Adam Horner

Agentic or hyperfunctional, I learned a lot of stuff last week on LinkedIn. I read a post about Agentic AI - it's software that makes decisions for you, that's not Solo. Solo looks at all of your tasks at every point in the day and tells you which is the most optimum to perform. Solo's algorithm is elminating this decision for you but it's a flexible system (and not a sorter). So I'm thinking what we've built is more hyperfunctional SaaS, you never have to decide what to do with Solo. You just have to sell. More to come, we're adding some important things this week. #solo

8

2 kommentarer

💃🕺💃 WE HAVE GOT A NEW PRODUCT 💃🕺💃 Introducing Ractivi Converse, an AI Conversational Builder Platform. It is an AI-Native reimagining of what learning/communications could be from the ground up, where you can: 🌴 Build a ‘Conversation Tree’ to easily control what you want the AI to say and show, and we will orchestrate multiple AIs in the background to get it done 🎨 Create Audio-Visual experiences for employees, where instead of walls of text, the AI draws concepts/diagrams and talks through them like a tutor on a whiteboard 📝 Explain questions to users visually, by having the AI fetch media/explanations and annotate them with circles, arrows, underlines, and other visual cues 💖 Use Multimodal-AI to explain with empathy, where the AI intuitively understands emotional cues of individual users' tone/words and adapts its teaching towards it 💌 Get granular feedback to iterate your training, where instead of waiting for users to write feedback, the AI anonymously explains how it believes users reacted to each part 🧩 Match your employees' unique learning needs, with AI that tracks and records what employee-specific teaching styles worked in memory to improve itself 🧑‍🏫 Use AI to underline learning intent in media. For example, you can click an element in a video, and the AI will understand it, navigate to it, and explain it when instructing. OpenAI’s GPT-4o and Google’s Project Astra have demonstrated the revolutionary potential of solving descriptive learning with Conversational AI that can deeply listen to you. Employees can describe any problem and instantly have an infinitely patient tutor talk them through the challenge. We have never had a tool that can deeply listen. A tool that not only understands the intent of your words but also perceives nuances in tone, emotional and body language. It is a magical unlock in data, creating an underrated opportunity to improve learning by adapting not only what buttons you click but also to diverse reactions and interactions. But, there is another type of learning: Prescriptive Learning. Companies educate employees on a range of concepts, from cybersecurity to mentorship preparation. Often, it is a frozen, cookie-cutter, one-way experience. As companies scale, they also lose visibility of mission-critical topics like scaling their culture with their team. AI is unlocking a platform shift. Good managers are going to use AI to reduce costs. Great managers are going to use AI to maximise the potential of their people. We need a new platform dedicated to augmenting human potential with AI through learning.     This specific product was just me over the past 11 weeks. The product is still not ready for prime time yet, but I am excited to test, validate and build the rest in public. If you are keen to give it a try as we test, let me know in the comments.

28

7 kommentarer

Lots of interest in preventing API Drift which happens in ~75% (!) of APIs according to APIContext. But there's also a lot of bad information about how to fix it and prevent it from happening again. The good news is Bill Doerrfeld has the info you need 👇 Read this article if you are making APIs. Bottom-line, treat APIs as you would a UI App. Build or adopt a design system (API style guide), create wireframes & prototypes (API designs & mocks) and then get feedback from actual users on them to make sure you build the right thing. And then, use automated testing to make sure you continue to deliver on that intent. Iterate: - Build - Measure - Learn ↻ But use API designs and testing to know what you are building, measuring and learning. I'm thankful to Nordic APIs for taking my input on this very important topic.

9

If you are interested in the Getting Things Done personal organization method, check out my latest article. It explains how to start populating data and create a solid personal management system: https://lnkd.in/dNHBQfua

1

Better LLM models have larger context windows, which aids them in better synthesis. Similarly, better leaders also have a wider context window, allowing them to retain a lot of information in their working memory for improved analysis and decision-making on a variety of issues.

27

2 kommentarer

Have you been caught with your hands in the AI Cookie Jar? 🆕 The EU AI Act came into force today. If you are a coach, consultant or therapist using AI and are not in compliance with the Act, get your hands out of the cookie jar 🍪. We have seen the rising use of Technology augmented Coaching over the last 10 years, with a significant uptake since the rise of Generative AI systems like ChatGPT. There has been a rush to use technology to enhance coaching, however, the coaching community needs to push the pause button on this. Some time ago I emphasised that learning has for some time been enhanced through technology, as a reflection of me working for 7 years as a University Research Fellow in Technology Enhanced Learning (TEL). Towards the end of my tenure I realised this also needed to be engaged with in a way that flipped TEL, Learning Enhanced Technology, LET. Earlier this year I declared that 'Technology Enhanced Coaching', as I would come to call it, also needed to be coupled with what I would call 'Coaching Enhanced Technology' (CET). I also realised that for these approaches to work with one another there would need to be a bridging framework, 'Coaching Aligned Technology' (CAT). I discovered these three approaches fed into one another in a cycle. The cycle began, not surprisingly, with Coaching, that would be used as a reference to enhance and align technology features or utilisation, through Coaching Aligned competencies and methods which would result in Technology Enhanced Coaching. Following my Emerging Paradigm Model, Technology Enhanced Coaching, done well, would feed back into Coaching for facilitating higher levels of Coaching Enhanced Technology, driving a virtuous loop. So, if you are a coach adopting technology to enhance your coaching practice, I suggest you might need to rewind back two steps or take a radical step forward. You either will need to step from Technology Enhanced Coaching back to Coaching Aligned Technology and then back to Coaching Enhanced Technology, or you might need to radically flip what you are doing altogether and engage in Coaching Enhanced Technology where you focus on a Coaching-First Principles approach to utilising such a technology as AI. Either way, to prevent yourself being caught not using AI in ethical, inclusive and compliance based ways, you are going to need to enter into a cycle of development and deployment that follows a clear pathway of value, insight, empowerment, transparency, accountability, integrity and robustness to quarantine yourself and your clients from harm so you can progress with realising the benefits of technology in coaching sessions. The pathway must be one that you frequently clear and clarify, through iteratively refined, agile timelines. The AI Coaching Alliance pathway, coupled with the EU AI Act aligned AICA Standard gets you there. The EU has given you 6 months to prepare. Commit to Compliance at the next AICAS Briefing. https://lnkd.in/ejcVdcKt

4

"Stop creating architecture diagrams no one understands." In this interview with Björn Helgeson, VP of Engineering at Mindler, he discusses why shared vocabulary should come before technical documentation, and how it enables better business-engineering alignment. Is your architecture documentation building bridges or barriers between teams? Read Björn's interview, featured in the latest edition of Exec Engineering: https://lnkd.in/eh9SYcJh

9

I feel leadership reviews help bring coherence amongst different teams concerning to prioritization, remove blockers, and execute fast with a shared goal. It also gives the leadership team early feedback on what's working and what's not. However, 100% agree with the "getting the document ready" part and it should not require hours to prepare it. It should be lightweight and could just be an automated metrics dashboard?

9

What about working with Agile in truly enhanced way .. 👉 Bring Agile & AI-Powered Assistant together 👈 How Stark Assistant Fits with Agile Teams: ✨ Sprint Planning just done in seconds! Full tasks and resources allocation. Everyone will know WHAT to do and WHEN to do it. ✨Variation Requests offer the flexibility to replan in seconds when changes happen. ✨ Real-Time progress tracking, efficiency metrics and intelligent insights keep you on track and help you prevent risks before they happen. ✨ Integrated Time Tracking helps monitor productivity and manage workloads efficiently. ✨ Team members can stay connected, share insights, and resolve issues faster, leading to a more cohesive and productive team environment. ✨ Motivate the team with a points-based system that rewards task completion and encourages productivity. ✨ Conduct sprint retrospectives using Stark Assistant’s intelligent analysis with full transparency. Make data-driven decisions to enhance team performance. Stark Assistant: https://lnkd.in/dwkwNmFj Follow us: https://lnkd.in/dVkaChbp #Agile #AgileManagement #Stark #StarkAssistant #ProjectManagement #ProductManagement #OperationsManagement #AIForManagers #AIForLeaders #FutureOfWork

2

Structured and reusable data is one of the power so Timezynk AB. Allowing you to add and link everything you need for solving the problem. If all your project needs is a name and a number, no problem. Would you rather have 30 fields with everything related to that order, no problem. We even visualize your data making it easy to understand where it all belongs and how it is related to each other. And if multiple connections are made, each new value overrides the previous one, so you can have timecode from a task on the shift, but the user can report a different one on the report still. There is almost no limit to what a register may contain. You then choose what information is shown for your users when they are booked, so they have access to everything they need to do their job. www.timezynk.com [email protected] 040475473

4

This is how we use Clay at Growth Engine X and how I think you should be using it too. We use Clay as an operating system, not an all you can eat smorgasbord of data. I feel some dissappointment in myself for my team that I worked at Clay and with the content I post now anytime anyone asks me "well how good is the data on Clay?" I don't know. How good are the houses on AirBNB? Even on a call today, people that were trying to implement Clay were asking me how it fits in with Lusha and what they are doing with Lusha in their sales team. And I told them it fits in perfectly and they should keep using both. Bringing all of the data providers in one place with conditional logic is the game changer part of Clay, not the marketplace you get access to with Clay credits. The credits are great for sure because you can use data providers that could require annual contracts and you just want to use it once. But I tell everyone to bring their own API keys into Clay so that they can keep doing everything they normally do in their workflow. Just automate it and make it repeatable where possible.

56

15 kommentarer

Here's the written version of a talk I gave at the Nordic APIs Summit 2024 in Stockholm. If you're interested in learning how to stop building web services like Fabergé eggs read on below. 👇 The original link to the talk is here -> https://shorturl.at/LfVnt for those who prefer the written word. Cheers!

4

Implement Micro-Segmentation In 30 Days! Ok, you probably scoffed at the headling? Well, take out time and read the referenced article below and lets show you how you can achieve frictionless micro-segmentation in 30 days.. Article, this https://lnkd.in/dHG7KD8H #microsegmentation #zerotrust #securityoperations.

6

Timesaver in crawling through the latest www-news Have you ever asked yourself why I read the article after it turned out it was a waste of time? Since the new #MacOS #Sequoia Beta 15.1 update, I have extensively used the built-in GenAI feature. It helps summarize or extract critical points from text, not only from websites. By having a quick look at the results, I can make a better decision to read it or not - in this example, I decided to read it because the key points were quite convincing, and the topic: The "AI Scientist" from sakana.ai . And why is this little feature useful: - it saves a lot of time, and you can concentrate on essential tasks - it helps to speed-scan through all the news, e.g., #GenAI, which might be a massive amount of articles to read daily - it is built in, available when I need it, and I do not have to use additional tools And Remember: 🚀 Innovation is racing ahead exponentially—it's time to adapt quickly or be left behind. 🚀 #GenAI #AIControlledByYou #Innovation #TheFutureIsNow #NeverStopLearning

10

At the highest level, for me, the development team productivity equation has two components: BANG and BUCK. Improving productivity then starts with either of two questions - how can we increase the value of the software created, and how can we reduce its cost? When you exclude developers from discussions around value (e.g., strategy, customer/user interactions, product design), and then demand greater productivity, you're asking them to do that with one arm tied behind their back. (And then, of course, tie the other arm behind their back by giving them no meaningful autonomy.)

17

4 kommentarer

OpenAPI Moonwalk GenAI is the new API consumer. And this needs changes to the API specifications. OpenAPI is working on version4 codenamed "Moonwalk" which should become available anytime in next few months. While each release of OpenAPI has enhancements, Moonwalk has Semantics introduced which provide purpose, whether the consumer is a human or an AI. OAI understands the importance of the traction that AI has within the space and has decided to act on it by incorporating a design constraint for it within the new major version of OpenAPI. With generative AI’s ability to understand natural language, OpenAPI can help bring the power of APIs to non-developers with a level of accessibility never seen before. To fully realize this potential, API producers should decorate their mechanical descriptions of HTTP requests with details that convey the meaning and purpose of those API operations. This, in turn, helps both people and LLMs to achieve better results. This is a great post for anyone wanting to go deep into what's coming. https://lnkd.in/dV96hsvV Let us all get ready for Moonwalk!!! 😉

13

Discover how workflow automation can streamline your daily tasks, boost productivity, and improve efficiency. Learn about tools, benefits, and best practices in our comprehensive guide. https://lnkd.in/dHQYH-tf

1

We really doing the "guy pointing at the thing" Youtube thumbnails now huh 😬 Also if you are curious I shipped two videos today: 👉 Using Notion Sites vs HelpKit to create your Notion-powered help center: https://lnkd.in/dNwUJSFq 👉 How my good friend Easlo is using HelpKit for his help center 😊 https://lnkd.in/drysBAHM

8