Should B2B software be self-serve or high-touch at first?

When I set direction for lightbreak's build, I was blunt about it:

"lightbreak is not going to be a self-service platform. This is going to be a high-cost, high-touch platform for the foreseeable future, so do not worry about Stripe or any automated onboarding. Just imagine I would do the onboarding for them, at least for the next 5."

At the time I had one client on it and a goal of 20 to 50.

Why go high-touch first?

• The buyer isn't technical. lightbreak's customer is a legacy local media company running several to 20 newsletters. Most are extremely un-tech-savvy, but a lot of them want to move forward.
• The real risks aren't billing. The questions are whether the systems, forms, events and email sending work in a new town, and whether writing quality and local sources hold up.
• The price can climb. My plan was to move lightbreak from $100 a month toward $200 to $300 as it went upmarket.

Going upmarket has a catch. The buyer gets further from the person who does the weekly work, so somebody has to actually run production after the contract is signed.

Who is the customer, if solo writers already have AI?

Solo newsletter writers are mostly served by the AI built into their email platform. There isn't a big enough need there. The target is local media companies, especially legacy ones that now run lots of newsletters, usually on Newspack plus Mailchimp or ActiveCampaign.

My idea for them: an editor says "write a new edition" in Slack and a draft shows up in their inbox to approve, without ever logging into my tool. An AI I was working with summed up the pitch well: sell "your Tuesday draft is already waiting," not the API.

When is self-serve worth building?

After the managed version works. I prefer selling a capability as a managed service first and productizing once the interface is solid. When greetform got self-serve plans, the requirements were concrete:

• Google sign-in, no Apple sign-in
• A "prove you're human" check
• A Help button with a message form that takes up to five photos and routes to a shared support inbox
• A polished enterprise inquiry form asking about number of publications, employees, revenue and goals
• Never save or show an empty question in the editor

What's the path from one client to investment?

My stated path for lightbreak: nail the pilot, generate referrals, reach $4,000 to $5,000 in monthly recurring revenue, then approach an AI-native technical partner or a local-media angel for a 10-15% equity stake rather than a straight investor. See lightbreak for what it does today.