Case Study: CrossCap (CSI)

SOFTWARE / ANALYTICS

Crosscap is a Constellation Software Inc. (CSI) portfolio company — Operating Group: Volaris Group Inc. (Volaris).

CrossCap partners with raia to launch an in-app, guided AI reporting experience across its flagship Distro Pro and Marketer Pro products—turning natural language into reliable business reports via CrossCap APIs and Metabase, and laying the groundwork for safe, auditable agentic actions.

Snapshot (planned targets)

Metric	Before raia	After raia (target)
Data accessibility	Reports require technical expertise or developer help	Self-serve reporting for business users via natural language + guided templates
Time to insight	Hours/days depending on expert availability	Seconds/minutes for common questions (targets set after baseline review)
User adoption	Adoption limited by “blank-slate” reporting complexity	High adoption driven by prebuilt templates + parameter prompts + conversational drill-down
Developer effort	High effort building one-off reports	Reduced custom-report burden; focus shifts to core platform + API enhancements
Security & auditability	High bar required	Tenant isolation + audit logs + SOC 2-aligned controls and observability

Note: This case study describes the planned solution and target outcomes. Measured results can be added once CrossCap’s rollout produces post-launch reporting.

1 | Company & Challenge

CrossCap (CSI) builds two products in scope: Distro Pro (flagship) and Marketer Pro (Calendar). CrossCap’s objective is to ship AI capabilities quickly—targeting Distro Pro live this year (ideally late November)—while preserving correctness, security, and a high-adoption user experience.

As CrossCap evaluated how to bring AI into reporting workflows, five compounding challenges emerged:

1. Technical barrier to insight — business users can’t easily self-serve answers and rely on experts, creating delays and bottlenecks.

2. High cost of custom reporting — developer time gets pulled into one-off report building instead of core product work.

3. Adoption risk of “blank-slate AI” — an open prompt alone tends to underperform; CrossCap preferred a guided crawl-walk-run approach (templates + parameters first).

4. Architecture constraints (API-first) — the database is “raw” and production views are composed in backend logic; direct DB querying is not the right path.

5. Multi-tenant + metadata complexity — each client has many user-defined attributes; the AI must read attribute metadata and ask clarifying questions, while enforcing strict tenant isolation (per-client servers/endpoints + API keys).

CrossCap needed a solution that could translate natural language questions into actionable, trustworthy reports, integrate seamlessly with an API-first + Metabase stack, and drive adoption through a guided UX.

2 | Why CrossCap chose raia

CrossCap selected raia to rapidly deploy an AI reporting layer that’s powerful, safe, and user-friendly:

• Guided, high-adoption experience: prebuilt report templates + parameter inputs first, then conversational drill-down—explicitly avoiding the “blank slate” problem.

• API-first integration: NL → CrossCap APIs (with server-side SQL generation) + Metabase for charts—no direct DB access.

• Phased fast-to-market rollout: start with Distro Pro reporting, then a low-lift Marketer Pro NL-to-CQL copilot, then agentic actions.

• Security and auditability: SOC 2 readiness, UAT access, logging, and traceability are built into the engagement expectations.

3 | Solution Design

The solution is a phased deployment of AI copilots integrated directly into CrossCap’s products—starting with read-only insights and evolving toward safe, auditable action.

Component	Role	Key Integrations
Distro Pro AI Reports & Insights (Phase 1)	Converts natural language into CrossCap API calls across assets, jobs, locations, orders; returns tabular results + Metabase visualizations; ships with a curated library of prebuilt report templates and conversational drill-down.	CrossCap Distro Pro APIs (incl. AI endpoints), Metabase, CrossCap auth
Metadata-aware clarifications	Uses system + customer-defined attribute metadata (labels, types, values) to resolve ambiguity; asks clarifying questions when needed.	Attribute metadata endpoints / data dictionary
Marketer Pro Calendar NL-to-CQL Copilot (Phase 2)	Converts natural language into CQL for Calendar (“blocks”), inserts into query builder for user refinement and execution.	Marketer Pro CQL interface + API
Distro Pro “New Location Wizard” Pilot (Phase 3)	Agentic copilot that can simulate then (with human approval) create orders for new store openings, with full auditability and idempotency controls.	Order creation + calculator APIs; webhooks/callbacks (as needed)

Example questions (Distro Pro)

• “What location is creating the most orders over the last 30 days?”

• “How many jobs do I have?”

• “Show order volumes by location for last quarter.”

Guardrails (by design)

• Scoped objects/fields (allowed functions only) + strict operator controls

• Metadata-driven clarifications for custom attributes and missing constraints

• Multi-tenant isolation (per-client servers/endpoints + per-key context; no cross-tenant leakage)

• Audit logs + history retention (traceable prompts → parameters → API calls → outputs)

• Rate limiting + error handling and safe fallbacks

• Read-only by default; human-in-the-loop for writes (Phase 3), with idempotency for safe order creation

4 | Deployment Timeline (phased)

1. Phase 1 — Distro Pro AI Reports & Insights (8 weeks): deliver NL-to-API reporting + 8–12 starter templates + UI embed + UAT. Target readiness for a client-facing update by late November (subject to contract start).

2. Phase 2 — Marketer Pro NL-to-CQL Copilot (4–6 weeks): strict CQL scoping, builder insertion, clarifications, pilot.

3. Phase 3 — Agentic Order Creation Pilot (6–8 weeks discovery/prototype): simulation-first, then guardrailed writes with human approval and full audit trail.

5 | Expected Impact (targets; measured results to follow)

CrossCap’s target outcomes include:

• Democratized data access: business users can self-serve answers with natural language and guided templates.

• Accelerated time-to-insight: reduce the lag from “question” to “answer,” especially for common operational reporting needs.

• Reduced developer overhead: fewer one-off reports; developers focus on platform work and API enhancements.

• A secure foundation for agentic AI: safe, auditable progression from read-only insights to action-taking workflows.

6 | Lessons for Prospective Clients

1. Solve adoption first: an open-ended prompt isn’t a strategy—templates + parameter prompts are the fastest path to early wins.

2. Correctness beats shortcuts: when production logic lives in the backend, the most reliable approach is NL → API, not NL → direct DB.

3. Metadata is the difference-maker: custom attributes require strong definitions or smart clarifying questions—build this in from day one.

4. Design for action later: start read-only, but architect for safe writes with simulation, human approval, and idempotency controls.

7 | What’s Next

After the initial Distro Pro rollout, CrossCap and raia will:

• Expand the template library and observability for Distro Pro reporting.

• Ship the Marketer Pro NL-to-CQL Copilot as a fast-follow, lower-lift feature.

• Productionize the New Location Wizard with simulation mode, audit trails, and human-in-the-loop confirmations before enabling broader write-path automation.

8 | Key Takeaway

CrossCap is building a phased, API-driven AI layer across Distro Pro and Marketer Pro—starting with a guided natural-language reporting experience (NL → API → Metabase) to democratize insights and reduce developer overhead, while laying a secure and auditable foundation for future agentic workflows like order creation.