---
title: "We didn't add AI to marketing - Atlan Frontier Labs"
url: "https://atlan.com/frontier/essays/we-rebuilt-marketing-around-it/"
description: "How a bet made at the start of Q2, cross-functional pods, and one shared context layer changed what our marketing team is for. The build log, the learnings, and what we got wrong."
---

> Atlan is hosting Context Conference, bringing together the leaders and builders at the frontier of giving AI the context it needs to understand their business. It runs online on October 28, 2026, from 11:00 AM to 2:00 PM ET. Atlan co-founder Prukalpa Sankar opens and closes the day. Leaders from AstraZeneca, BNY and Verizon share why they invest in context and what they get from it. Registrants get early access to The AI Context Gap, a new study from MIT Technology Review Insights. Register: https://atlan.com/context-conference/

An essay from Atlan Frontier Labs (essay 003) by Surendran Balachandran, Marketing, Atlan, dated 2026.07.29, 1,451 words. Title on the page: "We didn't add AI to marketing. We rebuilt marketing around it." How a bet, cross-functional pods, and one shared context layer (Marketing OS) changed what Atlan's marketing team is for: the build log, the learnings, and what the team got wrong. Part of [Becoming Frontier](https://atlan.com/frontier/).

## No decks: a room full of people building

In early July the marketing team came together from around the world at the foothills of Nandi Hills, outside Bangalore, for the H2 kickoff. No presentations, QBRs, planning or strategy decks. Instead, two things:

1. A debate on what marketing means in an AI Native world: when execution is not a bottleneck, how to raise the bar by focusing on taste, judgement and ambition.
2. A hackathon in thematic groups: 36 hours of building, ending with a showdown of the demos.

## The bet

When ChatGPT hit the world in November 2022, the author thought it would change everything, including how companies are built and run. The question 18 months on: to start 2026 as one of the most AI Native companies in the world, what needs to happen today? The question behind the bet: could a 30-person team do what usually takes a 300-person team?

Starting 2026, the team rebuilt marketing from scratch, on two foundations: a pod structure instead of a functional org, and Marketing OS, a shared context layer that humans and agents both work from.

## From tasks to outcomes

Marketing teams are usually organized around specialised functions: content develops a story, design gives it form, web publishes it, events create the experience, operations connect the systems. Function builds depth but leaves task owners and no outcome owner. A pod owns the outcome and carries enough cross-functional context to take something from idea to completion.

Example: in Q2 2026 the Media & Community pod carried three flagship moments: Activate (Atlan's biggest virtual product launch), Snowflake Summit, and Databricks Summit. Events have hundreds of moving parts, a dozen functions, and many hops between them; each hop adds friction, and timezones slow every feedback loop. This year the pod owned the whole outcome, with Marketing OS as the foundation for experiences across digital and physical worlds. An event reduces to a few primitives (theme, brand guidelines, personas, channels); once standardized, humans and agents could build everything on that foundation. With most execution handled by agents, the team focused on the experience, the programming, and the quality bar. Result: Atlan Activate, on April 29, became Atlan's biggest product launch yet, with several thousand people joining, overshooting its goal by more than 60%.

## The shared context layer for humans and agents

Pods reduce friction but don't fix context: five functions in one room still need a single source of truth (story, brand, personas, offers, proof points), or the handoffs just move inside the pod. Marketing OS is the marketing team's brain:

- One shared context layer for the team and its fleet of agents, with access to all the tooling, data, systems of record, and platforms.
- A curated knowledge and skills layer on top that captures tribal knowledge: every definition, method, process, and playbook, maintained by its owner.
- Guardrails, what great looks like, and evals, so agents can orient themselves and work backwards from the goal like a team member would.
- Skills stack like Lego blocks into workflows. Example: launching an event registration page daisy-chains skills for brand, messaging, positioning, copywriting, personas, past performance, page build, and design critique.

A figure shows the architecture: internal systems, analytics, ad platforms, social, and data sources feeding one shared, version-controlled context layer of skills and knowledge, with agents plugging in and outputs flowing to content, campaigns, enablement, reports, creative, and the website.

Context compounds: feedback from every campaign, report, launch, article, and creative flows back into Marketing OS. In the last few months the team created and refreshed over 300 skills. Citing Bret Victor's talk "Inventing on Principle" (creators need an immediate connection to what they create), Marketing OS reduces the time and effort between having an idea and seeing it come to life.

## What "AI-native" actually means

Most of what gets called "AI-native" is AI-adjacent: broad tool access, high usage, a wall of experiments, a productivity number. Usage is not the standard; if it were, the winner would be whoever burned the most tokens. The real standard is whether new capability turns into new behavior: whether the organization changes its shape around what it has learned it can do, and raises its quality bar to match. The team didn't get faster at the old marketing; it changed its shape, then what it held itself to. See [Becoming a frontier company, in the open](https://atlan.com/frontier/essays/becoming-a-frontier-company-in-the-open/).

## What we got wrong

- **Too many skills, too quickly.** Redundancy made quality hard to measure; at one point there were three copywriting skills (web pages, ad campaigns, announcements and blogs). Better: one copywriting skill that imports others, like tone and voice, and complements downstream skills like ad copy.
- **Skills grew too long, and quality degraded.** Skills work best modular and chained toward the outcome. One skill for the weekly review deck packed in data analysis, sprint review, deck building, and posting to Slack, and models started missing instructions. Splitting it into chained skills for analysis, review, and the artifact improved quality a lot.
- **Just because we said it doesn't make it so.** Early on, the content system invented pricing for another company's product; AI summaries in search picked it up and assistants repeated the mistake as fact, despite guardrails. Lesson: build solid checks and evals from the start.
- **Agents discover skills differently than humans.** Humans start with an open-ended problem and invoke skills along the way; agents start with a specific goal and load skills upfront, and every harness loads different files in a different sequence. Delivering the same quality across harness and model combinations is a work in progress.
- **Scaling taste and judgement with AI is hard.** Quality is enforceable when output is verifiable: website health monitoring and regression catching is 100% agent-driven. The same is not yet true for writing the title of an event.

## What's next

- More skills and agents in Marketing OS; evals to tighten the feedback loop for skills are high on the list. Early benchmarks across twelve tasks show the same model scoring higher with a skill than without on nearly every task.
- More autonomous agents, driven by feedback loops for continuous improvement. The author's favourite hackathon demo: Argus, a new agent that continuously evaluates skills and agents and sends personalised feedback and recommendations to the team.

Execution is not a bottleneck; this is an era of infinite marketing. So what does it mean to be a marketer in this era? "Taste, judgment, ambition, agency: we believe these are the new human frontiers."

## Author

Surendran Balachandran, Marketing, Atlan. Surendran leads Atlan's marketing AI taskforce and the Marketing OS: the operating model that rebuilt the team around cross-functional pods and one shared, governed context layer.

## Read next

- [We rebuilt our content engine four times](https://atlan.com/frontier/essays/a-self-learning-content-engine/) (2026.08.25, 12 min read)
- [What it took to build an agent our CSMs actually trust](https://atlan.com/frontier/essays/an-agent-our-csms-actually-trust/) (2026.09.18, 11 min read)
- [Becoming a frontier company, in the open](https://atlan.com/frontier/essays/becoming-a-frontier-company-in-the-open/) (2026.07.23, 7 min read)