How The Wright Gardner Trained Cody on Its Business Knowledge

The Wright Gardner’s work combines a physical service—designing and caring for indoor office plants—with the operational knowledge required to deliver that service consistently. In 2023, the company trained Cody on internal documents and guides so its team could bring more business context into everyday marketing and customer communication.

Evidence note: This case study reflects the company’s reported experience in 2023. It does not assert continued use, a quantified productivity gain, or results beyond the customer statement quoted below.

At a glance

• Organization: The Wright Gardner, an indoor-office plant design and care company serving Northern California
• Challenge: make internal business knowledge easier to apply across communication tasks
• Approach: train Cody on internal knowledge documents, guides, operations, and communication style
• Reported outcome: the co-owner described a more productive and energized team, without publishing a numerical measure

When business context lives in too many places

Most established service companies already have the raw material for an internal AI knowledge base. It may live in onboarding guides, service standards, proposals, plant-care notes, marketing documents, and examples of strong customer communication. The problem is that this knowledge is often fragmented. Employees need to remember where it lives before they can apply it.

Generic AI can draft fluent text, but fluency is not the same as business fit. A useful internal assistant needs the organization’s terminology, processes, audience, and boundaries. The Wright Gardner addressed that gap by grounding Cody in the company’s own material.

How The Wright Gardner trained Cody on its business

The company’s approach began with documents rather than isolated prompts. By adding internal guides and other knowledge resources, the team could give the assistant a shared base of context to draw from across different tasks.

“The team has been invigorated with the boost of productivity the tool has provided, as well as the excitement created by participating in what feels like a new work revolution. Our Cody AI knows a ton about our business, our operations, and our communication style because we’ve trained it using a bunch of our internal knowledge documents, guides, and more and we are adapting it to be useful across more and more internal functions.”

Nick Haschka, co-owner of The Wright Gardner

The statement identifies three useful knowledge layers: what the business does, how it operates, and how it communicates. Together, those layers can help an assistant produce a more relevant starting draft than a model working without company context.

Starting with marketing and customer communication

The original account highlighted marketing and customer-facing messages as early applications. Those are sensible entry points because they are frequent, reviewable, and informed by reusable company knowledge. An assistant might help a team outline a campaign, turn service information into a first draft, or prepare a response that reflects established terminology.

The human remains responsible for accuracy, tone, and the final send. The assistant should not invent a service promise, diagnose a plant from incomplete information, or commit the company to work that has not been approved. Its job is to retrieve relevant context and reduce the effort required to reach a strong first draft.

What the story supports—and what it does not

Haschka reported that the team felt more productive and energized by the experiment. That is meaningful qualitative evidence, but the original story did not provide hours saved, output volume, response-time changes, or a before-and-after benchmark. It would therefore be inaccurate to attach a percentage improvement to this case study.

The more durable lesson is operational: an AI assistant becomes more useful when it is taught the knowledge that employees repeatedly need. That usefulness must be maintained. Documents go stale, procedures change, and communication guidance evolves. A knowledge base needs ownership just like any other operating system.

A practical blueprint for an internal AI knowledge base

1. Pick a bounded workflow. Start with one repeatable task such as drafting customer emails or answering internal service questions.
2. Curate the minimum source set. Add current, approved documents and remove duplicates or obsolete versions before testing.
3. Separate facts from style. Use source documents for business facts and explicit instructions for voice, formatting, and response boundaries.
4. Test representative prompts. Include straightforward questions, conflicting sources, missing information, and requests the assistant should escalate.
5. Require review where it matters. Keep a person in the loop for customer commitments, safety advice, pricing, policy, and other consequential outputs.
6. Assign a knowledge owner. Set a cadence for updates and make it clear which version of a document is authoritative.

Design for reuse, not one-off prompting

The Wright Gardner story is less about finding a clever prompt than building reusable context. Once the knowledge base, instructions, and review process are in place, the same foundation can support multiple approved workflows without asking every employee to reconstruct company context from scratch.

Cody Assist helps teams upload or import business content, organize that knowledge, define response behavior, and share a source-grounded assistant with the people who need it. Begin with a focused use case, measure the quality of the reviewed output, and expand only when the evidence supports it.

Explore Cody Assist to see how your documents can become a practical internal knowledge resource.

Current organization context was checked against The Wright Gardner’s official company page.