# Marketing Has Too Many Humans In The Loop > Source: > Published: 2026-07-23 10:00:00+00:00 [Open Source CEO by Bill Kerr](../)- Posts - Marketing Has Too Many Humans In The Loop # Marketing Has Too Many Humans In The Loop ## An interview with Doug Tallmadge, Co-Founder & CEO at Gradial. 🗝️ 👋 Howdy to the 2,318 new legends who joined since our last edition! You are now part of a 513,456-strong tribe outperforming the competition together. **LATEST POSTS** 📚 If you’re new, not yet a subscriber, or just plain missed it, here are some of our recent editions. 🛡️ [The Agentic CRM For AI‐Native Companies](https://www.opensourceceo.com/p/keith-peiris-interview). An interview with Keith Peiris, Co-Founder & CEO at Lightfield. 📡 [The Last Industry Software Hasn't Eaten](https://www.opensourceceo.com/p/hermann-frank-interview). An interview with Hermann Frank, Co-Founder & CEO at Gigs. 🐝 [Unfiltered: beehiiv’s Build Or Die](https://www.opensourceceo.com/p/unfiltered-beehiiv). Inside the platform that ships faster than anyone its size. **PARTNERS **💫 Framer is the pro website builder for creators, teams, and businesses that care enough to get every detail right. Now with [Agents that work alongside you](https://framer.link/Ar93B9H?utm_source=www.opensourceceo.com&utm_medium=referral&utm_campaign=marketing-has-too-many-humans-in-the-loop) across the full website workflow—all without leaving the tool where the real site lives. Agents bring speed and scale; [you bring taste, judgment, and control](https://framer.link/Ar93B9H?utm_source=www.opensourceceo.com&utm_medium=referral&utm_campaign=marketing-has-too-many-humans-in-the-loop). Build and iterate faster, and ship updates in minutes. Scale content and keep metadata consistent across pages. Your attack surface is bigger than you think. [Arctic Wolf](https://arcticwolf.com/resource/aw/state-of-the-cybersecurity-attack-surface?utm_source=newsletter&utm_medium=cpc&utm_campaign=ADV%20FY27%20CPC%20NA%20Open%20Source%20Newsletter) analyzed more than 800,000 IT assets and found that one in three was missing a critical security control, while 19% had reached end of life. Explore the hidden visibility, protection, and [patching gaps expanding cyber risk](https://arcticwolf.com/resource/aw/state-of-the-cybersecurity-attack-surface?utm_source=newsletter&utm_medium=cpc&utm_campaign=ADV%20FY27%20CPC%20NA%20Open%20Source%20Newsletter), plus practical steps security leaders can take to close them. *Interested in **sponsoring these emails?** See our *[partnership options here](https://www.passionfroot.me/open-source-ceo?utm_source=www.opensourceceo.com&utm_medium=referral&utm_campaign=marketing-has-too-many-humans-in-the-loop)*.* **HOUSEKEEPING** 📨 If you ever over‑prepare for a big meeting and still find yourself rambling or freezing, you’ll like this. I’ve been using [Ultraspeaking](https://ultraspeaking.com/app/?utm_medium=Creator&utm_source=newsletter&utm_campaign=Bill_Kerr%20&utm_content=Free_App) to improve my speaking (found it through [Lenny](https://www.lennysnewsletter.com/p/master-public-speaking-tristan-de-montebello?utm_source=www.opensourceceo.com&utm_medium=referral&utm_campaign=marketing-has-too-many-humans-in-the-loop)), and they’re running a free live class on July 27 where they'll live‑coach three volunteers on how to: explain complex ideas clearly (even on the fly), recover when you get off track, and prepare without sounding stiff or scripted. You can grab a [free spot inside their app here](https://ultraspeaking.com/app/?utm_medium=Creator&utm_source=newsletter&utm_campaign=Bill_Kerr%20&utm_content=Free_App). I love Ultraspeaking. I find it incredibly valuable, and always come away with a number of interesting insights. Anyway, enjoy today’s interview! **INTERVIEW 🎙️** # Doug Tallmadge, Co-Founder & CEO at Gradial [Doug Tallmadge](https://www.linkedin.com/in/doug-tallmadge-14166075/?utm_source=www.opensourceceo.com&utm_medium=referral&utm_campaign=marketing-has-too-many-humans-in-the-loop) is the Co-Founder & CEO of [Gradial](https://www.gradial.com/?utm_source=www.opensourceceo.com&utm_medium=referral&utm_campaign=marketing-has-too-many-humans-in-the-loop), a Seattle-based startup building AI agents for enterprise marketing operations. He co-founded the company alongside three fellow Dartmouth graduates: Anish Chadalavada, Deip Kumar, and Anup Chamrajnagar. Before Gradial, Doug worked at SpaceX as a software engineering manager. The company raised a $5.4M seed round in February 2024, a $13M Series A in 2025, a $35M Series B in December 2025, and a $65m Series C in June 2026—led by VMG Partners, with Madrona and Pruven Capital participating—bringing total funding to $55M at a $350M valuation. Customers include AWS, T-Mobile, and Prudential Financial. Doug. Doug went from writing software for rockets to building agents that update websites and route campaign approvals, and his argument is that the two are not as far apart as they look. The real story at Gradial is less about the AI and more about what happens when a company has to compress a normal fundraising timeline by two-thirds, grow headcount fourfold in a year, and simultaneously dog-food its own product on the problems that growth creates. # What is Gradial doing for enterprise marketing teams? Gradial is an AI agent that focuses on solving the operational problems of running a large enterprise marketing team. It's an agent that actually goes and does the work within the workflow: updating the website, sending emails, reviewing copy docs, doing QA analysis on different things you're doing. It’s an operational agent that jumps in and [works on all the bottlenecks within enterprise workflows](https://www.gradial.com/?utm_source=www.opensourceceo.com&utm_medium=referral&utm_campaign=marketing-has-too-many-humans-in-the-loop). Source: [Gradial](https://www.gradial.com/?utm_source=www.opensourceceo.com&utm_medium=referral&utm_campaign=marketing-has-too-many-humans-in-the-loop). # You raised your Series B eight months after you raised your Series A. Why so fast? The AI space has been moving so fast that the growth we've seen (and that many companies in the space have seen too) has really kind of forced us to take the typical company-building playbook and run it even faster than you normally would. We've been lucky to expand to a number of leading enterprises and really grow the team out here in Seattle and across the country to support that customer demand. # What was the hardest part of going from a SpaceX engineer to a founder? SpaceX likes to describe itself as a startup, and there are a lot of aspects that still are very startup-like about how it operates, even though it's no longer a startup and it has 15,000 or 20,000 employees at this point. I was lucky to be there in 2017 when it was a little smaller than it is today. The similarities are that you've got [this need to really move fast](https://www.gradial.com/?utm_source=www.opensourceceo.com&utm_medium=referral&utm_campaign=marketing-has-too-many-humans-in-the-loop) and try to be scrappy. At SpaceX, we would think about every dollar we were spending. We were very frugal in how we operated, even though it was already on a multi-billion-dollar scale. Where it's different is that when you're operating within a company like SpaceX, you take for granted the fact that you can just go work with a recruiter to build out your team, that you have all this infrastructure, like someone literally keeps the lights on in your office. As soon as you start a company, you realize how much additional work there is behind the scenes, doing all those different tasks that you often ignore as an engineer. The Gradial team. I think there's actually more similarity between enterprise marketing and building a satellite than people would expect. There's a schedule, there's a rollout, there are people contributing from different places, and there are hard problems to solve along the way. But the thing you really see in a startup is just how many of those smaller tasks that aren't very difficult add up to something that operationally is quite challenging to scale. # What does a typical week look like for you? Meeting with customers—usually on the road somewhere—and spending as much time as possible thinking about what marketing is going to look like in 2027. Diving into ideas with the team about what's become possible as the models have gotten better. Source: [Gradial](https://www.gradial.com/?utm_source=www.opensourceceo.com&utm_medium=referral&utm_campaign=marketing-has-too-many-humans-in-the-loop). We spend a lot of time testing and experimenting with the latest generation models to figure out what use cases are becoming possible or on the edge of possible. We've really done that since day one. You'd get these new model releases, and there'd be certain use cases that went from maybe 20-30% to 70-80% accurate, and you start to develop an intuition for what's on the edge of possible. So a big part of every week at Gradial is around how we are going to unlock that as the models continue to improve, and figuring out what tricks we can come up with to unlock use cases and capabilities that aren't possible yet with the current generation. # How has your job as CEO changed since the seed round? In a startup, your job is going to be different every six months. That's true for the CEO and for everyone at the company. A lot of times, that change is really adding more responsibilities, not subtracting them. Learning to multitask across a larger number of different parts of the business is something all of us are doing, but I definitely feel it very acutely. At the seed, a lot of my time was spent building the product and just writing code. I still do a fair amount of that today, but I've learned there's a need to start spending less time on that and more time thinking about the direction of what code we should even write. The hardest problem is spending enough time thinking about what to delete. It's easy to come up with a scope and ideas. It's harder to figure out what ideas to ignore. This is the thing I've been most focused on personally: figuring out how I can spend more time deleting stuff across the company that isn't needed. [Simplifying the product, the vision, and the roadmap](https://www.gradial.com/?utm_source=www.opensourceceo.com&utm_medium=referral&utm_campaign=marketing-has-too-many-humans-in-the-loop). Especially when you have AI that can go and build all of your ideas, if you're not careful about eliminating the worst ones first, you could easily crowd out your garden of ideas. It's like throwing weeds in your garden. Those weeds develop, and they take away a lot of space from the things that really matter. | Simplifying in public. | A year or two ago, that was not a problem. At the seed stage, you have the clarity of really one thing you're working on, and you just have to deliver that one thing. It's actually very simple: you just show up and work on it all day. Yet, as you start to take on more customers, more responsibilities, you have to be really purposeful about not spreading things too thin. # Are you using your own agents inside Gradial's operations? We've been using our own agents since day one. And as you start to add more people, you start to see a lot of the [pain points that Gradial is designed to address](https://www.gradial.com/?utm_source=www.opensourceceo.com&utm_medium=referral&utm_campaign=marketing-has-too-many-humans-in-the-loop), which is the operational handoff across people on the team. Something that resonated for me as an engineer coming into the marketing space was that whether it's a marketing project like creating a new website, or an engineering project, a lot of the work actually becomes coordination as the team gets larger. It's been really satisfying to start using Gradial in a new capacity as we've grown, because you get into exactly the problems the workflow has to solve. Like, who's able to approve that update to the website? When we had fewer people, there was just one person handling the website updates; they could approve them themselves. When you get to a certain scale, suddenly you need a creative asset from one person and approval from another person to make the update. And that's exactly what we've built our agent to solve. Agent harmony. Source: [Gradial](https://www.gradial.com/?utm_source=www.opensourceceo.com&utm_medium=referral&utm_campaign=marketing-has-too-many-humans-in-the-loop). Then, in other parts of the business, whether it's customer experience analytics and optimization on the site, we have agents that can basically optimize your site, the journey, and try to understand where a customer might get stuck across your different digital experiences, whether it's email or web. As our surface area of digital experience has grown, we've been able to really leverage that. We have these always-on brand monitors where the agent is going and looking at your website, looking at all these digital channels, and basically just trying to catch things where there's an opportunity to improve the product or to improve the experience. That's another place where the models have really unlocked capability in the last six months. The latest versions are becoming smart enough to use computers at a high level, which has opened up some really cool use cases that were not really possible even 12 months ago. | # Describe the difference between an AI tool and an AI agent in practice? A tool is something that takes an input and delivers a specific output. It does one step in a chain of steps. An example would be: go and spell-check this document, and you get back a spell-checked document. An agent is something that takes many different steps across at least one system, and the output is undefined. It can take a range of inputs and deliver a range of outputs. It's not one-to-one, and it's not a single step. For example, customers use Gradial agents to enter a workflow and [execute a campaign in a marketing setting](https://www.gradial.com/?utm_source=www.opensourceceo.com&utm_medium=referral&utm_campaign=marketing-has-too-many-humans-in-the-loop). A campaign can have many different steps. You could have a step where you need to get legal approval from someone in another time zone to review a document you're creating with the agent. One thing a customer can do is take a screenshot or photo of a whiteboard with their workflow and give it to the agent. Then, the agent will execute the workflow according to a diagram they drew on a whiteboard. That's an agentic capability because it wasn't programmed in. It's not hard-coded. I can literally go to the whiteboard, draw the workflow I want, take a photo on my phone, give it to the agent, and it goes and executes. Working with the best. In a tool, thinking about the prior generation of SaaS, things like RPA, you would have had to spend many, many hours of configuration just to implement a single workflow. Typically you'd have to bring in an expert to do it because the product is too hard to configure. For example, if you've ever tried to set up even a simple CRM, you end up on the phone trying to learn how the documentation works because you actually can't configure the workflow. In a tool, all the intelligence is in the hands of the person using it. It's like a screwdriver—I have to turn the screwdriver. In an agent, the intelligence is partly in the agent itself. You're not providing all the force. It's like a power tool where the energy comes from somewhere else. Source: [Gradial](https://www.gradial.com/?utm_source=www.opensourceceo.com&utm_medium=referral&utm_campaign=marketing-has-too-many-humans-in-the-loop). # How do you get the best out of yourself? I wake up every day excited because I'm working on something I love. That's why I know a startup was the right place to be. I get that energy just from doing something I really enjoy: working on hard problems where there's no clear path. No one's ever built agents at this scale before. No one's ever done what we're doing. | That's been just really gratifying, to wake up every day and say there are many hard problems to solve and also so much value to deliver in a space that is pretty greenfield. There's a lot of opportunity to [make the marketing workflow way more fun](https://www.gradial.com/?utm_source=www.opensourceceo.com&utm_medium=referral&utm_campaign=marketing-has-too-many-humans-in-the-loop) and way more effective for so many different teams. Having a list of ideas that you could never possibly do all of them. It's like having a travel list you could never fully complete, but you're checking something off every day. To me, that's extremely gratifying. # Anything else you'd like to share? I think we're going to see the next six months go even faster than the last. That hasn't always been true; there's been this narrative that AI has just been continuously accelerating, and that's partly true, but there have been periods where the relative impact was not accelerating. From roughly January 2024 through December 2024, the impact in the enterprise was not accelerating. It was pretty static. People were still stuck in POC hell across most applications. We were able to break through most of that POC landscape, but there was a period where people were in try mode and not in use-case mode. That has 100% changed. Everyone I'm spending time talking to in the marketing landscape is [accelerating their AI adoption](https://www.gradial.com/?utm_source=www.opensourceceo.com&utm_medium=referral&utm_campaign=marketing-has-too-many-humans-in-the-loop) and fundamentally rethinking their entire stack. Everything is on the table right now. The way I see it, and the way I'm hearing it from customers, is that all of the infrastructure built up over 20 years to run these different marketing technology practices has to be reevaluated in the time period of AI.‘ What is actually going to scale with us as we prepare to re-architect our marketing technology for AI agents?’ Gradial co-founders from left: Doug, Anish, Deip, and Anup. | A big part of that is that more of the traffic is coming from AI agents, and that's going to continue to accelerate. So a lot of the tools, techniques, and priorities that made sense three years ago make way less sense now. From a Gradial perspective, that is very much where we're focused: 2027 marketing and what the fundamental data primitives needed to support the marketing agents are. # Extra reading [Gradial Raises $65M to Build the First System of Work for Enterprise Marketing](https://www.gradial.com/blog/gradial-65m-series-c?utm_source=www.opensourceceo.com&utm_medium=referral&utm_campaign=marketing-has-too-many-humans-in-the-loop)- June, 2026[Announcing GEO with execution built in](https://www.gradial.com/blog/announcing-geo?utm_source=www.opensourceceo.com&utm_medium=referral&utm_campaign=marketing-has-too-many-humans-in-the-loop)- March, 2026[From Hours to Minutes: How Agentic AI Gives Marketing Teams Their Time Back](https://www.gradial.com/blog/technology/from-hours-to-minutes-how-agentic-ai-gives-marketing-teams-their-time-back?utm_content=377037700&utm_medium=social&utm_source=twitter&hss_channel=tw-1700280173529300992)- April, 2026[Agentic Marketing Is Rising. Gradial Is Leading the Way](https://www.gradial.com/blog/agentic-marketing-is-rising-gradial-is-leading-the-way?utm_source=www.opensourceceo.com&utm_medium=referral&utm_campaign=marketing-has-too-many-humans-in-the-loop)- July, 2025 **BRAIN FOOD** 🧠 **TOOLS WE RECOMMEND 🛠️** Every week, we highlight tools we like and those we actually use inside our business and give them an honest review. [Today, we are highlighting Lightfield](https://lightfield.app/old-home?utm_source=newsletter&utm_medium=paid&utm_campaign=open_source&utm_content=primary_3-22-2026)*—an AI-native CRM that assembles itself from your email, calendar, and meetings. See the full set of tools we use inside [Athyna & Open Source CEO here](https://athyna.notion.site/Our-Complete-Business-Toolkit-2cdd8af0f44d4093a91f728bec4de4f7?utm_source=www.opensourceceo.com&utm_medium=referral&utm_campaign=marketing-has-too-many-humans-in-the-loop). **HOW I CAN HELP **🥳 # P.S. Want to work together? **Hiring global talent:**[If you’re hiring](/cdn-cgi/l/email-protection#0c68636f4c6d786475626d226f6361336f6f317e636e4c6d786475626d226f63612a6d617c377f796e66696f783145293e3c7b6d6278293e3c7863293e3c64657e69293e3c7b657864293e3c4d786475626d222a6d617c376e63687531446975293e3c48636f293e4f293e3c62656f69293e3c63626922293e3c45293e3b68293e3c60637a69293e3c7863293e3c64696d7e293e3c61637e69293e3c6d6e637978293e3c64637b293e3c4d786475626d293e3c6f63796068293e3c6469607c293e3c7b657864293e3c6175293e3c64657e65626b22)tech, business or ops talent and want to do it 80% less, check out my startup,[Athyna](/cdn-cgi/l/email-protection#bdd9d2defddcc9d5c4d3dc93ded2d082dede80cfd2dffddcc9d5c4d3dc93ded2d09bdcd0cd86cec8dfd7d8dec980f4988f8dcadcd3c9988f8dc9d2988f8dd5d4cfd8988f8dcad4c9d5988f8dfcc9d5c4d3dc939bdcd0cd86dfd2d9c480f5d8c4988f8df9d2de988ffe988f8dd3d4ded8988f8dd2d3d893988f8df4988f8ad9988f8dd1d2cbd8988f8dc9d2988f8dd5d8dccf988f8dd0d2cfd8988f8ddcdfd2c8c9988f8dd5d2ca988f8dfcc9d5c4d3dc988f8dded2c8d1d9988f8dd5d8d1cd988f8dcad4c9d5988f8dd0c4988f8dd5d4cfd4d3da93). 🌏**See my tech stack:** Find our suite of[tools & resources](https://www.notion.so/athyna/Our-Complete-Business-Toolkit-2cdd8af0f44d4093a91f728bec4de4f7?pvs=4&utm_source=www.opensourceceo.com&utm_medium=referral&utm_campaign=marketing-has-too-many-humans-in-the-loop)for both this newsletter and Athyna[here](https://www.notion.so/athyna/Our-Complete-Business-Toolkit-2cdd8af0f44d4093a91f728bec4de4f7?pvs=4&utm_source=www.opensourceceo.com&utm_medium=referral&utm_campaign=marketing-has-too-many-humans-in-the-loop). 🧰**Reach an audience of tech leaders:**[Advertise](https://www.passionfroot.me/open-source-ceo?utm_source=www.opensourceceo.com&utm_medium=referral&utm_campaign=marketing-has-too-many-humans-in-the-loop)with us if you want to get in front of[founders, investors and leaders](https://www.passionfroot.me/open-source-ceo?utm_source=www.opensourceceo.com&utm_medium=referral&utm_campaign=marketing-has-too-many-humans-in-the-loop)in tech. 👀