We think of ourselves as a lab.

We develop systems that fix the GTM problems teams face now. Then we run them, on each company’s own context, with a person approving what goes out.

The problems we’re working on.

Lean B2B teams are feeling the same five pressures at once. Each one is a reason the old way of growing stops working.

  • Tight budgets
  • Pressure to do more with AI
  • Good talent that’s hard to hire
  • Channels that keep performing worse
  • A flood of AI slop buyers tune out

Help lean B2B companies grow without growing headcount.

Here’s what we believe about how to do it.

Every business should own its context.

What it does, who it serves, how it talks, and what its words and numbers mean. AI models are rented and replaceable. Your context and workflows shouldn’t be.

Use AI for the scarce, hard work.

Research, targeting, signals and quality systems. Not spam comments, lazy outbound or generic content.

Keep a person on what customers see.

AI drafts scaled content like search pages, with an editor on every piece. A person writes every comment and social post, and approves every message.

How we got here.

  1. Late fall 2025

    Claude said Aptible had 4,124 customers.

    Henry connected Claude to Aptible’s data warehouse and asked how many customers they had. It handed back a polished, board-ready chart showing 4,124. The real number was just under 1,000. The data was right. The meaning was missing. Once Claude had Aptible’s actual definition, it got the answer right.

  2. January to April 2026

    A product marketer became an SEO engine.

    Michael, Henry and Aptible’s product marketer, Kelsey, wrote down Aptible’s company context: positioning, voice, customer calls, product knowledge, traffic and user data, and SEO data, all usable in Claude Code. Kelsey went from waiting on engineers to doing SEO herself. Organic qualified signups grew 3.4x.

    Read the Aptible story →

  3. April 2026

    Software got cheap for everyone.

    Henry left Aptible. The newest AI made it feel like we could build anything. Then the obvious hit: it made software cheap for everyone. Building stopped being the moat. Trust and attention became scarce, and AI slop started ruining every channel.

  4. Summer 2026

    Back to the lab.

    Our first product, for product marketers, didn’t land. We learned to keep experimenting instead of pushing, and to stop treating one result as the verdict. Henry calls it putting the lab coat back on.

  5. October 2026

    Protocol today.

    A fractional AI GTM team for lean B2B companies, plus the GTM systems we run for them every month. Every one of them runs on the client’s own company context.

Two of us, by design.

Companies try to hire one AI-native leader who’s great at marketing and great at AI. That person is a unicorn, and the actual unicorns are hiring them. We’re that profile, twice, with complementary strengths. Each of us can cover nearly the whole job.

Henry Hund

Henry Hund

Marketing strategy

Henry has spent 15 years as the operator inside B2B software companies, taking growth, marketing, sales, product and operations from zero to one. At Protocol he picks the job with you and coaches your team.

  • Aptible, 2017 to 2026. Growth, operations, and interim head of product, marketing, sales and customer success, ending as COO. Revenue grew about 7x with almost no marketing headcount or budget. He led the sale process, and Aptible was acquired in November 2025.
  • Stitch, 2015 to 2017. Helped launch the ETL platform as a founder for hire, and built a programmatic SEO program that produced more than 60% of its leads. Acquired by Talend.
  • RJMetrics, 2014 to 2015. Ran growth, and led a team of 5+ engineers building an outbound email system that sent hundreds of thousands of emails a quarter, before Apollo or Outreach existed. Acquired by Magento, now part of Adobe.
  • Before that, Morgan Stanley and a Wharton MBA.
Michael Peterson

Michael Peterson

GTM engineering

Michael has spent more than a decade as a growth engineer, building GTM systems before “GTM engineer” was a job title. He has real software engineering behind him and fluency building AI tools. At Protocol he builds the context layer and the workflows.

  • Henry and Michael have worked together for about 8 years.
  • At Aptible, Michael built the connection that lets the team use its company data and knowledge in Claude Code.

Work with us.

Start with one sprint. Leave your work email, confirm your website, and pick a time.

Rather we run it for you? See the Search Visibility System →

Is this your company’s website?

We take a quick look, then you pick a time.

Enable JavaScript and reload this page to continue.