AI Integration in Communication

We spend years learning how to code better. But very little time learning how to talk about what we built. For a long time, I thought — “If I build great things, people will automatically see the impact.” But in real life, that’s not how it works. If people can’t understand it, they can’t value it. Recently, I tried Prezent for a big presentation. Instead of long nights editing slides, the AI turned my technical work into something clear and easy to grasp. Everyone in the room could follow what I was saying. That changed the whole game for me. 👉 Great work matters. 👉 But clear communication makes it visible. As engineers, we often focus so much on building that we forget the power of explaining it well. Because impact isn’t just created — it’s also communicated. What’s been your biggest challenge while presenting your work to non-technical folks?

Got an email from a colleague I've known for three years. Drinks after conferences. Inside jokes. His daughter plays soccer. Subject line: Strategic Alignment for Q3. Flawless formatting. Perfect grammar. Professionally upbeat. Every bullet precisely spaced. I felt absolutely nothing. Closed it without responding. Here's what's actually happening: for decades, polish was proof of effort. A well-written message meant someone cared enough to craft it. AI severed that connection completely. Now a perfect email could be 30 minutes of real thought or 3 seconds of prompting, and the recipient cannot tell. So we don't trust any of it. Not dramatically. Not consciously. But in the slow, cumulative way that hollows out working relationships over time. Each frictionless message becomes a little harder to take seriously. Each exchange feels more like a transaction, less like a conversation. There's a concept in evolutionary biology called costly signaling. A peacock's tail is trusted precisely because it's expensive to grow. Cheap signals carry no weight. AI communication costs nearly zero to produce. The recipient, consciously or not, values it accordingly. And when everyone in an org uses the same tools, something stranger happens: the voices converge. AI is a probability engine. It gravitates toward average phrasing, standard structure, safest tone. Use it to smooth your communication and you're not saving time, you're deleting your own fingerprint. Before your next important message, ask one question: is there a single sentence here that could only have come from me? If no, the message might land. But it won't build anything. The polished email costs nothing to produce. That's precisely why it costs everything to trust. Link to the full essay in the comments below.

𝗢𝗻𝗲 𝗔𝗿𝗰𝗵𝗶𝘁𝗲𝗰𝘁𝘂𝗿𝗲 𝗗𝗶𝗮𝗴𝗿𝗮𝗺 𝗧𝗵𝗮𝘁 𝗘𝘅𝗽𝗹𝗮𝗶𝗻𝘀 𝗘𝘃𝗲𝗿𝘆 𝗔𝗜 𝗔𝗴𝗲𝗻𝘁 Every AI agent you've ever used follows the same pattern. ChatGPT. Claude. Copilot. Devin. Custom agents built with LangGraph or CrewAI. Even the autonomous multi-agent systems running inside enterprises right now. Strip away the branding and the frameworks and they all share one architecture loop: 𝗣𝗲𝗿𝗰𝗲𝗶𝘃𝗲 → 𝗥𝗲𝗺𝗲𝗺𝗯𝗲𝗿 → 𝗧𝗵𝗶𝗻𝗸 → 𝗣𝗹𝗮𝗻 → 𝗔𝗰𝘁 → 𝗢𝗯𝘀𝗲𝗿𝘃𝗲 → 𝗟𝗼𝗼𝗽 Here's what each layer actually does: → 𝗣𝗲𝗿𝗰𝗲𝗶𝘃𝗲 — The agent receives a trigger. A user message, an API call, a Slack notification, a sensor reading. Raw input gets converted into something the reasoning engine can process. → 𝗥𝗲𝗺𝗲𝗺𝗯𝗲𝗿 — Two types of memory working together. Short-term memory holds the current conversation and working state. Long-term memory stores learned patterns, past interactions, and retrieved knowledge from vector databases. → 𝗧𝗵𝗶𝗻𝗸 — The LLM at the center. It takes the input, pulls relevant memory, and reasons about what to do next. Chain-of-thought. ReAct. Plan-and-execute. The method varies but the function is the same — decide the next move. → 𝗣𝗹𝗮𝗻 — If the task can't be solved in one step, the agent breaks it into sub-tasks. Step 1 feeds into Step 2 feeds into Step 3. This is where simple chatbots end and real agents begin. → 𝗔𝗰𝘁 — The agent executes. It calls APIs, runs code, queries databases, sends messages, reads files — all through tool execution. In 2026, MCP (Model Context Protocol) is becoming the standard connector layer here. → 𝗢𝗯𝘀𝗲𝗿𝘃𝗲 — Every step gets traced. Logs, metrics, latency, cost, token usage. Without this layer, you're flying blind and debugging becomes guesswork. And running alongside everything → 𝗚𝘂𝗮𝗿𝗱𝗿𝗮𝗶𝗹𝘀 — permissions, approval gates, content filtering, human-in-the-loop checkpoints. The layer that keeps the agent from doing something you didn't authorize. Here's what matters about this diagram: The difference between a basic chatbot and a sophisticated autonomous agent is not the pattern. It's the depth of each layer. A simple chatbot has thin memory, no planning, no tools, and no observability. A production agent has vector-backed long-term memory, multi-step planning, 20+ tool integrations through MCP, and full trace observability. Same loop. Different depth. Once you understand this, you stop being overwhelmed by every new framework announcement. LangGraph, CrewAI, OpenAI Agents SDK, Google ADK — they're all implementing the same seven layers. They just make different trade-offs on which layers get the most engineering attention. The engineers who understand the pattern can pick up any framework in a weekend. The ones who only know the framework are stuck when the next one comes along. Which layer do you think is most underinvested in right now across the industry — and why?

Communications jobs are becoming more valuable because of AI. First it was The Wall Street Journal on storytelling. Now Business Insider and Amanda Hoover are making the same case in a recent article: the rise of AI has put a massive premium on strong communications professionals: ➡️ Anthropic has 3x’d the size of its comms team ➡️ OpenAI is paying 400K+ for comms roles ➡️ VC funds like Andreessen Horowitz have hired Erik Torenberg to lead the New Media team. AI slop is flooding the internet. But enterprises & brands need strong taste & judgement in their writing that AI doesn’t *quite* yet capture. For businesses - the ability to consistently produce high-quality copy is a competitive moat that builds connection with customers, amplifies the performance of owned and earned channels, and drives topline. It’s a counterintuitive insight (shouldn’t good writing be commoditized with GenAI?). Flooding the bottom of the market has put a premium at the top.

In the rapidly evolving landscape of AI, data has become a pivotal asset, offering unique opportunities for commercialisation and monetisation. The groundbreaking deal announced yesterday between OpenAI and Axel Springer underscores the emerging trend of data deals and licensing in the AI sector, heralding a new era of AI-driven journalism and content creation. This unprecedented partnership allows ChatGPT to utilise and summarise news stories from prominent media brands like Politico and Business Insider. This agreement is significant for several reasons. Firstly, it enables AI models to access and use high-quality, current content, enhancing their capability to provide relevant and timely information. Secondly, it sets a precedent for how AI companies can legally use copyrighted material, a concern that has become increasingly prominent as AI technology advances. The value of a company's data has never been more apparent. In the AI context, data is not just a resource; it's the lifeblood that powers these sophisticated algorithms. By entering into data licensing agreements, content creators can open new revenue streams, and help ameliorate the adverse impact of AI outputs competing with their own work. This shift is essential in an era where traditional revenue models, especially in journalism and media, are under strain. The deal between OpenAI and Axel Springer also brings into focus the legal nuances of AI and IP. As AI models are trained on vast datasets, including copyrighted material, questions around ownership and infringement become increasingly complex. This partnership shows a path forward where AI companies and content creators can mutually benefit while respecting IP rights. Looking beyond this specific deal, the concept of data licensing in AI opens a myriad of possibilities. For AI models to be effective, they need diverse, extensive, and current datasets. Data licensing agreements can ensure a steady supply of this crucial resource while providing a fair compensation model for content creators. The OpenAI-Axel Springer deal is a harbinger of the changing dynamics in the AI industry. It represents a shift towards a more collaborative, ethical, and legally compliant approach to AI development and deployment. As AI continues to integrate into various sectors, the value of data will only escalate, making data deals and licensing an essential aspect of the AI ecosystem. This partnership is not just a business deal; it's a blueprint for the future of AI, data management, and the potential symbiosis between technology and content creation. Data deals in the AI context are new. Organisations will need expert advice on how these will need to be drafted, taking into account representations and warranties, indemnities and liability carve outs which are specific to AI. The licence grants will need to be carefully considered and limited to ensure that the licensor’s interests are maintained. I wonder who could help with that…

Let me say something that may be uncomfortable for communicators: AI is not reshaping your tasks.    It is reshaping entire workflows — end to end. And the Comms function that doesn't reckon with that, fast, will not just lag. It will become irrelevant.   Here's what the data is telling us right now:   1️⃣ 91% of PR professionals now use generative AI in their workflows (Cision, 2026) 2️⃣ 75% of companies expected to invest in Agentic AI this year (Deloitte, 2026) 3️⃣ #1 most AI-impacted occupational group: media & communications workers (Microsoft Research)   That last stat is the one that should stop you mid-scroll.   Some may see this as the biggest threat to our profession. But shift your perspective just a little - and this is actually a great opportunity as communicators. 😉   When AI handles the heavy lifting of research, drafting, and real-time measurement, something becomes even more valuable: human judgment and human relationships.   The journalist who trusts your call. The executive you counsel in a crisis. The instinct to know when a message will land — and when it will backfire. No algorithm gets there.   And this is where I think we're failing the next generation: comms education still trains people to execute, not to advise. Senior presence comes with experience, but the advisor mindset has to start on day one. In an AI-powered world, a junior communicator who can't yet counsel will struggle to find their footing.   The Comms teams that will lead are those that use AI to move faster and smarter, while doubling down on the strategic counsel and relationships that only people can provide.   We are all at the beginning of this journey. The playbook is still being written. But one thing is clear: we need to move alongside the technology, not wait for it to pass us by.   #FutureOfComms #AITransformation #CorporateCommunications

Disney 🤝 OpenAI — why this matters beyond the headline It isn’t just about a tech partnership announcement. It’s about how the next generation of consumer touchpoints gets built, and who owns the emotional interface with fans. From a consumer + licensing POV, this deal is quietly transformative. 1️⃣ IP becomes interactive by default Disney’s strength has always been story + scale. OpenAI adds intelligence + responsiveness. That combination turns passive IP into: ⌙ Conversational characters ⌙ Adaptive storytelling experiences ⌙ Personalized worlds that respond to you, not just demographics This shifts fandom from watching to participating. 2️⃣ The next licensing frontier isn’t just products, it’s experiences Think beyond toys, apparel, or collectibles. AI unlocks licensable formats like: ⌙ Branded AI companions (characters that “live” with fans across platforms) ⌙ Smart play patterns for kids & families ⌙ Educational + entertainment hybrids tied to core franchises Licensing moves closer to software + services, not just physical goods. 3️⃣ Every consumer touchpoint becomes a brand moment Theme parks, streaming, retail, games, even customer service — AI allows Disney to: ⌙ Maintain character voice consistency at scale ⌙ Localize tone and storytelling by market in real time ⌙ Extend IP life cycles far beyond release windows This is IP always on, not campaign-based. 4️⃣ Data gravity shifts back to the IP owner In an AI-powered world, whoever controls: ⌙ Narrative rules ⌙ Character behavior ⌙ Ethical guardrails …controls the fan relationship. The Walt Disney Company partnering early here is about protecting brand trust as much as accelerating innovation. 💡The big takeaway: This isn’t Disney chasing AI hype. It’s Disney reinforcing that the future of entertainment is intelligent, personalized, and emotionally resonant, and that their IP is the safest, most scalable playground to build it in. Licensing teams should be paying close attention. So should anyone who thinks the next Mickey Mouse won’t talk back. #Media #Disney #OpenAI

A few weeks ago, while everyone was talking about the OpenAI & TBPN deal (including myself), HubSpot acquired Futurepedia, a network of 17 AI-focused YouTube channels. And now I'm realizing the two deals might be more closely connected than I originally thought. See, what makes this deal fascinating is the fact that HubSpot has been sponsoring these channels for three years and watching how well they convert for their products and services. Instead of continuing to rent the audience, they decided to buy it altogether. As Barbara Abraul from my team, brought up in our team Slack, this presents a really interesting development in branded content, especially when you compare it to the recent TBPN acquisition by OpenAI. Both deals show how brands are moving from traditional advertising relationships to outright ownership of media properties that work for them. As brands experiment with creator marketing, they inevitably find creators who are punching way above their weight class in terms of conversion and audience quality. At some point, it starts making more financial sense to acquire that media property rather than continue paying for sponsorships indefinitely. Compared to OpenAI dropping a fat bag to show-off, HubSpot's approach is particularly smart because they've built the infrastructure to actually sustain these investments long term. After acquiring The Hustle in 2021, they continued investing in niche media brands like Mindstream and Starter Story, plugging them into a broader media network where they can continue to grow their reach. This creates a fascinating dynamic where successful creator partnerships could naturally evolve into acquisition opportunities. Instead of just buying ads or sponsoring content, brands are starting to buy the entire media operation when the numbers make sense. This just goes to show how important it is for brands to actually be doing YouTube creator integrations in the first place. Otherwise, maybe they'll never know who they should be acquiring down the road 😏 #hubspot #creatoreconomy #acquisitions

Last week, I was in a meeting where a young analyst presented an incredible AI-generated report. Beautiful graphs. Accurate predictions. Everything was perfect. But when the CEO asked, “So… what does this mean for us?” the room went silent. The AI had all the data. But it couldn’t explain the insight, connect the dots, or guide a decision. And the analyst — brilliant, talented — struggled to articulate the message. In that moment, one thing became painfully clear: AI can give you information. Only communication gives you influence. The future of work won’t be led by the people who know the most. It will be led by the people who can: • make others understand • simplify the complicated • inspire action • manage emotions • resolve conflicts • bring clarity when chaos hits AI can write, code, design, analyse. But it cannot build trust. It cannot speak with empathy. And it definitely cannot lead humans. If you want to stay future-proof, invest in: ✔ communication ✔ storytelling ✔ executive presence ✔ emotional intelligence Because in a world full of AI-generated noise, a human who can communicate clearly will always win. #CommunicationSkills #FutureOfWork #Leadership #ExecutivePresence #Anecdote #SoftSkills #AI #CareerGrowth

A funny thing happens when humans interact with AI—they start talking like robots. You might think it’s ironic, given how much suspicion there is around AI taking over human jobs or changing the way we live. But the reality is, when faced with a piece of advanced technology, many of us revert to our most basic, stiff communication patterns, almost as if we’re trying to meet the machine on its own terms. Let me give you an example. Imagine asking AI to write five Instagram posts. You get results, but they're flat and uninspired. Now, imagine doing the same thing, but with a twist—introduce yourself, adding some personality, and interact as if you're speaking to a new team member. The difference is night and day. Why? Because you brought humanity into the conversation. If you've ever caught yourself sounding like a robot while chatting with AI, don't be too self-conscious. It's more common than you think—and it's been happening to professionals learning new skills long before AI came on the scene. I’ve spent fifteen years teaching innovation at Stanford, working with both students and professionals. One thing I’ve observed consistently is how learners—especially when they’re focused on achieving a specific outcome—tend to become robotic. Whether it’s during user interviews or while delivering a perfectly-scripted presentation, there’s a noticeable stiffness that creeps in. It’s almost as if the goal of “getting it right” takes precedence over being human. This isn’t just a problem limited to AI interactions. It’s something I’ve seen in countless learners. And it’s a behavior that can hinder creativity, collaboration, and, frankly, effective communication. So, here’s a simple piece of advice: if you're wondering what to say next, simply ask yourself, "What would a human being say?" And then say that. This simple shift can make a huge difference—not just in innovation work, but in your interactions with AI as well. During a recent Beyond the Prompt interview, Jenny Nicholson highlighted something that really struck a chord with me: bringing humanity into AI interactions isn’t just nice to have—it’s a requirement. She astutely pointed out that the only truly new element in any AI interaction is the particular human’s input—your experiences, thoughts, and personality. As Jenny put it, “It's helping people realize that (they have to bring their humanity to the conversation) is a non-negotiable. What is in the model is what is in the model... The only thing that is quote unquote new is anything that you bring.” This underscores the point that AI can only work with what you give it — after all, every interaction starts with the human prompt — and the more you bring your humanity into the mix, and continue to contribute your humanity to the mix, the more powerful and relevant the model’s ultimate output will be.