How to use the Build Business Case agent skill

What the skill makes the agent do

An agent skill is a folder of instructions and reference files that an agent loads when a task matches the skill's description. This one covers business-case work tied to a specific customer, workflow, initiative or decision. The agent starts from the customer's own evidence, such as notes, metrics, CRM records and prior value cases, rather than from product features.

The skill maps the workflow to value buckets: Enhanced Productivity, Cost Reduction, Risk Reduction, Revenue Acceleration and Time to Market. Every material number is labeled Known, Inferred, Assumed or Missing. If the evidence is too thin for real math, the agent produces a clearly labeled structural case and a short list of what to validate.

• Works from the customer's metrics, discovery notes and account records first
• Maps each use case to the fewest applicable value buckets
• Shows formula, inputs, labels and low, base and high scenarios for finance-ready cases
• Does not handle pure presentation design or CRM writeback

Install the build-business-case skill

Download the ZIP for this skill from this page. Then choose one of the two install options below.

Option 1, personal install: unzip the folder into ~/.claude/skills/ so the skill is available across your projects.

Option 2, project install: unzip the folder into the .claude/skills/ directory inside a specific project, so the skill applies only there.

Option 3, Claude apps: upload the skill in the Claude apps under Settings → Capabilities → Skills.

• The unzipped skill folder should contain SKILL.md, agents/openai.yaml and a references/ folder
• Restart or start a new session if the agent does not pick up the skill immediately

Trigger the skill with natural requests

You do not need a special command. Describe the customer and the decision in ordinary language and say what you want produced. The skill's description covers requests for business cases, ROI or value models, pricing or investment rationales, executive summaries and customer-ready value stories.

The more account context you include, the stronger the case. Name the customer, the workflow, the buyer and any numbers you already have.

• Build a business case for [customer] on their claims processing workflow using the discovery notes I pasted.
• Draft an ROI model with low, base and high scenarios for the pilot with [customer].
• Write an executive summary of the value case for [customer]'s CFO, based on our account plan.
• Explain the pricing rationale for the expansion path at [customer] and label which inputs are missing.

What to provide for the best output

The skill is built to work with uneven inputs. It can start from pasted notes, links, exports or a clearly anchored thread. Useful inputs include discovery call transcripts, customer metrics, CRM summaries, product usage data, the customer website, annual reports or investor presentations, internal account strategy materials and prior internal business cases.

If the customer and workflow are both unclear, the agent asks one focused clarifying question before going further. If the customer is ambiguous, it offers up to five candidates and asks you to choose.

• Customer or account name and the workflow or decision in question
• Any metrics the customer has confirmed
• Pilot amount, expansion path, budget range or close timing if you have them
• Access to CRM, meeting transcripts or knowledge files, where your setup allows it

What the skill needs and what it does on your machine

The skill's dependencies are optional or blocking depending on the case. CRM access is blocking for customer-specific cases unless authoritative account and commercial context is already grounded. Meeting transcript access is optional for discovery evidence. Knowledge and files access is optional for account plans and prior value cases. Sales intelligence and public web research are optional and used only to close specific gaps.

Safety notes: the skill runs no scripts. It does use the network, because it can reach CRM, transcript, knowledge and public research sources when those are connected. Public research is used to sharpen strategic priorities and executive wording, and it never replaces customer-confirmed metrics.

• Runs scripts: no
• Uses network: yes, through connected CRM, transcript, knowledge and research sources
• Public and enrichment context cannot substitute for customer-confirmed pain or metrics
• Review outputs and validate numbers before anything goes to a customer

Tips for better value cases

Start with the strongest account artifact you have, such as an account-specific value case or a pilot target. The skill looks there first and stops once the core case is supported.

Treat the labels as a working list. Each Assumed or Missing input is a validation task, and the skill lists the smallest questions that would make the model finance-ready.

Keep the customer's own words and metrics in the inputs. The skill ranks customer-provided metrics highest in its evidence order.

• Paste the customer's own metrics rather than summaries of them
• Ask for low, base and high scenarios when you need a finance-ready model
• Ask for a structural case first if the data is thin