super.work

Super is your team ultimate AI search engine and assistant, based on all company data.

Founder

Slite

Founder and CEO 🛰

Staffit

Launched in college, Staffit let recruiters without technical skills assessing their candidates.
We were offering proper tests based on projects, rated with peer assessments (candidates reviewing each others in a controlled manner).

Co-founder

CEO at Slite (YC18) | We build the first knowledge base on autopilot, in remote since 2016!

Currently hard at work building Super 🔥

Super is our ultimate AI search and assistant. It plugs to your company data and comes with the recipes and tools for your team to get instant benefit from AI at work
 👉 super.work


And of course we keep growing Slite! Your team dead simple yet crazy powerful knowledge base, since 2017. 

I'm fortunate to lead a diverse team of exceptional individuals spread across 10+ countries 🇮🇳 🇬🇷🇱🇹 🇦🇷 🇬🇧 🇳🇱 🇫🇷 🇳🇴 🇺🇸 🇪🇸 🇩🇪 🇨🇦 🇵🇹 . Our journey includes going through YC W18 batch, raising $15M+ in funding from top-tier investors like Spark Capital and Index Ventures, and serving tens of thousands of teams monthly.

Christophe Pasquier

OpenAI's top 30 customers list got leaked?
And by the way things look like
SaaS competition is going to be BRUTAL 

Look at who's consuming the most tokens: Notion, Salesforce, HubSpot, Zendesk, Shopify, JetBrains. These aren't simple products. Far from it. They're the most complex, feature-rich platforms in their categories. They've spent years adding features, building depth, creating power. And in doing so, they created a massive problem: 

Their products became too complex for most users to navigate.
AI is their answer to that problem.

Eg: Salesforce has thousands of features across dozens of modules. Most users touch maybe 5% of what they're paying for. But with AI, you can just ask "show me all deals likely to close this quarter" instead of building a custom report.
Same with Notion. Same with HubSpot. AI becomes the interface that makes complexity accessible.

At Slite and super.work, we've got our hats in the ring too. Knowledge management tools have historically struggled because finding the right information means navigating structured and unstructured data from ALL your apps. We made Slite into the best Knowledge Base but Super will actually solve the information asymmetry problem.

If you're building software today, you're automatically in the AI business now. It doesn't matter if you're in productivity, dev tools, legal, healthcare, e-commerce. AI isn't a feature category anymore. It's infrastructure.

The companies on this list figured that out.

CEO at slite (W18) & super.work | Deep AI search & agents connected to all your work apps

Every knowledge base sells you on centralizing everything.
Then you realize your docs can't answer half your team's actual questions.
We spent 8 years building Slite before accepting this.

Let's say your documentation is perfect with every tag, folder, and doc.
But then someone asks "why did this customer churn?" and you're suddenly opening Intercom, Slack, and HubSpot because the answer lives across all three.
Good questions will always need good context. That's why we built Super.

Slite gives you verified, official documentation. Super finds everything else.
For a year, we sold them separately. Watched teams buy both because they needed both.

So here's the Knowledge Suite.
Slite at $10/user/month. Super at $15/user/month.
Together: $20/user/month.

You're getting both for less than competitors charge for one mediocre product.

Companies are choosing the suite it because fighting reality is exhausting. Your knowledge will always live in multiple places. The suite lets you stop pretending otherwise.

Document what should be official in Slite. Search everything that exists with Super.

If you're tired of choosing between perfect docs and complete context, book a demo. 👉 https://lnkd.in/eSwiyTjp

The lost prospect that pissed me off the most?
A French fintech company we'd been courting for months.
Perfect fit on paper, great logo to have, their team was excited about everything we built. Then at the last minute they chose our biggest competitor.

All because our competitor was selling Kanban boards + databases + project management + wikis.

I realized that the leads we were getting wanted an all-in-one solution and couldn't tell the difference between a standard all-in-one and Slite. But Slite had always been opinionated about just being a great knowledge base, but somehow our leads had hesitations because they wanted more.

The answer wasn't to ship more features and catch up to our competitors. That would lose our opinionated feature set of being decluttered and beautiful. We needed to communicate more clearly about who we were.

We chose our lane and repositioned fast. Instead of competing on breadth, we doubled down on being the best knowledge management tool that actually works. We focused on speed, search quality, and AI-powered assistance.

The pivot accelerated our growth. Once we clearly stated that we were for teams who cared about having the best documentation app, the magic happened.

Prospects started choosing us because we solved their specific problem better than anyone else. We still kept getting prospects that wanted an all-in-one, but the new positioning and messaging let them self-disqualify and saved the pressure on our sales team of trying to sell to customers that never wanted something like Slite in the first place.

That lost deal helped us do that. Sometimes you'll lose a deal so right that it'll shake you to re-think everything you thought you knew. And that's when magic in positioning and product thinking happens.

But damn, that rejection stung at the time.

Leaving this ✨ crew after 1 week of outstanding work and now ➡︎ on my way to San Francisco!

Last week we were preparing the future of Slite and super.work, this week I plan to show it to as many potential buyers and founders to see if our next iteration resonates!

If you are in town and 
- need Super for your team
- want to boost your team speed with instant business answers
- need a more accurate and cheaper alternative to your Guru, Glean Notion AI, etc
- work on RAG and AI productivity

Hit me up! 

And for all the rest stay tuned, we have the 2nd version of Super almost ready for show time, it truly will be the best company search in the world (we measure it).

The SF 996 work culture gets criticized a lot, but honestly? It's the default for most founders I know.

It also applied to most startup employees before ~2018, before the massive COVID funding wave.

In 2012, most of my friends out of college joining startups had way lower salaries than others, they were genuinely excited about innovation and building things that hadn't existed before, they wanted to be here for the ride.

My cofounder Pierre still tells me stories of running migrations at nights, working and sleeping in his first company office. And I don't think he'd have traded this for anything in the world.

If you ask anyone who went through this in their 20s in the last decade, I believe the story would have been the same. The long hours weren't glorified or bashed, they just were, but I believe for the right reasons.

It felt sustainable because everyone was doing it for the right reasons (and most people were too young to have families 😅).

996 is a taboo issue now because the "tech world " norm went from startup to scale-up, from bootstrapping to being funded day 1, from joining and be part of a mission to optimising your career by joining the right rocketship. Something shifted, and we began to criticise 996 because we were suddenly judging startups as mature companies.

But, making a startup work is a game against all odds. It requires a crazy amount of attention and care. It's not for everyone, and not everyone should join a startup.

But if you're a startup founder, looking to create generational products, at least for a given point in time, 996 is necessary. And if people join and enjoy this rhythm for a few years, and get rewarded for it, so be it.

It should not be the norm of work, but startups are not the norm, they're the outliers.

People don't understand Claude skills but I couldn't be more pumped
10y ago, I wouldn't have thought I'd be hyped about folders with MD files 😅
But hear me out, Claude Skills are great.

We've been trying to figure out how to package AI expertise for the last few years.

Custom GPTs were too rigid and quite limited.
MCPs are just tools, and need to be wrapped to deliver value.
Claude Projects were quite basic, just a preset of files and simple RAG to prompt on.

Skills might crack it because they're stupidly simple.

So what's a Skill: folder with context + optional code ready to be run. Claude reads a 2-line description. When you need it, Claude loads the full thing to get the extra context + get your custom code tools ready.
That's it.

Here's why that matters.

Your designer spent weeks getting your brand guidelines perfect. Fonts, colors, spacing, logo placement. Now every deck needs to follow them.

Old way: Copy-paste the guidelines into every prompt. Hope people remember. Get inconsistent results.

Skills way: Package it once. Everyone on your team asks for a deck, Claude applies it. No copy-paste. No drift.

You need a chart in there? add some python code to do it. Or generate end embed images, just running a script to the midjourney api.
Need custom team data to feed this deck? use Super MCP in the loop.

Same for code review standards. Support ticket responses. Meeting notes that become action items. Contract redlining rules.
Build it once. Share it. It works the same for everyone.

Skills validate something: simple modules that load when needed, executable when required, shareable across teams.

Anthropic is testing abstractions. This might not be the final form. But it definitely feels like a natural way for how LLMs work. And we'll definitely take inspiration from them if not use them as a basic format standard in our products.

Only 10-15 books on startups have achieved cult status. Every founder has either read it or has it on their shelves. 
But 2 of them set off my internal red flags every time I see them.

They are

1. OKRs - John Doerr's Measure what matters
All companies will implement OKRs at least once in their life.

Maybe it's the new VP at the big org who does it to give the illusion of impact.
But soon the idea spreads and suddenly, you see anxious first-time founders will opt for it as a solution. And it makes perfect sense to them on paper - OKRs signal you're a serious company with clear structures.

But no one tells how to do it, your startup evolves every 3 months, your team is dynamic, and half your tracking for Key Results is a farce.

On the other hand, a large company doesn't have that kind of internal volatility. They're probably an established market player, with a 4+ level hierarchy system where the groundwork becomes opaque to the ones at the top floor. Moreover, they don't have a sense of culture to propel them forward. So OKRs come in, ensure every department knows their lane and HOW they'll impact the bottomline.

You, as a startup, need a strong culture, flat hierarchies, and constant iteration on what you're shipping and how you're selling it.

2. Marty Cagan's Product books
Every PM has read Inspired. It's gospel in product circles.

The ideas make sense when you're at a company with real scale. You've got product-market fit, dedicated teams, engineering resources that aren't stretched thin.

Discovery becomes valuable because building the wrong thing is expensive.
You can afford to say no to features. 
But early-stage founders read the same books and think they need to "professionalize" their product process.

The founder-PM creates a discovery backlog while also doing customer support and trying to close deals. You start separating discovery from delivery when the whole point of being small is that you can do both simultaneously, faster than anyone else.

Cagan's frameworks prerequisite a stability you don't have. They're tools for optimization, not for finding your way in the dark.

And I get it. As a new company, you're always wondering what you're doing wrong. Reading books help you quiet that nagging fear. But as a startup, there's really only 3 things to take care of:
1. Culture
2. Shipping speed
3. Finding PMF

So read wisely, selectively, and don't worry so much - you'll do just fine.

This customer is beating us at our own playbook 🔥🔥

Uscreen is a 100 person team, spanned across 73 countries, fully profitable.

And they've already nailed an issue most companies suck at: Context. 

Their knowledge is perfectly retrievable. Be it an employee or an agent, they can just ask a question about anything related to the company, and get the context to unblock their work.

Sure, it saves the Slack pings of hunting for files, but more importantly, that's the foundation that makes AI tools actually work when you deploy them.

In 9 days, Jaimie (their Chief of Staff) is joining me live to walk through how they built this.
How a distributed team across 73 countries keeps their knowledge accessible.

You'll see how they
- structured their documentation so it's actually findable.
- built the culture where writing things down is just how the team works
- built a team that never blocks each other with information assymetry

This is peer-to-peer learning. No pitch deck. Just a 45-minute conversation with someone who solved this problem.

When: October 29, 2025
Time: 5 PM CET / 10 AM EST

Register here: https://lnkd.in/emBXRsf8

So pumped to have you on this one Jaimie!

Your sourcing agent is hunting for a product marketer. It finds two candidates:

Profile A: 5 years at Google, launched 3 major campaigns, MBA from Wharton, expertise in growth marketing and analytics.
Profile B: 3 years across 2 startups, built marketing from scratch at a 10-person company, no formal degree, specializes in scrappy growth tactics.

The AI agent sees both profiles and thinks: "Perfect matches!"

A human recruiter immediately picks Profile B. Not because of the resume, but because of context. The human recruiter has been in your Slack chats, seen the long threads about moving fast and avoiding corporate processes. They've watched candidates from big tech companies struggle with your lack of structure. They know that your last Google hire quit after three months because they couldn't handle the chaos.

This accumulated wisdom can't be captured in a prompt. You can't distill years of hiring mistakes, cultural observations, and team dynamics into four sentences and expect AI to make the same nuanced decisions.

Yet that's exactly what we keep trying to do. We feed AI generic job descriptions and wonder why it returns generic candidates. We give it surface-level requirements and expect deep insights about cultural fit.

The problem isn't the AI. The problem is context.

AI that truly understands your company needs access to everything your human recruiter knows. The Slack conversations about what went wrong with the last hire. The performance reviews that revealed why certain personalities clash with your team. The CEO's offhand comments about the kind of person who thrives in your environment.

That's what super.work does. It doesn't just read your job description. It understands your company's hiring DNA from every rejection email, every successful onboarding, every cultural misstep. It crafts prompts you wouldn't think of because it knows context you've forgotten.

You need better context delivered fast, otherwise you'll never capitalize on AI's potential.

Without it, you'll just create more noise.

4 months into selling our first AI product and first learning: you need someone in Slack DMs with your biggest prospects.

Palantir Technologies deployed Forward Deployed Engineers 10 years ago to be close to the customer and iterate to a perfect deployment with the customer.

Right now, all AI companies need the same motion to close their biggest customers.

We've closed half a million in revenue since launching Super, and still, the main thing that helped us win all the pilots (beyond core quality ofc) was...
🥁🥁🥁🥁
... our white gloved, 3-month pilot phases.

Any time a Super prospect comes in
→ We offer them a trial
→ They get a Slack channel with us
→ They run into a wall? They text us and we jump immediately
→ They spot a bug? We resolve within 24 hours
→ They get a crazy idea request? We try to ship a proof of concept the same week if it was part of our plans.

The same trend is happening in enterprise deals too. FDE roles are skyrocketing, with GTM over-promising and product under-delivering, you need to do some plumbing to truly embed AI in their team.

Onboarding needs to stay close-knit until the AI category matures to the point where customers know exactly what to expect and product can deliver that with 100% reliability.

Until then, it only makes sense for B2B sales to adapt, even for mid-market
- A real Solution eng + AE to close
- A Product specialist come work hand in hand post sales for internal deployment

In this weird early phase of AI product, this supplements an enormous amount trust to your customers that they're heard, that their concerns are valid, and that you'll help them implement the tool and scale with their team.

Once you prove that you're trustworthy, you automatically get a leg up from your competitors in the trust race. Even if they offer 1 more integration or a better UX for now, your customers will KNOW that you're there, hearing them, working with them to make your product into something they can safely roll out to their team.

All enterprise search tools look the same 🫠
We all promise enterprise search, custom assistants, and transforming how work gets done.

While we're the best in class in terms of answer accuracy, we too sound the same. 

Any advice would be VERY appreciated 😫

Animalz

Director of Marketing and Innovation

Senior Content Marketing Manager

Becoming Superhuman

Writer and Editor

Saent

Cofounder and Product Manager

KaiOS Technologies, Inc.

Marketing Director (global)

Happylatte

Head of Marketing and Publishing

DanceTelevision

Managing Director

Yourzine

Sherpa Media

Co-Founder & Creative Director

Timmetz Media Productions

Freelance video creator (music, travel, media)

Production Manager

📣 Marketing @ Animalz | ⏳ Productivity @ Saent

Always navigating the crossroads of marketing, creativity, and productivity. Current roles and projects:

Crafting high-quality content at Animalz for tech customers like Parabol.
👉️ animalz.co/blog
👉️ parabol.co/blog

Writing a about personal productivity for Saent and Bercoming Superhuman.
👉️ saent.com/blog
👉️ becomingsuperhuman.science

Co-creating Lifeline, a macOS Pomodoro timer that boosts focus and productivity.
👉️ https://apps.apple.com/app/apple-store/id1526186940?pt=118017664

I'm on an indefinite management sabbatical—previous roles and projects:
☎️ Marketing Director (Global) at KaiOS, a mobile operating system.
🎵 Managing Director at DanceTrippin, a global music platform (online + broadcast).
📣 Managing Director at Yourzine China, a digital marketing agency.
🕹️ Head of Publishing and Marketing at Happylatte, a mobile gaming studio.

I'm originally from The Netherlands and work remotely from Koh Samui, Thailand, where I live with my wife and son.

If you've made it this far, let's connect! I enjoy discussing marketing, productivity, AI, and remote work.

Tim Metz

You open a document from a coworker. Looks professional at first. Formatted nicely, plenty of text. But as you start reading, something feels off.

The content is vague.
Key details are missing. 
You're not sure what you're supposed to do with this.

Then it hits you: they probably just had AI generate this without thinking it through.

If this sounds familiar, congrats, you've been workslopped!

On social media, low-quality AI content is called "AI slop." For the work context, researchers from BetterUp Labs and Stanford have coined this 'workslop:’ AI-generated materials that masquerade as good work but lack the substance to meaningfully advance a task.

The researchers surveyed 1,150 full-time employees and found that:
- 40% received workslop in the last month
- Each incident takes nearly 2 hours to deal with
- For a 10,000-person company: $9 million/year in lost productivity

That’s not all though. The interpersonal cost might be worse.

Half the people who received workslop now view the sender as less creative and capable.
42% see them as less trustworthy.
37% see them as less intelligent.

One person in finance described it like this: "It created a situation where I had to decide whether I would rewrite it myself, make him rewrite it, or just call it good enough."

Workslop shifts work downstream. From the person who should have done the thinking to the person who now has to decode the content, infer the missing context, and figure out what to do.

Have you been workslopped recently?

(Full Harvard Business Review article in the comments.)

Ever since ChatGPT came out, I've been trying to create an AI style guide checker. It's never worked, until I tried with Claude Code last week.

The reason for trying again is a bit shameful: our Animalz style guide was outdated, and I'd sometimes finish articles so close to deadline that I couldn't run them past our copy-editors anymore.

(Deadly sin, I know.)

In trying to have a solution for such situations in the future, I gave Claude Code a challenge: reverse-engineer our style guide from scratch.

1. Analyze three blog posts from 2021, 2022, 2023, 2024, and 2025 (15 in total) along factors that style guides and copyeditors typically pay attention to.

2. Build a new style guide from the patterns.

3. Flag conflicts for me to resolve.

Critically, I told Claude Code to use multiple agents, each one focused on a different aspect: headers and capitalization, punctuation, grammar, etc. And since Claude Code can search the web, it fetched all 15 posts directly from our blog -- no manual downloads needed.

With this done, I asked Claude Code to edit an article for me based on this guide. Eight agents launched in parallel, each enforcing their piece of the guide. It actually worked, and I got back a copy-edited version of my article!

You can replicate this yourself with Claude Code, whether you’re working from an existing style guide or reverse engineering one from scratch based on content you already have.

(Not sure how to get started with Claude Code? We’ve just released a getting started guide on the Animalz blog -- link below in the comments.)

Remember the AI style guide creator I built with Claude Code? It's now a published plugin. In an Animalz plugin marketplace. That didn't exist 30 minutes ago.

And I didn't build any of it.

Here's what happened:

Anthropic announced Claude Code plugins a few days ago. I thought: "Maybe my style guide creator could be one of those."

So I opened Claude Code and said: "Can you analyze how existing plugins work, convert my blog style guide creator into a plugin, and also create a marketplace to distribute it?"

It did all of it. In one session, with one click:
- Analyzed existing plugin structures from GitHub
- Converted all my agents and commands to plugin format
- Added proper documentation, examples, and copy
- Created marketplace infrastructure
- Published everything to GitHub
- Created a v1.0.0 release

For context: Last week I reverse-engineered our Animalz style guide by having Claude Code analyze 15 blog posts. Then I had 8 specialized agents review my articles against it.

Now that workflow is a plugin anyone can install.

Launch Claude Code, then use these commands:

/plugin marketplace add animalzinc/claude-plugins
/plugin install blog-style-guide-creator

That's it. It's open source and free on GitHub: https://lnkd.in/gY_7tzPH

More plugins coming soon I guess! 😃

P.S.1. Claude Code then also wrote 80% of this LinkedIn post.

P.S.2. Not sure how to get started with Claude Code? We just published a getting started guide on the Animalz blog (link in comments).

taplio

The Battle

Christophe Pasquier

Tim Metz

Followers

Posts

Engagements

Likes

Performance ScoreChristophe Pasquier's LinkedIn Performance

Performance ScoreTim Metz's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Christophe Pasquier

Tim Metz

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);}Christophe Pasquier's LinkedIn Performance

Performance ScoreTim Metz's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreChristophe Pasquier's LinkedIn Performance