Home/Growth experiments/AI and Digital Marketing for B2B Startups
More growth from the same marketing budget
We bring AI and digital marketing together, directing agents on research, writing and campaign work. Experienced founders and engineers put the hours saved into more growth experiments. You get that support for less than the cost of a junior marketing hire.
AI and digital marketing now share the work
Agents can handle much of the research, drafting, design and campaign production behind digital marketing. People choose what to test and direct that work toward the company's buyers. The output still needs judgment: a draft can sound plausible while describing the product incorrectly, and a report can mistake automated visits for interest. How the work is divided affects both the time saved and what reaches buyers.
Running AI yourself gives you direct control of the tasks and their output. Hiring an agency by the hour gives you people for an agreed amount of time. We take responsibility for the work and use agents for production, putting time saved into further tests. The comparison is who gets the benefit of that time and who takes responsibility for the quality of what it produces.
Hourly billing keeps the AI saving from you
When you buy agency hours for work AI can help produce, saving production time also changes what the agency can bill.
- Automation competes with billed hours. An hourly model gives the provider less reason to reduce the time each task takes.
- Small budgets can lose experience. Busy agencies can assign junior staff to smaller accounts, leaving less experienced people choosing the work.
- Unchecked drafts cost trust. A page that reads as AI-made can weaken a B2B buyer's confidence in the company.
Paying for work with people responsible for its quality gives production savings somewhere useful to go.
We choose the tests and direct the agents
We decide which campaign or page to try, then assign its research and production to agents. Our people remain responsible for getting the experiment into market.
Choose the experiments
We examine where customers come from and identify the tests worth running next.
Direct the production
We give agents the research, writing, design and campaign tasks needed to carry out those tests.
Check the buyer-facing work
People on our team read each page, post and ad before it goes out.
Run the chosen channels
We handle search, LinkedIn, outbound and paid campaigns, with people writing the social posts.
Each week starts from what buyers just did
We look at what happened after a page or campaign went live before assigning more work. The next test starts from that response.
First
Set the first tests
We use the growth plan to choose the campaigns and pages that should receive attention.
Then
Prepare the test material
We direct agents on the research, drafts and campaign builds each experiment requires.
Next
Settle the important choices
You approve the important steps with the people responsible for carrying the work into market.
Each week
Use the buyer response
On the call and in Slack, we decide what to continue, change or stop from what happened.
Example: a bot crawl read as growth
On our own projects, we found that an agent's browser testing was counted as traffic. A bot crawl appeared as "28x" growth, and Search Console was connected to the wrong domain. We caught these errors while reading the reports, before the apparent growth could set the direction of the work.
Hours saved by agents go into more tests
We are paid for the growth work, with no long-term commitment. If agents save time preparing a campaign, we can spend that time on another test. Our people decide which idea deserves it and examine what happens.
- Follow the growth question. We choose tasks for what they can teach us about reaching buyers and building demand.
- Spend saved time on tests. Production savings let us try more of the ideas in the company's growth plan.
- Reserve attention for judgment. Our people choose where experiments go next and examine the evidence that supports those decisions.
It fits founders who want agents run for them
Hand over agent-led production
Founders see useful applications for AI but want someone to run the marketing work for them.
Add experienced marketing direction
Pre-seed to Series A B2B companies have customers and no experienced marketing lead.
Explain an AI-enabled service
Companies selling expertise through software and AI need buyers to understand both the service and the people.
Talk to Tin
Let's look at the marketing work filling your week and which research, writing and campaign tasks our people and agents could take on.
FAQ
Should I run AI marketing agents myself instead?
Yes, if you want to direct the agents and maintain their work. Our open-source system runs through a coding agent and can be self-hosted. Founders who ask us for help generally want someone to choose the experiments and carry them out, which is the work our people take on.
Do you work with the tools we already use?
Yes, where a connection exists today. During setup, we establish which of your search, site, email, advertising, payment and analytics accounts we can use. We read the data from those accounts to choose experiments and examine what happened. The campaigns and their records remain in your company's accounts.
Will our content read as AI-written?
We write for your buyers in your company's voice, starting from the actual product and the point each piece needs to make. Agents gather the material and draft the outline. People shape the writing so a page or post gives the reader a clear, accurate explanation of your work.
Who keeps the accounts and the work?
Yes: the accounts, the data and every workflow we build belong to your company. The work develops in your company's accounts, with reports, research and project records kept as files. That gives your team the material behind the campaigns and decisions as well as the finished pages and content.
Talk about where agents fit in your marketing
Speak with the founders and engineers who direct the work about how your marketing gets done today. We can discuss where agents could handle production and where experienced people should spend their time choosing the next experiments.