Unibrix

In 2024, your CRM’s AI feature suggested an email draft. In 2026, it qualifies the lead, writes the outreach, schedules the follow-up, updates the pipeline, and flags the deal as at-risk if they go quiet — without a human initiating any of it. And this is a structural change in what a CRM is. A few numbers that put this in context: → Salesforce Agentforce: $540M ARR, 18,500 customers, 3 billion agent workflows per month → HubSpot moved to outcome-based pricing in April 2026: you pay $0.50 per resolved conversation — only when the agent actually solves the problem → 83% of sales teams using AI reported revenue growth last year versus 66% without it → AI-native CRM startups have raised $200M+ in the past 12 months (Lightfield, Reevo, Monaco, Aurasell) But here’s the thing that’s not in the press releases: AI agents amplify the quality of your existing CRM data — good and bad. The biggest hidden cost in CRM AI is not the software. It is the data cleanup that happens before the agents can perform reliably. A few more things the full article covers: → The three generations of CRM AI: predictive (2016) → generative (2023) → agentic (2025-26) → What the Atlas Reasoning Engine actually does inside Agentforce → The CRM-native challengers and what architecture thesis they’re betting on → Why “system of record” is being replaced by “system of action” The CRM market is splitting between platforms where AI is a feature and platforms where the agent is the product. Full read → https://lnkd.in/d-y3X8FX

How do you keep hotel content consistent when the same photos and information have to live across Booking, Expedia, Google, Instagram, and partner channels? 🧳 For Keeps, Unibrix works on the software behind a travel CMS that brings those channels into one place. We maintain and improve the existing product, build new features, and develop the API integrations that keep everything connected. The idea is simple: update the content once, then make it useful everywhere guests might find you. ✈️ For hotels and travel brands, that means less manual content work and a much easier way to keep their digital presence up to date. See the Keeps case study 👇 https://lnkd.in/dUzpgDnb #Unibrix #TravelTech #APIIntegration

Fun fact: the “whole development team” is increasingly just one person. In Stack Overflow’s 2026 Developer Survey, 10.5% of respondents said their organisation was literally just them, up from around 4% the year before. Maybe the modern tech team is sometimes just one developer wearing five hats: dev, QA, PM, designer… and occasionally tech support too 💼 #Unibrix #SoftwareDevelopment

Stablecoins have been around for years, but what’s changing now is where the real activity is happening. Visa says nearly 17% of its stablecoin-linked card volume already comes from business and commercial programs, while Circle is pushing USDC and EURC further into enterprise workflows through the SAP ecosystem. That makes the conversation much less about crypto as an asset class and much more about the infrastructure around it: settlement, treasury, payments, compliance, and the APIs connecting all of that together. We looked at what that shift means for fintech builders in our latest article, The Stablecoin Shift: Why Fintech Infrastructure Is Quietly Changing. Read the full piece 👇 https://lnkd.in/dzAwwbNr #Unibrix #Stablecoins #FintechInfrastructure

The line between scrolling and shopping keeps getting thinner. TikTok and YouTube are pushing more shopping directly into video. SocialAF came at it from the other side: instead of adding commerce to social media, it added the social experience to shopping. We worked on a mobile platform where users could discover products through video, shop directly from that content, and browse together with friends. Social media used to send people to the shop. Increasingly, the social experience is the shop. Full case 👇 https://lnkd.in/deZdRWxa #Unibrix #Ecommerce #ProductDesign

Energy is one of those systems we rarely think about until reliability becomes a question. In Ukraine, across Europe, and globally, resilience has become a much bigger part of the energy conversation. And that resilience isn’t only physical — there is also a digital layer behind contracts, pricing, market data, trading, and everyday energy operations. That’s where Jules comes in. We helped build the product foundation and internal tools behind an energy platform designed around flexible contracts, pricing, and operational workflows. Another side of modern energy infrastructure that often stays behind the scenes. Take a closer look at the case 👇 https://lnkd.in/d7Wpjw65 #Unibrix #EnergyTech #EnergyInfrastructure #SoftwareDevelopment

Today is the Day of Defenders of Ukraine 🪖🇺🇦 Above all, today is about gratitude. To everyone defending Ukraine — on the front line and across medicine, communications, cybersecurity, engineering, technology, logistics, and countless other roles. Ukraine’s defenders protect our future while contributing to the security of Europe and beyond. We are deeply grateful to everyone serving, and we remember those who gave their lives for Ukraine. Thank you for defending our today, so that Ukraine can have a tomorrow✊

Yesterday was International Day for Universal Access to Information. A good excuse to bring back one of our enterprise cases — a bidding system built to make procurement data easier to access, compare, and actually use. Less digging through scattered information. More clarity on contractors, bids, and decisions. Full case here 👇 https://lnkd.in/d4pefTKf #Unibrix #EnterpriseSoftware

Tech hiring is a bit chaotic right now. There are fewer openings, hundreds of applications on some roles, AI can already write a decent amount of code, and every second post seems to announce either the end of software engineering or the beginning of its golden age. So we went looking at the numbers 📊 The first interesting thing is that the hiring slowdown itself does not seem to be primarily an AI story. LinkedIn still attributes most of it to the wider economy and cautious hiring. But underneath that slowdown, companies are clearly changing what they are hiring for. AI Engineer hiring has grown dramatically, while System Engineering and Data Engineering are already among the biggest areas connected to traditional software engineering. Jobs asking for specific AI skills are growing much faster than the overall jobs market. And here is the part we find more interesting. Knowing how to use AI probably will not stay a competitive advantage for very long. GitHub says 80% of new developers already use Copilot in their first week. When almost everyone has access to the same tools, simply saying you work with AI tells an employer less and less. The bigger opportunity may be around everything AI still needs before it becomes useful software: clean data, APIs, cloud infrastructure, security, testing, integrations, architecture, deployment and people who know how to put all of those pieces together. Which also changes the hiring question… For someone trying to enter tech, a small project that actually runs, connects to real data, handles failure and can be deployed may soon say much more than another certificate saying you know an AI tool. And for companies, replacing junior hiring with a search for ready-made senior talent has its own problem. PwC’s research suggests that some of the routine work people used to learn through is disappearing first. If companies remove the junior role at the same time, they may also be removing part of the pipeline that used to create experienced engineers. So after looking through the data, we are less worried about whether software development survives AI. The more interesting question is where the next generation of good developers is supposed to come from — and what companies will actually need them to be good at. #Unibrix #SoftwareDevelopment #TechJobs LinkedIn Economic Graph — Building a Future of Work That Works, 2026. https://lnkd.in/gTWzsUgc PwC — 2026 Global AI Jobs Barometer. https://lnkd.in/gMHFyxvD GitHub — Octoverse 2025, updated February 2026. https://lnkd.in/gyjSM5yn

Enterprise software usually does not become slow because of one bad decision. It gets there gradually, while teams are busy shipping, fixing, integrating, and moving on to the next thing. At some point, though, you notice the pattern. A “small change” that should take two days somehow turns into two weeks. Someone opens an old system, realizes that one integration depends on three others, half the documentation is outdated, and there is one part nobody really wants to touch because the last time someone did, something else broke. That is what technical debt often looks like in practice. Not one dramatic failure, but years of small compromises slowly making every release, integration, or product change harder than it needs to be. And once that starts happening across a large enterprise system, it stops being just an engineering problem. It starts affecting how quickly the business can actually move. In our latest article, we look at why this happens, what technical debt really costs, and why fixing it does not always mean rebuilding everything from scratch. Read the full article on our website 👇 https://lnkd.in/dhdk3c7E #Unibrix #EnterpriseSoftware #SoftwareDevelopment