The Use-Case Markdown file tells Synap what your agent does, who it serves, and what it should remember. Synap uses it to generate and continuously refine the Memory Architecture Configuration (MACA) for your instance.
When you create an instance, Synap needs to understand your agent’s purpose to generate a useful Memory Architecture Configuration (MACA) — the configuration that governs what gets extracted from conversations, how it’s stored, and how retrieval ranking works.The Use-Case Markdown file is the primary input for that process. It’s a plain Markdown document you author and upload. The more detail you provide, the better the resulting MACA.
Uploading a use-case file is optional but strongly recommended. Without it, Synap falls back to a generic default configuration that may not match your agent’s actual needs.
Describe what your agent does and the problem it solves. Be specific: include the domain, the workflow, and the end goal.
## Agent ObjectiveOur agent is a customer support assistant for a B2B SaaS platform.It helps users troubleshoot integration errors, understand billing,and navigate the product. The goal is to resolve issues withoutescalating to a human agent whenever possible.
Describe who interacts with the agent — their roles, technical level, and how they typically use it. This helps Synap understand the right granularity for memory scoping.
## Target UsersTechnical leads and developers at mid-market companies (50–500 employees).Users are generally technical but not deeply familiar with our internal systems.Most sessions are one-off troubleshooting requests, but power users returnfrequently and expect the agent to remember their stack and past issues.
Provide 3–5 representative examples of real tasks the agent handles. For each, show a typical user message and what the agent should do. This is the most impactful section — concrete examples let Synap extract the right signal types.
## Task Examples- **User**: "My webhooks stopped firing after I rotated my API key yesterday." **Agent**: Identifies the API key rotation as the likely cause, walks through re-registering the webhook endpoint, and stores the user's webhook configuration for future reference.- **User**: "What's included in the Pro plan?" **Agent**: Answers from the knowledge base and remembers the user is evaluating an upgrade, surfacing relevant context in follow-up conversations.- **User**: "We're migrating from v1 to v2 of your API. What do we need to change?" **Agent**: Provides the migration guide and remembers the user is mid-migration so future sessions can pick up where they left off.
List explicit do’s and don’ts. This section shapes how Synap filters and weights extracted memories.
## Behavioral Guidelines**Do's**:- Remember the user's tech stack and integration details across sessions- Escalate to a human when the issue requires account-level access- Confirm understanding before suggesting destructive actions (e.g., key revocation)**Don'ts**:- Never share one customer's configuration details with another- Don't store payment or billing card numbers in memory- Don't make commitments about roadmap timelines
Clarify who the key actors are in your setup. This directly maps to Synap’s memory scope hierarchy (Client, Customer, User) and helps Synap isolate memory correctly across tenants.
## Role Descriptions- **Client** (you): Acme SaaS Inc — we build and operate the platform- **Customer**: Companies that have purchased a Acme subscription (each has their own workspace)- **User**: An individual employee at a customer company — the person chatting with the agent
List any regulatory constraints, PII handling requirements, or data retention policies. Synap uses this to configure what should and should not be persisted in memory.
## Compliance & Data Sensitivity- GDPR compliance required — EU customers must be able to delete all their memories on request- No PII in memory: names and email addresses should not be stored as facts- No payment data: card numbers, billing tokens must never be extracted or stored- Data retention: memories older than 12 months should be eligible for expiry
Tell Synap what your agent should prioritize remembering. This shapes retrieval ranking and extraction confidence thresholds.
## Memory Priorities- **High**: Integration configuration, past errors and resolutions, product plan and limits- **Medium**: Feature interests, upgrade consideration signals- **Low**: Greetings, small talk, generic product questions with no user-specific context- **Disable**: Emotional state tracking — not relevant for a technical support agent
Anything else that helps — deployment environment, integrations, technical constraints, or custom sections you define.
## Additional Context- Agent is deployed as an in-app chat widget (not a standalone product)- Sessions are typically short (< 10 minutes) but users return frequently- We integrate with Salesforce for account context — agent should defer to CRM data for billing questions- Staging instance shares the same use case but memory should not cross-contaminate production
You are not locked in to the file you uploaded at creation time. You can update the use-case file at any time:
Navigate to your instance in the Dashboard
Go to Settings > Use-Case
Upload a new file and click Save
Synap will re-evaluate the file and regenerate the MACA. The updated configuration goes through the standard approval workflow before taking effect.
Updating the use-case file does not alter memories that have already been stored. It only affects how future ingestion and retrieval behave under the new MACA.
Be specific in Task Examples — vague examples produce generic MACA output. Real user messages and real agent actions give the best results.
Define roles clearly — if your Client, Customer, and User are the same person (e.g., a personal AI), say so explicitly. Synap will collapse the scope hierarchy accordingly.
Name your compliance constraints explicitly — “GDPR” or “HIPAA” is enough; Synap maps known frameworks to the appropriate memory handling rules automatically.
Start with the three required sections — a concise, accurate use-case file with only Agent Objective, Target Users, and Task Examples outperforms a long but vague file with all eight sections.
