Danny Singer

Breezy! has introduced #Broker, a new platform giving industry buyers direct access to verified margin VAT stock with full transparency and trust. Following the likes of Dipli, NorthLadder and CompaRecycle. Breezy! has launched #Broker, a multi-buyer platform designed to strengthen transparency and reliability across the secondary tech market. The platform offers direct access to verified margin VAT stock sourced entirely through Breezy!’s omnichannel trade-in ecosystem, which collaborates with major partners across Europe. With its first registration wave open until the end of 2025, #Broker is positioned to support companies looking for consistent, high quality supply in a competitive market. The platform makes use of Breezy!’s extensive trade-in infrastructure, including automated production facilities and detailed device evaluation processes. Broker allows users to observe real-time price dynamics, enabling faster and more informed decisions. This transparent bidding environment operates under a single rule set for all participants, including Breezy’s own retail division. For buyers seeking financial flexibility, Breezy! provides credit lines that cover the full sales cycle. Devices can be shipped on credit, giving partners greater control over cash flow and purchasing schedules. The platform also incorporates double grade verification, ensuring each device is evaluated during the trade-in stage and re-inspected at Breezy’s own facility before an order is finalised. With more than 14,000 accepted device models, including Apple iPhone models, laptops, smartwatches and gaming consoles, Broker serves marketplaces, e-commerce companies, retailers and mobile operators. Breezy! encourages early registration for companies seeking reliable access to margin VAT stock as the next wave of onboarding begins in April 2026. Great move Breezy!, Andrii Kosar and team. #SecondaryMarket #RefurbishedTech #CircularEconomy

23

8 Comments

Attention is All Your Clients Need. We, at Exthalpy are excited to announce that our paper is now live on SSRN discussing and providing possible resolutions for how modern agentic #AI and video agents can be the market disrupters in client success and customer relations for businesses. Our goal was simple - map leading tech and order them with the magnitude of NPS increase and ROI increment. Read our paper here -> https://lnkd.in/dRzR_ypr #Research #AI #Engineering #ClientSuccess #Support #Aftersales

22

2 Comments

The two big types of agents you should think about having for your business Stateful vs stateless agents I'm sure you are seeing dozens of "you need this lead magnet generator agent" or "content agent" or "linkedin dm agent". All useful when done right. I am talking about the fundamentals here. The better you understand first principle thinking for this the better you'll be able to deploy agentic leverage. In software when you have an app or piece of code that "retains state" it keeps track of what's happened, how and what next logic steps are open. This is really big for frontend apps for example, since it needs to keep track of how you interact with it (like your email app knowing what emails you've opened, replied to, and what drafts you've written) For agents this really means do they remember what you've told them over all your queries and runs, not just that workflow execution. We really haven't see large scale adoption of stateful agents, for many reasons, but I'm betting they'll become foundational load bearing context nodes in everyone's intelligence layer (all your context sources, tools, and agents to connect them together). You'll still want stateless agents to execute data ingestion and stuff that can and should be isolated, but stateful agents offer a whole new paradigm. An example I'm working on now: fireflies transcript --> triggers stateless agent --> stateless agent summarizes, classifies (if this is a lead call generate a SoW, if I'm just jamming with a work friend summarize the topics we talked about, etc.) stateless agent sends instructions w/basics---> stateful agent that knows what I care about and what I've fed into it for the past week generates the actual doc and connects the dots across my whole tech stack and context sources (literally I want it to be able to check through my linear and @ things that would be so cool) The third advanced type is a Stateful workflow. Stateful workflows theoretically have the best of both: you can run them concurrently (like Claygent) but over each run it stores what it learned in its memory (either immediate memory blocks or long term archival) These will be gigantic value drivers when pointed at the right problem, but I think it will extremely difficult to figure out what problems they're best suited for, how and why. If this was helpful to you I'd love to know why and what sort of problems you think you'll solve using both stateless and stateful agents have a great weekend all! P.S. Diontay Santiago what a day 🫡

18

2 Comments

“If I have 50 agents, how many can your AI replace?” It’s the most common question we hear. It sounds strategic: cost-focused, ROI-driven, but it doesn’t help us see the full picture. Let’s say your contact centre gets 200,000 calls a month. You only have the capacity to handle 150,000. That leaves 50,000 patients unheard, unanswered, or dropped altogether. Now you bring in AI. Don't be in a rush to "replace" your agents yet. It picks up the slack—the 50,000 calls you couldn’t get to before. Suddenly, more patients are getting help. But internally, your team still feels the same level of workload. Nothing has “freed up” because these customers were not getting help before to begin with. And once patients realize they no longer have to wait 40 minutes to speak to someone, more of them start engaging. The number of calls might go up. Not because AI failed, but because you just made access easier. That’s demand the statistics didn't tell your team about. The goal should not only be to shrink headcount. It should be expanding service capacity, lifting quality, and enabling your team to focus on high-value cases that actually need a human intervention. But all of that only works if you look at the system holistically, from operations to experience to long-term sustainability.

59

12 Comments

For years, I’ve kept a spreadsheet of new deals done by UK businesses in contact centre outsourcing. I use it so that I can understand what’s happening in the industry. I’ve never analysed it - with the exception of occasionally thinking “why are they using them?” However, I was looking at the list this morning from the past 6 months as the same things are appearing in different rows on the same column and so here’s some basic analysis. (1) almost every company offshoring for the first time is going to Africa (mostly South Africa with some interesting work going to Egypt (2) a number of companies who have been outsourcing to South Africa have moved to other locations (especially to India & Philippines). This is all in sharp contrast to previous years where there has never been such clear trends on my spreadsheet. Clearly I don’t have details of the whole of the market but I thought it was interesting.

134

39 Comments

Agents are not next. They are now. SnapLogic just showed that nine out of ten tech leaders already tap GenAI tools and half have real agents running today. The average firm runs thirty-two agents and plans thirty more in the next year. Budgets cross two million and trust in the results beats human work for many tasks. That means one thing for builders like me: the window is open but the rules are set. I build every agent in n8n, a no-code canvas that lets one workflow talk to another at light speed. Drag, drop, ship. Yet the fundamentals from the report remain the same. Start with a single use case that saves real time. IT help-desk tickets, data entry, sales research, pick one pain. Keep the loop tight. Input, logic, output, human check. No mystery math. Protect the data first. SnapLogic shows security and privacy top the worry list. In n8n that means secrets live in environment variables, webhooks stay behind auth, and logs rotate fast. Train the team early. The study says skill gaps stall twenty-nine percent of projects. Most companies in the survey expect payback inside eighteen months because they track the numbers. Do the same. Expand only when the first agent works in the wild. SnapLogic firms plan to double their fleet but they are adding, not guessing. Clone proven flows, swap credentials, review edge cases, then launch. Keep a Center of Excellence even if you are a team of two. Document prompts, share test cases, store reusable snippets. The goal is speed with safety. n8n makes the build simple. The discipline makes the difference. Agents can cut nineteen staff hours a week, slash outage calls by twenty percent, and flip research days into minutes, but only if we follow the playbook. The frontier is wide open. The rules are clear. Build fast, guard data, count the wins, and your agents will not just run. They will run the business.

53

33 Comments

Israeli startup Second Nature has raised $22 million in a Series B round to expand its AI sales training platform. Second Nature uses conversational AI to help companies train and coach their sales and service teams. The platform creates realistic, data-driven roleplays with virtual avatars that can mimic customer moods, objections, and complex scenarios. After each session, users receive personalized feedback and performance scores, enabling sales increases of over 20% and sharply reduced onboarding times. The round was led by Sienna Venture Capital, with participation from Bright Pixel Capital, StageOne Ventures, and customer-investor Zoom. This brings the company's total capital raised to $80 million. The new funding will be used to further develop the platform and scale operations as it transforms how companies coach their teams. Congratulations to co-founders Ariel Hitron, Alon Shalita, and the Second Nature team. Calcalistech: https://lnkd.in/d7pxikNg Deals like this are reshaping the AI landscape. For weekly analysis on the funding, valuations, and strategic moves that matter, subscribe to the Charted Growth newsletter: chartedgrowth.beehiiv.com #AI #SalesTech #ConversationalAI #SalesEnablement #CorporateTraining #EdTech #Funding #SeriesB

16

5 Comments

A powerful thought worth repeating... if agents are going to be 10x cheaper in 12 months... we don't expect 10x more volume, we expect more. As costs drop new capabilities come online. With longer agent chains they can do more work. Things that are not considered today become practical. Ideas that seem wild or crazy seem normal. This concept means engineering leaders who are championing AI-first companies should be making decisions today, not for what the systems can do today, but what they will be doing in the next year.

29

9 Comments

𝐖𝐢𝐭𝐡 𝐠𝐫𝐞𝐚𝐭 𝐩𝐨𝐰𝐞𝐫 𝐜𝐨𝐦𝐞𝐬 𝐠𝐫𝐞𝐚𝐭 𝐫𝐞𝐬𝐩𝐨𝐧𝐬𝐢𝐛𝐢𝐥𝐢𝐭𝐲 That quote isn’t about superheroes anymore. It’s about AI agents. Meet 𝐌𝐨𝐥𝐭𝐛𝐨𝐭 🤖 Launched in November 2025 and already downloaded 600,000+ times. A new kind of AI that doesn’t just answer questions, it takes action. Most AI tools today work like this: 𝐘𝐨𝐮 𝐚𝐬𝐤 → 𝐢𝐭 𝐫𝐞𝐬𝐩𝐨𝐧𝐝𝐬 → 𝐲𝐨𝐮 𝐜𝐥𝐢𝐜𝐤. In Moltbot : 𝐘𝐨𝐮 𝐬𝐞𝐭 𝐚 𝐠𝐨𝐚𝐥 → 𝐢𝐭 𝐩𝐥𝐚𝐧𝐬 → 𝐢𝐭 𝐞𝐱𝐞𝐜𝐮𝐭𝐞𝐬 → 𝐢𝐭 𝐤𝐞𝐞𝐩𝐬 𝐠𝐨𝐢𝐧𝐠 𝐧𝐨𝐧 𝐬𝐭𝐨𝐩 𝐰𝐢𝐭𝐡𝐨𝐮𝐭 𝐇𝐮𝐦𝐚𝐧-𝐢𝐧-𝐋𝐨𝐨𝐩 1. Reading emails 3. Interacting with apps 5. Managing files 7. Operating browsers 9. Acting without constant human approval That’s not just AI. That’s delegated autonomy. Which explains why it went from “new project” to half a million downloads in a few weeks. If used responsibly gives huge productivity gains if used carelessly then its unexpected (sometimes risky) outcomes Important nuance (before we panic 😄): • On iPhones, this kind of “rogue bot” behavior isn’t really possible (heavy sandboxing). • On computers, permissions and guardrails matter a lot. My takeaway: AI isn’t becoming scary because it’s “thinking.” It’s becoming powerful because it’s doing. So the real question isn’t : “Is AI smart enough?” It’s: 𝐇𝐨𝐰 𝐦𝐮𝐜𝐡 𝐜𝐨𝐧𝐭𝐫𝐨𝐥 𝐚𝐫𝐞 𝐰𝐞 𝐰𝐢𝐥𝐥𝐢𝐧𝐠 𝐭𝐨 𝐠𝐢𝐯𝐞 𝐢𝐭? Would you trust an AI to act without your approval : yes or no? Drop your take in the comments. #AI #AIAgents #ArtificialIntelligence #Automation #FutureOfWork #TechTrends 

42

🚀 Small Business Owners: Want to Save Time & Cut Costs with AI? 🚀 AI isn’t just for big corporations—it’s transforming small UK businesses right now! From automating tedious admin tasks to streamlining workflows and enhancing customer service, AI can be your secret weapon for efficiency and growth. 🎯 ChatGPT – Craft compelling content and handle customer queries effortlessly 🎯 Zapier & Make – Automate workflows and connect your business apps 🎯 Midjourney & Suno – Create professional visuals and music for marketing 🎯 Claude & Gemini – Analyse data and get instant insights for better decisions Curious about how AI can work for you? I’ve broken it down in my latest blog post! ⬇️ 👉 https://lnkd.in/eqCxQEGH 💬 Have you tried any AI tools in your business yet? Let’s discuss in the comments! 👇 #AIforBusiness #SmallBusinessGrowth #Automation #Productivity #MacawVA

12

6 Comments

Here's why you should always have your Voice Agents state that its an AI Agent at the start of a call I've been guilty of this in the past and I see other people trying to hide the fact it's AI Which is a bad idea and I'll tell you why: It's legally required (in some places) 🚔 In the US, outbound AI calls have to disclose they're AI. It's the law. And the rest of the world is following suite and for good reason to stop the scammers. Better to get ahead of it now than scramble later. It sets expectations ✅ When someone knows they're talking to an AI, they're way more forgiving if it mishears something or asks them to repeat themselves. If they think it's a human and it fumbles, they get frustrated and hang up. Transparency = patience. You can actually use it to your advantage 🤖 I prompt my agents to say something like: "Just so you know, I'm an AI assistant so if I mishear something or get it wrong, feel free to correct me" It's disarming. And it gives you leeway if the agent isn't perfect (which, let's be honest, they never are 100% of the time). I also, depending on the client, throw in a bit of humour when I say its an AI Agent which disarms people and sometimes even gets a laugh 😆 It builds trust 🤝 Trying to trick people into thinking they're talking to a human is weird. And when they figure it out (which they usually do), you've lost trust before you've even started. Just be upfront It works better

67

20 Comments