What a week in Cleveland! Our fourth year sponsoring #MAICON was our best yet, mostly because the conversations got so much more real. If you lead a marketing or sales team, here's what we heard on the ground: → The pilot phase is over. Teams are already using AI across strategy, operations, content, analytics and customer experience. Now they're being asked to show it works at scale. → Scale breaks things. Ten people experimenting with AI is fine. Five hundred people using it with no shared rules is a brand, compliance and data problem. → Governance is what lets you move fast. The leaders scaling quickest built control in from day one, so they can say yes to more use cases across the board. That's what Carlo Ammendolia showed on stage. In a live demo, he walked through how WRITER turns everything your company knows about your brand, your products, your customers, and your rules into a universal context layer. Every agent reads from it and writes back to it, so the work stays on-brand and accurate no matter who, or what, is doing it. Thank you to every leader who stopped by and shared what's working (and what isn't). See you next year, MAICON! 👋

LVMH 🤝 WRITER Last week we spent a day with @LVMH at the historic @joinstationf in Paris, where we introduced the Agentic Atelier, a concept built for luxury and powered by WRITER. For an LVMH Maison, everything comes down to craft, heritage and voice. They’re the whole moat. But digital commerce now demands more content, localisation, personalisation and product enrichment, across more markets and languages, than any team can produce without diluting what makes the Maison distinctive. The Agentic Atelier is our answer. We encode each Maison’s brand, craft and tone of voice into a governed platform. Marketing and revenue teams then build agents that create work that sounds like the house, while IT keeps full control of access, observability and security. We loved seeing the LVMH team engage and lean into the concept, and we're so excited to work with their Maisons over the coming months. A huge thank you to Selma Delic, Maxime Biraud, Antoine Rizk, Olivier Le Garlantezec, and the entire Tech & Digital Partnerships Team and Advisory Board for the trust placed in WRITER. Here’s to building many more Ateliers together!! ❤️

Every week, there’s a new headline about an AI agent going rogue. As the fear grows, it’s worth separating two very different problems: an evil AI, and an autonomous worker with overly broad permissions and weak guardrails. Look past the headlines, and nearly every case is an agent doing exactly what it was asked, with no limits on how far it could go to get there. It had write access when read-only would have worked, or a “please check with me first” that lived in a prompt instead of in the system itself. That last part is where teams get caught out. Telling an agent what not to do is a request. Taking away its ability to do it is a control. If you work in marketing or sales, your agents may already be writing to prospects, updating the CRM, or touching ad spend. So for each one, do you know what it can change, and what it can do without asking? Benjamin Popper mapped the four levels of agent autonomy and the minimum controls each one needs before it touches a customer: writer.com/blog/rogue-ai-age…

WRITER scored the highest in capabilities in 2026 IDC MarketScape for Worldwide Agentic Marketing! 🏆 IDC defines agentic marketing platforms as software that doesn’t just inform the people doing the work, but acts on the work directly. We built WRITER for that reality, side by side with marketing and revenue teams at some of the world’s biggest brands. Their bar is high and specific, and meeting it started with two choices we made early: 1️⃣ 𝐖𝐞 𝐛𝐮𝐢𝐥𝐭 𝐨𝐮𝐫 𝐨𝐰𝐧 𝐦𝐨𝐝𝐞𝐥𝐬. We didn’t want to rent our intelligence from providers who don’t understand enterprise customers. Today, our Palmyra models rank near the top of independent benchmarks for translation and in-context learning, with better inference economics than frontier models. That means predictable AI costs for our customers, and data that stays theirs. 2️⃣ 𝐖𝐞 𝐛𝐮𝐢𝐥𝐭 𝐠𝐨𝐯𝐞𝐫𝐧𝐚𝐧𝐜𝐞 𝐢𝐧 𝐟𝐫𝐨𝐦 𝐝𝐚𝐲 𝐨𝐧𝐞. IDC highlighted our controls across the full agent lifecycle that keep every output on-brand and grounded in a company’s own data. That’s what buyers want: more than 60% rank transparency into business, IT and security risk above every other requirement. Marketers can move fast, and the guardrails hold. Thank you to IDC for the recognition, and to the team for delivering on the capabilities that shined through in the report ❤️ writer.com/blog/writer-score…

Everyone watches for the malicious MCP connector, one that's over-scoped and pointed at the wrong host. That's the obvious threat, and it gets the attention it deserves. There's a quieter risk, however, we see far more often, and it doesn't leak sensitive data. The connector is perfectly legitimate, but its tools just have bad descriptions, too many parameters, or vague instructions. Your agent doesn't get hacked so much as it confuses itself. It grabs the wrong tool, invents an argument, and burns through your token budget before you even get a chance to respond. From the user's side, both results look the same: "the agent did something I didn't want." But these are two distinct failure modes. One leaks data, while the other quietly erodes trust in your service. We think both deserve the same attention and discipline. Our MCP gateway was designed to deal with both. Every tool description gets scanned for prompt-injection poisoning before it's admitted to the registry. OAuth scopes get winnowed down by ~25% to fight context bloat. Tools get deterministic names so two connectors can expose "create_record" without forcing the agent to guess between meaningless suffixes, and plumbing stays hidden from the model so it can't get a required-but-irrelevant field wrong. A connector that passes every security check and then ships a tool called "do_thing" with a dozen unexplained parameters isn't an attack from a malicious actor. It's just a different way of losing control. @kapilduraphe breaks it all down right here: writer.com/engineering/mcp-g…

Last week, we spent half a day in London with @MarriottIntl's EMEA marketing team, rethinking how marketers can put AI to work. By the end, the team was building their own AI playbooks in minutes. Katie Cox, VP of EMEA Marketing at Marriott International, brought her team in for an Ideas Workshop with one goal: step back and think differently about how AI can transform marketing. Here's what we worked on together: → A prompt engineering framework that helps marketers articulate the goal, the context, and what success looks like → Building playbooks live, no IT ticket required → Brand compliance built directly into every playbook, so every output stays on brand We loved seeing the Marriott team get hands-on in the solution design sessions, mapping out their real workflows and spotting where AI could help most. Huge thanks to Katie Cox, Clara Bayona Morales and Sarah Dubash-Wisler for their partnership in bringing this workshop to life and driving adoption, and to the wider team for welcoming us and engaging so actively in the session. We can't wait to keep building momentum and developing AI champions across EMEA. ❤️

WRITER has been named the Best AI Platform at the 2026 @Digiday Technology Awards! 🏆 Generic AI creates generic companies. Most tools know nothing about your brand, your rules, or your team's hard-won decisions. They're built for individuals, so what one person knows rarely carries across the team. But no customer meets just one person. They move through dozens of teams and touchpoints. Winning depends on everyone moving together on the same shared foundation. We built WRITER for that reality. Our agents plan, execute, and scale multi-step work across the whole customer journey. Brand and compliance are built in. Context carries across sessions and teams. Supervision is granular at every step. Our work with @Qualcomm brings this win to life. Qualcomm's marketers protect more than 1,200 trademarks, their brand portfolio and their IP across 12+ product launches a quarter, meaning legal and brand compliance is an essential workflow. With WRITER, they’re able to tackle their biggest bottleneck head-on, reinforcing brand and legal rules at generation so compliance is baked in from the start. And because brand guidelines live inside every output, every handoff (from brief to creative to content) is already on-brand. That means Qualcomm marketers can speed up and scale entire processes while taking on more work from start to finish themselves. Together, we’re setting a new standard for brand excellence. Thank you to Digiday for the recognition and to Qualcomm for showing what agentic marketing looks like in practice. ❤️

The average knowledge worker switches apps 100+ times a day. We've all done the ritual: copy the question from Slack, open a new tab, paste it in, re-explain the backstory, then copy the answer back into the thread. It's exhausting just to type. 😮‍💨 Until now! WRITER now lives inside @SlackHQ and @Microsoft Teams. Add it to a channel and it just… knows the conversation. Your playbooks, your brand voice, your connected apps, all of it. One message and it's off doing the actual work, landing the result right back in the thread where your team can see it, reply to it, build on it. No new tab or re-explaining needed. Not even a "let me pull that up." It's on your phone too, so a quick voice note on the way home means the follow-up is already drafted by the time you're back at your desk. The easiest place to start real work is the app you already have open.

WRITER + Snowflake = your data, straight into campaigns 💪 Our native Snowflake connector, built on @Snowflake's MCP Server, pulls live governed data straight into WRITER, so segments, LTV scores, and churn signals land in every marketer's hands, not just the analysts. From there you generate the brief, the assets, the landing pages, and the copy, all on-brand with voice, approved terms, and compliance baked in. Once the campaign is live, results land back in Snowflake alongside the rest of your governed data. WRITER uses that full picture to spot what's converting, where spend is leaking, and who's fatiguing, then recommends the next move. A human approves, and WRITER puts it into action. More here: writer.com/product/snowflake…

"When agents handle more of the work, how do you make sure the work that's left is more human, not less?" That's the question Sara Brooks has been working through at BetterHelp, the world's largest online therapy platform, with 6 million people helped and more than 30,000 therapists. When your brand is built on human connection, marketing that feels automated just doesn't work. So her team built BetterHelp’s brand standards into every marketing workflow agent, making sure the work still sounds like them. Then they tested it. One campaign built with AI performed 15% better than the traditional version. Another came together in 45 minutes, instead of the three weeks it would have taken with an agency. On Oct 1, Sara joins our CMO @diego_lomanto to share how her team reorganized around outcomes instead of tasks, and what it took to get there: go.writer.com/betterhelp-web…

Every time a meeting ends and someone says "I'll send notes," a small part of you dies. Your team's best insights and decisions get created in a meeting  room, then get buried in notes no one ever opens again. WRITER Meet turns conversations into agentic action. Add it to Zoom, Microsoft Teams, or Google Meet and it transcribes and summarizes in real time, so you can stay present instead of typing. Then it goes further. Run a Playbook and WRITER Agent takes it from there: recaps land in the team channel, plans update, and the next steps get handled across WRITER and your connected apps, all within minutes.

Did you spot us around San Francisco last week? 👀 WRITER was proud to sponsor @Dreamforce 2026, and we showed up in a big way. From LED trucks cruising San Francisco to our story on screen at Moscone North, column wraps across campus, and a mango smoothie station keeping the community fueled, we brought WRITER to life all over the city! But the real highlight was the conversations. We loved hearing directly from the Dreamforce community about what's possible with enterprise AI, and it reminded us why we do this work. Thanks to everyone who stopped by. Until next year! ☁️

Sara Brooks, Chief Growth Officer at @betterhelp, didn't ask her team to just adopt AI. She built her own agent first. When adoption still lagged, she cleared 50% of her most curious team member's time and built an AI excellence function from scratch. What happened next: an A/B test that outperformed, a campaign built in 45 minutes instead of three weeks with an agency, a set of guardrails that keep every agent's output sounding like BetterHelp, not generic AI. Real people, real results, radically simpler and more human. This is what it looks like when a marketing team stops using AI to go faster and starts using it to work differently. BetterHelp already did it. And we're SO excited to help tell that story. On Oct 1, Sara sits down with our CMO @diego_lomanto to unpack the six-function framework behind it all. go.writer.com/betterhelp-web…

Your AI agent just wrote a launch email that could've come from any of your competitors. Word for word. Almost two years into agentic AI and teams are shipping 10x more content, but with the same generic output. We built 𝐄𝐧𝐭𝐞𝐫𝐩𝐫𝐢𝐬𝐞 𝐁𝐫𝐚𝐢𝐧 to fix that. It's the universal context layer that captures what makes your company distinct, and makes it usable by every agent, in real time. It gives your agents four things they've never had: 1️⃣ 𝐀𝐠𝐞𝐧𝐭 𝐌𝐞𝐦𝐨𝐫𝐲: Everything your teams learn as they work. Decisions, customer signals, the judgment calls that live in Slack threads. Captured and applied to every task. 2️⃣ 𝐁𝐫𝐚𝐧𝐝 𝐚𝐫𝐜𝐡𝐢𝐭𝐞𝐜𝐭𝐮𝐫𝐞: your messaging, voice, visual identity, colors, and logos, automatically enforced so every agent works from the full picture of how your brand looks and sounds. 3️⃣ 𝐂𝐨𝐧𝐭𝐞𝐧𝐭 𝐥𝐢𝐛𝐫𝐚𝐫𝐲: your team's best work, turned into context. Agents build on what's already approved instead of starting from scratch. 4️⃣ 𝐂𝐨𝐦𝐩𝐥𝐢𝐚𝐧𝐜𝐞 𝐥𝐨𝐠𝐢𝐜: your rules and legal guardrails, applied at creation rather than bolted on after. Approval cycles move faster. Every agent reads from Enterprise Brain before it produces anything. Every campaign writes back to it when it's done. The system gets smarter, but more importantly, the work stops being generic and starts being unmistakably yours.

Most memory systems are built to remember the personal things: people, places, events, the name of your dog, Biscuit. Enterprise work has a different shape. When someone asks "how does this team typically qualify a sales opportunity?" it doesn't matter what your favorite TV show is, or that you take Biscuit to the groomer on Wednesdays. This context has its own structure: a stage, an owner, a scoring system, an outreach plan. We built Agent Memory on an opinionated ontology shaped around enterprise work: tasks, requirements, deliverables, the projects they roll up into, the conventions a team settles into along the way. This means the key context isn't a pile of docs, spreadsheets, and slideshows every retrieval pass has to re-read. It has shape. We still remember Biscuit. She's in the picture, just not the focus. What matters is how your company works, and how AI can help you get work done. Full writeup on how we built it, the ontology, the write operations, retrieval, and team-level scoping, here: writer.com/engineering/build…

Most sales and marketing teams run on relay. Marketing updates the messaging, and it takes multiple emails and training sessions for Sales to hear about it. A rep learns something on a call, and it takes a Slack message, a meeting, or someone just remembering for that insight to ever reach marketing. Most of the time, it doesn't. With Enterprise Brain, that relay is gone. A field call surfaces a priority nobody saw coming. By that afternoon, marketing is already writing to it. No meeting. No handoff. Nobody had to remember to pass it along. Because now sales and marketing are pulling from the same source: a shared context layer that gives agents everything a company knows about its brand, customers, and go-to-market. The same context that keeps a rep's deck on-brand and compliant is what lets a marketer see field signals the moment they happen. writer.com/blog/enterprise-b…

Introducing Enterprise Brain: the shared GTM context layer, for an always-on customer journey that's distinctly yours. We rebuilt our platform to turn how your team works, its brand, its ways of working, the things only your company knows, into shared context for humans and agents across sales and marketing. With Enterprise Brain, agents stop producing generic work and start producing *your* work. And the system only gets smarter with every campaign and every deal, making a company running on WRITER harder and harder to copy. What's new: → 𝐀𝐠𝐞𝐧𝐭 𝐌𝐞𝐦𝐨𝐫𝐲: Transparent, team-level memory that learns your team's facts, decisions, and preferences, and puts them to use wherever they're needed. No more context dumps at the start of every session. → 𝐁𝐫𝐚𝐧𝐝 𝐚𝐫𝐜𝐡𝐢𝐭𝐞𝐜𝐭𝐮𝐫𝐞: Goes beyond tone and style to your messaging, content, visual identity, colors, logos, sub-brands, and sub-voices, so every output actually sounds like you. → 𝐀𝐩𝐩𝐫𝐨𝐯𝐞𝐝 𝐜𝐨𝐧𝐭𝐞𝐧𝐭: All your best decks, one-pagers, and more, ready to be personalized, so no one has to start from scratch or dig up the right file. → 𝐂𝐨𝐦𝐩𝐥𝐢𝐚𝐧𝐜𝐞 𝐥𝐨𝐠𝐢𝐜: Agents already know your legal guardrails by industry and region, so first drafts meet your standards from the start. PLUS: WRITER for @SlackHQ & @MicrosoftTeams turns team conversations into real work and shared context, right where you collaborate. And 𝐖𝐑𝐈𝐓𝐄𝐑 𝐌𝐞𝐞𝐭 turns meetings into shared context everyone can act on, not transcripts stuck in a folder. No custom engineering, no weeks of setup. Share a few work samples, upload your brand guidelines, point it at your website. Enterprise Brain takes it from there and keeps up as your brand evolves. The result: sales and marketing become one unified GTM function, shipping content aligned to the same messaging and standards, learning from each other in real time, and giving your customer one distinct experience at every touchpoint. writer.com/blog/enterprise-b…

This one's spicy 🌶️ We put an industry voice, an enterprise CMO, and an AI CEO in a room and told them to hash out the thorniest questions in agentic GTM. What came out as obvious was... agentic GTM isn't a tool problem. It's a systems problem. Bolting an agent onto a funnel that's already leaking doesn't fix the leak. @shellypalmer put it bluntly, “we're all in the same ocean, but no one's in the same boat.” @donnymac pushed back with where he thinks the industry is actually getting it right. And @may_habib called out the instinct CMOs keep leaning into, the one that's quietly holding marketing back. Three vantage points. One conversation. Full fireside at our product launch event on September 9: go.writer.com/september-prod…

With AI budgets ballooning, new ideas for managing cost are taking shape