What if, instead of starting with a campaign brief, you could open a personalized feed of the most important opportunities for your lifecycle team? And what if that feed was unique to your business, your customers, your products, your campaigns, your journeys, and your goals?
The feed might surface a lapsing audience, a fatigued journey, a cross-sell opportunity, or a campaign you need to refresh. And what if each opportunity came with the context and tools needed to act on it immediately to create personalized, on-brand experiences for your customers?
Solving that problem is the reason I bet on joining Hightouch in 2021 when we were fewer than 20 employees. In 2026, this idea is now very much a reality, thanks to something called the marketing harness.
Why AI models alone are not enough for lifecycle
It feels like there is a new foundational AI model almost every week. The models are getting remarkably good at reasoning, generating, analyzing, and creating almost anything. But the model that is most capable today might be just as easily surpassed by another one tomorrow. LLMs also tend to index toward the aggregate average. That can be useful, but average does not necessarily equal good, and it rarely equals great marketing.
Marketing is extremely complex and nuanced. It isn’t a binary discipline where 1 + 1 always equals 2. While an LLM can draw inferences based on the context that you provide, it doesn’t explicitly know whether a message is on brand, whether or not an offer conflicts with another promotion, or if a customer is overlapping in two different campaigns.
Put simply, LLMs can reason, but they cannot run entire lifecycle marketing workflows on their own.
What is a marketing harness?
An agent harness is everything built around an LLM that makes it possible to do multi-step, enterprise work. A marketing harness is purpose-built for marketing workflows. I like the way our CEO, Tejas Manohar, explains it:
“Imagine you hire a new marketer on the team. They’re brilliant and best-in-class at what they do. But you give them no access to your tools, tell them nothing about your brand, your business, and what you’ve tried in the past. Then you say: “ok, now go launch a campaign.” There’s a good chance something will go wrong. That’s what most of us are doing today with AI. We’re expecting super powerful models to do well without the right inputs. That’s the role of the harness.”
Tejas wrote a full deep dive on the marketing harness, and I highly recommend reading it. But here’s my quick summary of the four components he says every harness needs:
Work with your team’s existing process: help agents and people move through briefs, creative, revisions, and approvals without throwing away how the team already works.
Understand your business, brand, and goals: build a shared marketing brain that connects customer data, brand knowledge, business rules, and campaign history, and learns from feedback and outcomes over time.
Work across your tools: give agents the connections and expertise to build audiences, configure journeys, and launch approved campaigns in the systems your team already uses.
Enforce guardrails and swim lanes: limit what agents can access and do, require the right approvals, check that decisions are grounded in real sources, and keep a record of their actions.
Lifecycle marketing is a particularly demanding specialization because the relevant context doesn't live in one profile. It spans customer behavior, business goals, brand rules, campaign history, journeys, and the systems where work gets done.
The marketing harness is the operating layer that connects that context through one workflow and one interface, so teams can use it at every step of the marketing process without ever losing context. And the beauty of this system is that teams move away from a collection of isolated AI sessions to a collaborative system where knowledge compounds and campaigns improve over time.
Why lifecycle marketing needs a harness
Lifecycle marketing encompasses a lot of use cases, a lot of channels, a lot of campaigns, and a lot of decisions. In fact, lifecycle marketing at its core is a series of continuous decisions:
Which campaign should we run?
Which customers should receive it?
What should we say?
Which channel and timing make sense?
What should happen next?
Today, marketers translate these tasks manually into a data request, an audience, a campaign brief, a creative brief, a new journey, a new QA process, and a bunch of other steps to approve the campaign and get it live.
Every campaign starts from scratch, and the learnings from those campaigns rarely get captured in a way that is reusable and broadly applicable to the rest of the organization. Not to mention, the work required to execute an idea limits how many ideas a team can even pursue. In all my conversations with lifecycle teams, this is perhaps the most common frustration.
There are four core workflows in lifecycle marketing, and the harness speeds up every one of them:
Opportunity identification: Agents identify the audiences, insights, and moments worth acting on; lifecycle marketers decide which ones to pursue.
Campaign creation: Agents draft the brief, audience, copy, and creative; the campaign owner and other stakeholders refine it.
Channel execution: Agents build and QA the campaign across channels; legal and brand teams correct it, and lifecycle marketers approve launch.
Learning and optimization: Agents connect results back to the decisions behind them, and the lifecycle team decides what to change next.
Every iteration makes the next campaign better, which is how idea-to-launch time keeps shrinking.
How the marketing harness impacts the lifecycle workflow
A marketing harness for lifecycle takes each of these components and turns them into one shared workflow that connects customer and marketing context to the tools, skills, people, and systems required to bring a campaign to life.
I’m going to use our own Agentic Marketing Platform at Hightouch as an example because it really grounds this whole discussion into something tangible.
Hightouch is (1) the marketing brain that brings all of your data, campaign history, brand knowledge, and business context in one place so you can surface insights and build audiences and (2) the workflow execution layer, which actually builds campaigns and executes them across any marketing channel.
Here’s how that would work in a real-life scenario:
Imagine you’re a lifecycle marketer at a home improvement retailer. You log into Hightouch on Monday morning, and the agent has already flagged an opportunity: a severe freeze is forecast across several states next week, and you have high inventory on space heaters, pipe insulation, and weather stripping. It’s also identified the customers in the regions most likely to need them.
You then draft a short prompt based on your desired goal:
“Build me a campaign brief for this upcoming freeze. Prioritize customers who browsed cold-weather products but haven’t bought yet. Lead with preparation tips, not discounts, since we don’t want to look like we’re cashing in on a storm. Exclude anyone who already bought in the last 30 days or is in an active win-back journey. Messages need to land by Wednesday, before the freeze hits.”
Hightouch then turns that into a campaign brief for you to review, pulling from your customer behavior, inventory, journey state, and past campaigns. It splits the audience into groups, like customers who browsed space heaters and customers who bought outdoor plumbing supplies, and presents the offer, channel, and timing for each, with the reasoning behind every recommendation.
Once you’ve finished reviewing and adjusting the campaign brief, you then move to production, and the platform turns those opportunities into real messages and campaigns using pre-approved creative and templates. In this case, that might mean heater browsers receive a product-focused message while plumbing customers receive pipe-protection guidance.
Hightouch would then match the channel to the audience and even manage the sending directly through your customer engagement platform of choice or our own infrastructure. For example, users who browsed heaters on the mobile app might receive a push notification, while plumbing customers who don’t use the app might get an email.
And like any good harness, it runs on guardrails: your team sets what the agent can access and do, it verifies every message against real data, and it keeps an audit log of every action. Before anything goes live, your brand and legal teams sign off on the messages.
In the past, the freeze would have come and gone by the time this campaign made it out the door. Now, it takes a few hours to go from idea to launch. And the best part is the engagement and performance data gets captured so it can inform every future campaign.
So a year later, when another freeze approaches, the marketer can start with a new goal, but the harness will remember everything from last year. It would be able to identify that customers who bought space heaters also bought weather stripping within the following week, and recommend pairing the two products in the next freeze campaign.
Note: If you want a closer look at how this works step-by-step within Hightouch, my colleague Anthony Chiulli walked through the full workflow in this post.
What marketing harness unlocks for lifecycle teams
Every marketing team has access to the same powerful LLMs. But for most teams, that access hasn’t yet transformed lifecycle marketing. If you want to see the real impact of AI, the harness is the key. It’s the marketing brain that houses everything unique to your business around the model, and everything your team does gets captured and carried into the next campaign. You never have to translate your business into a prompt or lose context again.
And if you want to see what this looks like in action, I’m going to be breaking it all down in a live workshop on October 23rd with Peter Oleson (ex-Iterable). Make sure to secure your spot before it’s too late.