Brandon Redlinger

I still remember the first time I presented to a board. I spent the entire week convincing myself I wasn't ready. I rewrote the deck every night, practiced every answer I thought they might throw at me, and second-guessed every slide. The morning of the meeting, I sat in my car thinking, "What if they figure out I'm not supposed to be here?" I don't even remember the presentation itself. What I remember is waiting for my confidence to magically appear before I walked into the room. But it never did. The questions came, and some I answered well, but a few caught me off guard. Nobody laughed or called me a fraud or anything like that. The meeting ended, and I had survived. The next board meeting felt a little easier, and so did the next one. Presenting to executives started to feel normal, then leading bigger teams, then speaking on stage. It took me years to realize I'd had been thinking about confidence completely backward. I always thought it was something successful people had before they did hard things. In my experience, it shows up after you've done the hard thing enough times to know you'll survive it. I talk to marketers all the time who tell me they're waiting until they feel ready – ready to lead a team, ask for the promotion, present to the board, become a CMO. I don't think that feeling ever shows up on its own. You earn confidence one hard conversation, one presentation at a time. Even now, I still get nervous before the biggest meetings, but the difference is I've collected enough evidence to know I'll figure it out. That's what confidence is for me now: proof I've done hard things before, and I can do them again.

They say outbound is dead, but those are the people who are not using signal-based selling or using it wrong. For most teams, a signal comes in, then a task in Salesforce gets created or Slack message gets dropped in a channel. This is as far as the automation goes and both get ignored, then teams get mad because they’re getting < 1% response rates. I spent last week building the full automation and outreach system that’s getting 7.7% response rates. Here’s what I built: 𝟏. 𝐒𝐢𝐠𝐧𝐚𝐥𝐬 𝐚𝐧𝐝 𝐫𝐮𝐥𝐞𝐬 𝐟𝐨𝐫 𝐰𝐡𝐢𝐜𝐡 𝐬𝐢𝐠𝐧𝐚𝐥𝐬 𝐞𝐚𝐫𝐧 𝐚𝐜𝐭𝐢𝐨𝐧 A competitor's name disappearing from a target account's subprocessor page is worth a call within the hour. But a whitepaper download from a student is worth nothing. That first one is my favorite signal in B2B right now. Any company selling into the EU has to disclose which vendors touch customer data, and most publish it on a page they're obligated to keep current. When a name comes off that list, the account churned from that vendor and announced it themselves. Almost nobody is watching those pages. Rank your signals by that logic, or your reps stop trusting all of them inside a month. 𝟐. 𝐓𝐡𝐞 𝐫𝐞𝐬𝐞𝐚𝐫𝐜𝐡 𝐡𝐚𝐩𝐩𝐞𝐧𝐬 𝐛𝐞𝐟𝐨𝐫𝐞 𝐭𝐡𝐞 𝐒𝐃𝐑 𝐢𝐬 𝐧𝐨𝐭𝐢𝐟𝐢𝐞𝐝. The moment a signal triggers, the account gets researched. What changed, who to contact, what they're dealing with, and whether sales is already working them. That's 10-20 minutes of prep that used to happen only after someone decided the signal was worth chasing. 𝟑. 𝐓𝐡𝐞 𝐨𝐮𝐭𝐩𝐮𝐭 𝐢𝐬 𝐚 𝐝𝐫𝐚𝐟𝐭. Previously, it was just a notification but now we’re taking it a step further and saving them more time. The signal should produce a written opener, a call script, a reply that's ready to send. For the reps, it’s just light edit and go. 𝟒. 𝐒𝐨𝐦𝐞𝐭𝐡𝐢𝐧𝐠 𝐡𝐚𝐬 𝐭𝐨 𝐥𝐞𝐚𝐫𝐧 𝐟𝐫𝐨𝐦 𝐰𝐡𝐚𝐭 𝐰𝐨𝐫𝐤𝐞𝐝. This is the eval loop. Every week the results get def back into the system. Which signals turned into conversations, which drafts got replies, which ones didn’t, etc. The next batch gets built on that. Skip this and you have automation that never gets smarter, which is faster mediocrity. For the SDR, now the morning starts with a ranked list of who to contact and why, the research already done, the first message already drafted, and the CRM already updated. The day goes to conversations instead of assembling context. Signal detection got commoditized two years ago. Everyone can see the same signals at the same time now. And now once again speed to lead + relevant (non-AI slop) outreach wins.

I almost talked myself out of hiring one of the best marketers I've ever worked with. On paper, she wasn’t the strongest candidate. Another person had the bigger brands on their résumé, more years of experience, and a title everyone on the panel recognized. The decision felt obvious, but I just couldn't shake one thing from the interviews. One candidate kept talking about what they'd already done. The other kept talking about what they wanted to learn. Every answer came back to curiosity. She asked thoughtful questions, she wanted feedback, and when she didn't know something they just said so instead of bluffing through it. So I took a chance on her. The first few months weren't smooth. She asked a lot of questions, needed coaching, and made mistakes. But every week she got a little better, because she treated every project as a chance to improve. A year later she was leading work I wouldn't have trusted her with in month one. Eventually she became one of the people I relied on most. I've thought about that decision a lot since. Experience matters, and skills matter, but a strong résumé only tells you where someone has been, and that's not the same as where they're going. Some of the best marketers I've ever worked with had exactly that kind of hunger and determination and drive. It’s that kind that keeps you learning long after everyone else decides they've got it figured out. I still look for experience when I hire, but I spend even more time looking for curiosity and drive. You can teach someone a new channel, your messaging, your process, etc. It's a lot harder to teach someone to stay curious and driven after they think they've made it. And I guess I have a lot of empathy for this candidate profile because someone took a chance on me early in my career, and I'm grateful he did. I've tried to pay it back by hiring for potential, even when it doesn't fit neatly on a résumé.

They just admitted that the hardest part of enterprise AI is implementation and adoption. Microsoft is committing $2.5 billion to forward-deployed engineers and embedding 6,000 of them inside its customers. Amazon made a similar investment a few days earlier. These are obviously 2 of the biggest companies on earth, and they’re basically saying that the hardest part of enterprise AI is implementation and adoption, and I think they're right. And they're not the only ones. This conversation keeps coming up with enterprise businesses. Everyone likes to cite MIT's NANDA report, which says that 95% of enterprise AI pilots show no measurable P&L impact. And I used to complain about the methodology and question the number itself, but I think it’s directionally right – everyone is running pilots, while very few have much in production. And for B2B marketers, that changes what we’re selling. A big portion of product marketing budgets I see still go to fancy demos, product storytelling, demo environments, demo data, etc. However, the part of the sales cycle that I see slows enterprise AI deals down is when the buyer asks who is going to make this thing work inside their environment. They don’t double the product because they see the power of AI, but they want to know when it will be live inside their environment and who would make it happen. Harvey figured this out early. Every deployment has a forward-deployed engineer and a lawyer. In other words, every account has a tech expert and a domain expert during implementation. I run fractional marketing at a few enterprise AI companies (and I’ve had a few AI clients in the past), and the enterprise AI companies that are seeing the most success are moving GTM budget into 2 things: 1 – Implementation case studies with real timelines, because if you can get them live 4-6 weeks, it beats any feature page or demo. 2 – The deployment team in the content, with names and faces. Right now they're a line item that probably nobody even sees. If you want to go further, publish a time-to-value number that you can deliver on. So, here are some things to think about if you're the one running marketing at an enterprise AI org: – Review your budget allocation. If demo assets are disproportionately more than deployment proof (at most companies I've looked at, they are), think about moving some of it. – Book time with your implementation team and find the deployments the customer bragged about. Use the story, with the real timeline and the number it moved, in your marketing. – Establish a time-to-value metric if you don’t have one. How are you thinking about this?

Most marketers I talk to are still not using agents because they think agents are for developers, but I'm using agents daily to help me 10x my work. The good news is they're pretty easy to learn and understand. Here are 10 agent features with some notes on how I'm using them. 𝟏. 𝐁𝐚𝐜𝐤𝐠𝐫𝐨𝐮𝐧𝐝 𝐬𝐮𝐛𝐚𝐠𝐞𝐧𝐭𝐬 Spawn smaller Claudes in parallel, each with its own memory. Example: My client onboarding uses 5 subagents at once, and now two weeks of research is done in one hour. 𝟐. 𝐓𝐡𝐞 𝐛𝐮𝐢𝐥𝐭-𝐢𝐧 𝐬𝐩𝐞𝐜𝐢𝐚𝐥𝐢𝐬𝐭𝐬 Explore and Plan ship with the tool, and almost nobody knows they exist. Example: I point an Explore agent at 200 interview transcripts and get the top 5 objections with quotes. A Plan agent maps the approach before anything runs. 𝟑. 𝐂𝐮𝐬𝐭𝐨𝐦 𝐬𝐮𝐛𝐚𝐠𝐞𝐧𝐭𝐬 Define your own specialist agent in one file. Example: I have one that runs a brand-voice check on everything that it writes. 𝟒. 𝐌𝐨𝐝𝐞𝐥 𝐫𝐨𝐮𝐭𝐢𝐧𝐠 Assign each agent its own model, Opus down to Haiku. Example: Use Opus for strategy and narrative, Haiku for pulling first names out of a CSV. 𝟓. 𝐌𝐂𝐏-𝐜𝐨𝐧𝐧𝐞𝐜𝐭𝐞𝐝 𝐚𝐠𝐞𝐧𝐭𝐬 Give each agent its own toolkit and nothing more. Example: I have an agent connected to HubSpot only, another to Slack only. 𝟔. 𝐀𝐠𝐞𝐧𝐭 𝐭𝐞𝐚𝐦𝐬 Multiple agents on one project, handing work to each other. Example: I have a research agent, a copywriter, and an editor in sequence. It's the closest thing to a marketing pod that doesn't need headcount. 𝟕. 𝐌𝐮𝐥𝐭𝐢-𝐚𝐠𝐞𝐧𝐭 𝐫𝐞𝐯𝐢𝐞𝐰 Several agents review the same asset at once, each from a different angle. Example: 4 agent reviewers analyze a web page together: CRO, SEO, voice, positioning. 𝟖. 𝐖𝐨𝐫𝐤𝐭𝐫𝐞𝐞𝐬 Isolated copies of your project, side by side. Example: I tested 3 homepage rewrites in parallel, kept the winner, and deleted the rest. Nothing touched the main version. 𝟗. 𝐀𝐠𝐞𝐧𝐭 𝐩𝐞𝐫𝐬𝐢𝐬𝐭𝐞𝐧𝐜𝐞 Continue an agent instead of starting cold. Example: I ran one across a full week, sharpening the ICP each morning, and it held the whole thread. 𝟏𝟎. 𝐒𝐜𝐡𝐞𝐝𝐮𝐥𝐞𝐝 𝐚𝐠𝐞𝐧𝐭𝐬 Agents on a schedule, no code. Example: A Monday content-calendar audit that flags what's missing for the next two weeks and posts it to Slack before you're awake.