How to Sell Client AI Systems Built with Claude Code

TL;DR

Build client-facing AI systems by giving Claude Code a system prompt and temporary Lerty API access, then describing the databases, agents, dashboards, and sample data the client needs. Focus each offer on a measurable business problem, charge according to its value, and use Lerty organizations to separate client environments and deliver access through one platform.

Transcript

In the world of AI at this very moment, there is pressure like we haven't seen before. Not pressure on people like you and I who are builders who are looking to use these new tools, but pressure on people who don't really know about AI yet. The plumber, the HVAC technician, the lawyer, the dentist office, all of these people who don't really have t... Read More

Key Insights

• Claude Code is the system-building layer that can create Lerty databases, agent structures, and overall architecture from conversational instructions after it receives the appropriate system prompt and API access.
• Lerty is the client-delivery layer that combines organizations, AI agent chat threads, relational databases, and dashboards in one shareable platform, making completed systems accessible through invited user accounts.
• Each Lerty organization can represent a separate paying client, and the agency plan described in the transcript supports up to 50 organizations for managing multiple client environments.
• A sellable AI system must address a meaningful business problem, and a premium offer should help the customer make money rather than merely save time or provide a convenient extra feature.
• The proposed pricing range is at least $500 to $1,000 per month for the listed agentic systems, with the system's scope and business value determining where it fits within that range.
• Temporary API access reduces credential exposure because builders can give Claude Code broad permissions only during construction, set the key to expire, and delete it after the system is built.
• A coaching accountability engine is a practical offer because it can organize client records, preserve session history, manage action items, and help a coach prepare for meetings by identifying past events and patterns.
• Realistic seeded data improves a system demonstration because prospects can immediately see how records, histories, tasks, agents, and dashboards work without evaluating an empty workspace.
• More videos from this channel:
  • How to Automate Marketing with Claude and Higgsfield
  • How to Build Claude Skills with the DBS Framework
  • How to Onboard Claude as a Business Employee
• More from this channel: AI Foundations

Questions & Answers

Q: How do you build an AI system with Claude Code and Lerty?

Create a client organization in Lerty, generate an API key, and insert that key into the provided system prompt. Paste the completed prompt into Claude Code so it understands how to interact with Lerty. Then describe the business outcome, required databases, agents, dashboards, workflows, and sample data in plain language. Claude Code can use those instructions to build the system architecture inside the selected organization.

Q: What does Lerty provide for client-facing AI systems?

Lerty provides organizations, AI agent chat threads, relational databases, and dashboards within one platform. Builders can invite users by email and give clients access to the completed environment. Agents can use a connected model, custom instructions, starter prompts, MCP servers, or webhook connections to automation tools. Dashboards provide an attractive interface for viewing and changing information stored in the system's tables.

Q: How can an AI system be priced at $500 to $1,000 per month?

The transcript proposes charging at least $500 per month for systems that create enough operational value, such as saving a coach hours and improving client satisfaction. Offers closer to $1,000 per month need to do more, particularly by helping the business generate money instead of merely saving time or delivering a convenient feature. Pricing should therefore reflect the commercial result and depth of the system.

Q: How should Lerty organizations be used for agency clients?

Each Lerty organization can function as a separate environment for one agency client. The agency plan described in the transcript includes up to 50 organizations, so a builder can separate systems, data, agents, and access among multiple customers. The example creates an organization called Elite Coaching before any databases or agents are built, establishing a dedicated workspace for that coaching business.

Q: How should API keys be handled when Claude Code builds a Lerty system?

Create an API key within the target organization's settings and grant the access needed for the build. Paste it into the system prompt before sending that prompt to Claude Code. The transcript recommends treating the credential as temporary: set an expiration time if desired and delete the key after construction. This avoids leaving the build credential active after Claude Code has finished editing the organization.

Q: What should a coaching client accountability engine include?

A coaching client accountability engine should track client records, session notes, progress, and action items between sessions. It should let a coach review what happened previously and identify patterns before an upcoming meeting. The example also requests four realistic clients at different stages: one getting started, one struggling, one on track, and one performing exceptionally well, supported by credible session histories and tasks.

Q: Why should an AI system be seeded with realistic sample data?

Realistic sample data makes a newly created system feel established instead of empty. In the coaching example, four clients at different stages receive enough session history and action items to demonstrate how the system behaves in real use. This gives a prospective customer meaningful records to inspect and helps show how the databases, tracking process, and meeting preparation features work together.

Q: What tools and accounts are needed to follow this workflow?

The workflow requires a paid Claude account with access to Claude Code and a Lerty account. Claude Code must be installed and available through Terminal or PowerShell after signing in through Claude Desktop. Lerty starts at $24 per month for a builder or power user, according to the transcript, while sellers need the agency plan because it provides organizations for separate client environments.

Summary & Key Takeaways

• Claude Code handles much of the system-building work after receiving a dedicated prompt, a Lerty API key, and a clear description of the desired outcome. Builders can use conversational instructions to create databases, agent structures, dashboards, sample records, and other components for clients in industries such as coaching, law, home services, or marketing.

• Lerty provides the delivery infrastructure for client systems, including separate organizations, customizable agents, relational databases, chat threads, and dashboards. Its agents can connect to models, MCP servers, webhooks, and automation tools. The agency plan supports up to 50 organizations, allowing each organization to serve as a distinct client workspace.

• The demonstrated offer is a coaching client accountability engine. It tracks coaching clients, session notes, progress, and action items between meetings. The requested sample implementation includes four realistic clients at different stages, plus enough historical sessions and tasks to make the environment feel established and useful during a sales demonstration.