What is Navless Guide and how does it work?

Navless adds a configured Guide to any landing page via one snippet. Each Guide gets its own knowledge base, brand kit, skills, and analytics, calibrated for the campaign. The Guide also captures contacts directly via a CTA that triggers a HubSpot form mid-session; the page handles structure, the Agent handles the conversation.

How do I structure a repeatable campaign workflow with agent pages and subdomains?

Every campaign gets its own page, its own configured Guide, and its own subdomain. The current campaign sits at agentpages.navless.ai; the next campaign sits somewhere else. Same workflow, fresh instance per campaign.

What tools make up the campaign stack in this approach?

Page in Lovable, Guide in Navless, distribution in HubSpot, attribution in Google Analytics, and Claude as the surface layer to manage it all.

How do I track campaign attribution across channels?

GA closes the attribution loop. Every campaign link carries a clean UTM string so posts and emails roll up to the same campaign, and GA shows which channels drive the right traffic while Agent analytics show what visitors asked once they arrived.

How I use Lovable, Navless, HubSpot, Google Analytics, and Claude to run a marketing campaign on a loop

The author describes a scalable marketing campaign workflow using Lovable to build individual campaign pages on subdomains, Navless's Guide to create personalized, agentic content experiences with integrated contact capture and conversion triggers, and HubSpot to manage contacts and email sequencing, all designed to replace the traditional slow, project-based campaign launches with a repeatable, modular system.

A few months ago I stopped treating each campaign as a fresh launch project and started treating it as a known workflow.

The old pattern wasn't going to scale at a lean AI startup: a designer queue, an engineering ticket, six weeks of build, two weeks of relevance, then stale.

The new pattern: every campaign gets its own page, its own configured Guide, and its own subdomain. The current Agent Pages campaign sits at agentpages.navless.ai. The next campaign sits somewhere else. Same workflow, fresh instance per campaign.

This is the stack I'm currently using. One way to run the loop, not the only way. Navless covers the agentic layer — that's the piece I'd call essential, and it's the one we sell. The rest is interchangeable. Use whatever your team already runs.

Lovable — the page layer

Each campaign's page gets built in Lovable on its own subdomain. agentpages.navless.ai is the current one. It is great that they allow you to easily host things. It's my go to for solo landing pages.

Guide, By Navless — the agentic content experience layer on top of the page

Full disclosure: I'm Director of Growth Marketing at Navless and part of the founding team. We built this layer because it didn't exist for what marketers actually need to ship.

Navless adds a configured Guide to any landing page via one snippet. Each Guide gets its own knowledge base, brand kit, skills, and analytics, calibrated for the campaign running on the page. Same Agent Page, different experience per visitor based on what they came to figure out.

The Guide also captures contacts directly. A CTA configured inside the Guide triggers a HubSpot form mid-session, making the agentic layer its own conversion surface alongside the page's CTA. Visitors convert when their intent shows, not just at the static form on the page.

The page in Lovable handles the structure. The Agent on top handles the conversation.

HubSpot — the contact and sequencing layer

HubSpot is where contacts, lists, and email Workflows live. LinkedIn launch posts schedule through HubSpot's Social Publishing so they tag to the same Campaign object as the emails and the microsite. Everything rolls up to one campaign dashboard instead of being stitched across tools.

Google Analytics — the attribution layer

GA closes the attribution loop. Every campaign link carries a clean UTM string, so a tagged URL in a LinkedIn post and a tagged URL in an email roll up to the same campaign. GA tells me which channels drive the right traffic. The Agent Pages analytics underneath tell me what those visitors actually asked once they arrived.

Claude — the layer where I run the campaign

Claude is the layer where I personally manage the campaign. I draft copy in Claude, scope decisions there, and run the project across the team in Cowork (Claude with file access and connectors). The campaign's voice rules, brand discipline, and prior decisions live in the project so the work stays consistent across surfaces and across campaigns. The marketer in the loop is me. Claude is the workspace.

How the loop actually runs

1.Plan a new campaign. Define the audience, angle, and offer. Brief in Claude.
2.Spin up a new subdomain for the campaign. Build the page in Lovable.
3.Configure the Guide in Navless. New knowledge sources, brand kit calibrated for the campaign, skills calibrated for what this audience will ask, CTA logic per the funnel stage. Deploy on the new Lovable page via the snippet.
4.Distribute via HubSpot. Workflows for email, Social Publishing for LinkedIn, all tagged to the Campaign object.
5.Measure in GA and HubSpot for traffic and attribution, in the Agent's analytics for in-session evaluation behavior. Read at Day 30 against whatever baseline I have. Adjust the configured Guide, the segmentation, or the distribution mix based on what the data says.

→ Spin up the next campaign on the same workflow. New subdomain. New page. New Guide. Same stack, fresh instance.

What I'm excited about for this current campaign

This is the first campaign that runs end-to-end on my fully connected stack: page in Lovable, Guide in Navless, distribution in HubSpot, attribution in GA, and using Claude as my surface layer to manage it all (within its approved capabilities per tool). It's a different work experience than bouncing between tools and much easier to get one clear picture.