖 Agentic security risks & how to address them  by Emilia Korczynska • 4 min read

⛔ Strictly NOT for humans 🦞 - 🤖 Agentic security risks & how to address them

We got hung up on listicles lately, but the “Self-promotional listicles” Google “penalty” has far more reaching implications - since producing content has become so much easier (if your uber driver is telling you about their Claude Code skills, it’s time to pause & reflect) - the number of pages to be indexed has ballooned - and more indexed pages means more storage space, more computing power to analyze, classify, and monitor them. This leads to increased operational costs (even if efficiency was improving). But not necessarily a better index. So there’s no upside either for Google or Search users to index all those pages. 

The question is - which ones should be prioritized. 

And here it gets tricky - in the past, there was a quite a big spread in page quality and signals between different pages. AI tool have become so much better that AI-produced content needn’t necessarily be of obviously worse quality at a glance. But if there’s little actual information gain - and everyone is just regurgitating the SERP with AI and plugging themselves in - Google needs to apply site-wide quality scores to judge which of the pages to index and which not. 

A 2024 Google exploit by Mark Williams-Cook Willams-Cook discovered a "site_quality" metric. The score determined at least whether your site would be eligible for showing in different SERP features. It’s a site wide weighted blend of multiple signals - from brand searches, through click probability, to topical focus & consistency (aka “domain relevance” - how topically aligned a domain is with a specific subject area) 

The crawl budget is already getting more limited (to 2 MB) and content freshness seems to play a stronger role (with pages not crawled in > 130 days more likely to be deindexed) 

And here’s the crux of the issue: most of us used the same SEO/organic traffic growth playbook for years: 

- Cover all the “BOFU” keywords - any KW indicating consideration, product awareness, high buying search intent “{{competitor}} alternatives”
- Cover all the KWs related to your product and use cases 
- Expand up-funnel to MOFU and TOFU keywords, targeting ones with higher SV first, plugging your product into more and more “far-fetched context” - whatever was vaguely relevant to your target audience. 

Now Google seems to be cracking down on that. It seems to be scoring your site first to evaluate whether you have the “license” to cover a specific topic in your content. 

But how is this license earned exactly? And won’t search become more monopolised by large brands? 

Organic traffic used to be this egalitarian space, this unlimited, renewable resource - now it’s gonna resemble performance marketing more and more, with limited slots, occupied by stronger brands.

I looked at the effective cost per click to landing page*  of our TLAs yesterday - and it was 5 x LOWER than that of Single Image Ads 🤯 (*NOT the CTR - look at the "Clicks to landing page" in the "Engagement" column view, and divide it by the number of impressions!) 

Should we replace our image ads with *only* TLAs though? 🤔

Not so fast...

Running ABM programs on Linkedin is like playing swings & roundabouts - I will be sharing my experience running them for the past 2 years at the meetup in London on Tuesday - alongside such ABM legends as Casper Emil Sciuto Rouchmann, Katya Tarapovskaia and Joana Petrova! 

Grab the last remaining seats from the link below! 👇

📅 Feb 24 | 6:30 PM | Shoreditch, London

taplio

The Battle

Adamou Boubacar Amadou

Emilia Korczynska

Followers

Posts

Engagements

Likes

Performance ScoreEmilia Korczynska's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Adamou Boubacar Amadou

Emilia Korczynska

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);}Emilia Korczynska's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreEmilia Korczynska's LinkedIn Performance