> For the complete documentation index, see [llms.txt](https://docs.startdeliver.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.startdeliver.io/health-and-metrics/how-to-set-up-health-metrics-and-assessment.md).

# How to set up health metrics & assessment

#### The model

Startdeliver assesses every customer automatically, continuously, and without your team having to do anything manually. The AI reads all available signals and context — product usage, customer feedback, support activity, relationship engagement, and more — combines them with what it knows about your specific business and customers, and produces a plain-language verdict your team can act on immediately.

This is not a scorecard your admins maintain. It is not a weighted formula you tune once and forget. It is an AI that reads everything it knows about each customer and tells you where they stand and what to do next.

***

#### Key indicators

The assessment draws on all the data connected to Startdeliver. The more signals flowing in, the richer and more accurate the picture. Key indicators include:

**Product usage** How well the customer is actually using your product. Measured against goals you define per product — the right level of engagement for your specific use case, not a generic benchmark. Tracks individual users and rolls up to the customer level. → \[Product usage health]

**Customer experience** How the customer feels about your product and service. Driven by survey responses pulled automatically from your feedback tools. The AI reads sentiment patterns across respondents, not just individual scores — identifying potential advocates as well as customers sending an early warning signal. → \[Experience health]

**Support activity** The state of the customer's support relationship — open tickets, delayed tickets, resolution history. Unresolved support issues are often the earliest visible signal of a customer losing confidence. → \[Support health]

**Relationship engagement** The full picture of how your team and your customer are actually interacting — emails, messages, meeting notes, internal comments, Shared Space activity, task completions, document interactions, and collaboration. A customer who is active in the product but silent in every other channel is a different risk profile from one who is communicating regularly and engaged in their Shared Space. The AI reads the presence and recency of relationship activity as a signal of how alive the partnership actually is — not just whether the product is being used.

**Goal progress** Whether your customers are on track toward the business outcomes they are trying to deliver. Goals reflect what your customer sold their own customers — a CRM company's customer goal might be growing sales, an HR company's might be reducing cost-per-hire, an e-sign company's might be faster contract cycles. The AI reads goal progress alongside usage, feedback, support, and relationship signals — connecting product adoption to business trajectory. A customer with strong usage who is falling behind on their stated business goal is a different situation from one who is simply inactive. The AI reads that distinction and flags it.

**And more** As you connect more of your stack — your CRM, your billing tool, your product analytics — the AI has more context to work with. Every connected data source makes the assessment more complete and the recommended actions more precise.

***

#### What the AI does with all of this

The AI does not average signals into a number. It reads everything together as a complete picture of the customer relationship, in the same way a good CSM would if they had time to review every data point on every customer every day.

A customer with strong usage, poor survey feedback, an overdue support ticket, and two months of silence from your team is not a percentage-healthy customer. They are a customer with a specific pattern that calls for a specific response. The AI reads that pattern and tells your team exactly what it sees and what to do about it.