Agentic Coding Workshops
AI coding agents can take real load off your Salesforce team across Apex, Flows and configuration. For their output to fit your org, they need clear guidance on your conventions and processes. In our workshops, your team works out that guidance with us. Afterwards, we accompany your team until it works with AI coding agents confidently and on its own.
Why does your team need this workshop?
AI coding agents write code that compiles but often does not fit your org. That code often follows a different trigger framework, unfamiliar naming conventions or an older Flow style. Your team then reworks that code by hand, and much of the time saved is lost again. The harness supplies the missing context. It holds the specs, skills, rules, subagents and hooks that define how the agent works in your org and which actions are blocked.
Your team builds this harness with us and develops it on its own afterwards. It turns what your Salesforce admins and developers already know about the org into context an agent can use.
What does working together look like?
1–3 team sessions
Lay the foundations
We show how AI coding agents work and what context they need, and we explain spec-driven development. Then we walk through your development and deployment process together, from sandboxes and source control to reviews and releases, and agree where agents can start first.
As needed
Build the harness
We define the artifacts between the steps, from business requirement and technical specification to implementation and review, and keep business requirements apart from technical specs. Then we write the skills, rules, subagents and hooks for your org with your team and try the harness live in your sandbox on real tasks from your team.
Ongoing, as needed
Accompany
We support your team while it implements with AI coding agents. The harness grows from your team's experience, takes on new use cases and, if you book Recurring Platform Reviews, follows their findings. Your team writes the changes, and we coach and review.
This agenda is an example. We agree in the first call whether one session is enough or five make sense.
What do the workshops cover?
At the centre is the harness, the files in your repository that guide a coding agent while it works. The building blocks below are typical topics, not a fixed list. Which ones your team needs depends on your org, and because the tools keep evolving, we update the topics regularly.
Which coding agents is the workshop available for?
Focus
Claude Code
Our focus. The content and examples of the workshops are designed for Claude Code.
Standard
Codex
If your team uses Codex, we tailor the workshops to it. Where Codex lacks a building block or names it differently, we show the closest equivalent.
On request
Other runtimes
If your team works with another coding agent, we check on request whether we can offer the workshops for it. Which building blocks are available varies by agent.
What does your team take away?
Our free, open-source Flow plugin shows in a small example what individual building blocks of a harness look like in practice.
Common questions
No, but we recommend it. The assessment shows up front which patterns and conventions apply in your org, so we align the harness with those findings from the start. Without it, your team starts the first session with a short inventory of the org instead.
No, your team implements. We accompany it by coaching, reviewing the changes and developing the harness together with your team, so the knowledge stays with the people who work on your org every day.
Everyone taking part needs access to a coding agent, typically Claude Code or Codex, under your own subscription or API account. Your team also needs a sandbox and source control for your Salesforce metadata. If your team does not use source control yet, we offer an additional session beforehand that walks through the basics together.
Your team works with its own accounts, so your company's agreement with the provider applies. What reaches the model provider is what the agent reads while it works, such as the files in your repository, your org's metadata and the prompts your team writes.
Your company's policy decides which data the agent may read. In the first session, we align the harness with that policy.
Yes, a single foundations session is a sound start, and you decide afterwards whether to build the harness further. The sample agenda above is an example, not a package.
What does the scope depend on?
We set the scope in the first call. No two teams start from the same place, so we agree the number of sessions and the depth of the accompaniment with you rather than selling a fixed package.
What shapes the scope
- Team sizeHow many Salesforce admins and developers take part, and whether they work in one team or several.
- Number of sessionsWhether one foundations session is enough or the harness grows over several sessions and months of accompaniment.
- RuntimesWhether your team uses Claude Code, Codex or another coding agent.
- Your deployment processHow your team moves changes today, from sandboxes and source control to reviews and releases, and how much of it the harness should cover.
- With or without an assessmentWhether a Platform Assessment already exists that the harness can build on.
Our recommendation
We suggest starting with the Platform Assessment. The assessment captures the current state of your org across 15 areas, such as automation, Apex code and integrations. On that basis, we align the harness with your org's findings instead of opening the first session with a generic example. Without it, your team starts the first session with a short inventory of the org.
Let's talk about how your team can use AI coding agents
Book a free 30-minute discovery call. We'll look at your current setup together and discuss what scope makes sense for your org. You'll then receive a proposal tailored to your situation – with no obligation.
If you'd rather start smaller, a free 60-minute quick check looks at your org from an AI agent's side first.