Founders, developers, and content operators
Build a marketing website backed by Slab5 CMS
Marketing content often lives in code, spreadsheets, or a separate CMS disconnected from leads and operations.
Use Slab5 CMS for pages, sections, use cases, blog posts, campaigns, and publishing states.
CMSCRMTasksActivity Log
Agent workflow
An agent drafts page sections and blog posts through MCP while humans review and publish.
API workflow
The website reads published CMS entries server-side and falls back to local content if Slab5 is unavailable.
Outcome
The site, content workflow, lead capture, and follow-up tasks share one workspace.
Founders, marketers, and developers shipping a product website without standing up a separate CMS stack.
Build a marketing website backed by Slab5 CMS
Most early-stage teams either hardcode their marketing site in a Next.js repo, plug into a hosted CMS that charges per editor seat, or maintain content in a doc tool that nobody trusts. None of those options play well with AI drafting agents, CRM lead capture, or a single audit trail.
Use the Slab5 CMS module as the backing store for your marketing site. Define collections for pages, sections, blog posts, FAQs, and pricing. Render the site by reading entries through the REST API. Let internal drafting agents propose updates through MCP. Capture forms straight into the CRM module in the same workspace.
cmscrmtaskssupportassets
Agent workflow
Grant a drafting agent an MCP client with `cms:write` and `cms:publish:request` scopes. The agent searches existing entries, drafts new revisions, and submits them for review. Every revision is logged against the agent identity in the audit log.
API workflow
The website server renders pages by calling `GET /v1/cms/entries?collection_id=...&status=published`. New leads from the contact form post to `POST /v1/leads`. A webhook endpoint mirrors `lead.created` events into your internal Slack workspace.
Outcome
One workspace owns the website content, the editorial workflow, the lead capture pipeline, and the audit trail. There is no separate CMS subscription, no separate CRM, and no shadow Google Doc.
Private beta teams and product founders
Capture beta requests into a headless CRM
Beta requests need structured follow-up, qualification, notes, and auditability.
Submit beta forms to Slab5 leads with metadata, UTM context, company, role, use case, and consent.
CRMTasksActivity Log
Agent workflow
An agent can review new leads, summarize intent, and create follow-up tasks.
API workflow
The site posts to `/v1/leads` and creates tasks for high-intent submissions.
Outcome
Lead capture becomes an operational workflow, not a static inbox.
Founders and product teams running a private beta or waitlist who need to triage inbound interest without bolting on a CRM.
Capture beta requests into Slab5 CRM
Inbound beta requests usually land in a typeform export, a Slack channel, or a marketing-tool CRM that the rest of the team cannot reach. Follow-ups slip. Status is invisible. The handoff to onboarding is manual.
Point your beta-request form at `POST /v1/leads`. New leads land in the CRM module with the source, intended use case, and email. A workflow creates a follow-up task in the Tasks module for the on-call founder. Once the beta call happens, the lead is converted to a contact and a deal stage on the onboarding pipeline.
crmtasksactivity_logintegrations
Agent workflow
A triage agent with scoped MCP tools (`create_lead`, `search_companies`, `create_task`, `log_activity`) deduplicates the new lead against existing companies, enriches it with a short summary, and creates a `Reply within 24h` task assigned to the right teammate.
API workflow
`POST /v1/leads` from the website backend. A webhook endpoint listens for `lead.created` and pings Slack. `POST /v1/tasks` creates the follow-up. `POST /v1/leads/{id}/convert` promotes the lead once a deal opens.
Outcome
Every beta request is captured, deduplicated, assigned, and followed up on inside one workspace. The audit log shows exactly who (or which agent) touched each lead.
Teams that want their existing AI assistants to read and write real business data without exposing a raw database.
Connect ChatGPT and Claude to a business workspace
General-purpose AI assistants are useful in chat but blind to operational data: deals, tickets, CMS content, tasks. Plugging them into a workspace usually means writing a custom integration server, copying data into a vector store, or giving the model raw SQL access.
Create a Slab5 workspace, add a scoped MCP client (or OAuth subject grant for org-wide rollouts), and point the assistant at `mcp.slab5.com/v1`. The assistant gets a typed catalog of tools — search contacts, create tasks, draft CMS entries, search tickets, run BI queries — governed by the scopes you grant.
crmcmstaskssupportproduct_analytics_bi
Agent workflow
Grant `crm:read`, `tasks:write`, and `support:read` to the assistant. The user asks `which contacts haven't been followed up on in 30 days?` — the agent searches the CRM and Activity Log, then creates follow-up tasks. Every read and write is recorded against the agent's identity.
API workflow
Server-side automations continue to use API keys against `api.slab5.com/v1`. REST and MCP both operate on the same workspace records — nothing has to be duplicated.
Outcome
Your assistant becomes operationally useful inside the business, with explicit scopes, revocable credentials, and a complete audit trail.
Support and operations leads who want AI to triage tickets without letting it ship customer-facing replies unsupervised.
Build AI-assisted support triage with scoped MCP tools
Inbound support is high-volume and unevenly urgent. Agents spend more time classifying, routing, and tagging than actually replying. Generic AI auto-reply tools either go off the rails or get walled off entirely because the team cannot reason about what they can or cannot do.
Use the Support module for queues, tickets, messages, and SLA targets. Grant a triage agent an MCP client with `support:read`, `support:write:comments`, and `tasks:write` — but not `support:write:messages`. The agent classifies tickets, posts internal triage comments, and creates follow-up tasks. Humans send customer-facing replies.
supporttasksactivity_log
Agent workflow
When a ticket arrives, the agent reads it, classifies priority and queue, posts an internal comment summarizing the request, sets priority, and creates a follow-up task for the assigned human. Every action is auditable.
API workflow
Application code uses `POST /v1/support/tickets` from the contact form and webhook endpoints for `ticket.created`. Dashboards read from `GET /v1/support/tickets` for SLA monitoring.
Outcome
Triage is fast and consistent. Customers still hear from humans. The agent never has the scope to send a message it should not.
AI app builders and enterprise operators
Run governed AI workflows with approval gates and run history
Agentic work is hard to trust when it happens in one-off chats without durable state, approvals, or execution history.
Use AgentGrid to define reusable workflows, connect tools, run manually or on schedules, pause for approval, retry failures, and inspect every step.
AgentGridMCP ClientsAudit LogUsageNotifications
Agent workflow
A manager agent coordinates specialist agents, calls allowed tools, records outputs, and pauses before sensitive side effects.
API workflow
Apps and internal systems invoke workflow runs, inspect status, and correlate request IDs with audit and usage records.
Outcome
Teams can automate repeatable business processes while keeping human review, run logs, cost visibility, and accountability.
Agencies and consultants
Run client operations with one isolated workspace per client
Agencies need repeatable campaign and operations workflows without mixing client data, approvals, assets, CRM context, or reporting.
Use separate Slab5 workspaces for client records, CMS drafts, campaign tasks, assets, approvals, and BI reporting.
CMSMarketingTasksAssetsAgentGridData & Insights
Agent workflow
Agents draft campaign content, prepare reports, summarize assets, and queue approval gates before client-facing actions.
API workflow
Client portals and agency tools read and write workspace records through scoped REST API keys.
Outcome
Teams reuse workflow templates while keeping each client isolated, auditable, and reportable.
Small operations teams that need one place to see leads, deals, tasks, tickets, and webhook health — without paying for five different observability tools.
Build an internal operations dashboard
Operational visibility usually lives in five tabs: CRM dashboard, helpdesk dashboard, task tool, marketing campaign tool, webhook delivery logs. Nobody can answer 'how are we doing this week' without copy-pasting screenshots.
Build a single internal dashboard on top of REST: open deals, open tickets, tasks due today, recent webhook delivery failures, MRR from invoices, and BI metrics on top of everything. Schedule recurring snapshots through the BI module.
crmtaskssupportintegrationsproduct_analytics_bi
Agent workflow
An MCP agent with read-only scopes (`crm:read`, `support:read`, `tasks:read`, `bi:read`) generates a daily standup summary in Slack and creates a Task for anything that's behind SLA.
API workflow
The dashboard calls `GET /v1/deals`, `GET /v1/support/tickets`, `GET /v1/tasks?status=open`, `GET /v1/webhooks/deliveries?status=failed`, and `POST /v1/bi/query` for aggregate metrics.
Outcome
Operations leadership gets one screen for the state of the business. The same workspace powers leads, support, tasks, and reporting — so the numbers actually agree.
Operators, analysts, and AI builders
Turn operational activity into governed BI and data-agent insights
Operational data is often scattered across tools, making it difficult to report on CRM, support, content, usage, and agent execution health.
Use Data & Insights for datasets, dashboards, reports, metrics, queries, exports, analytics governance, and data-agent execution jobs.
Data & InsightsAnalytics GovernanceCRMSupportCMSAgentGrid
Agent workflow
Data agents produce insights, recommendations, dataset versions, lineage, and knowledge artifacts for review and retrieval.
API workflow
Apps run BI queries, create exports, validate event payloads, and connect reports to workspace records.
Outcome
Workspace activity becomes measurable, governable, and reusable across humans, apps, and agents.
Services, agency, and operations teams running real workflows out of Google Sheets and Notion databases.
Replace spreadsheet-based operations with workspace modules
Spreadsheet-based ops do not scale. There is no audit trail, no API for agents, no role-based scope, no webhook fan-out, and no way to enforce uniqueness or required fields. The first time an intern overwrites a column, the workflow breaks.
Map each spreadsheet tab to a Slab5 module or a CMS collection with a typed schema. Contacts and companies move into CRM. Project tasks move into Tasks. Editorial content moves into CMS collections. Custom domain models live in CMS collections with custom fields. Everything is reachable from REST and MCP.
crmcmstasksactivity_logintegrations
Agent workflow
Run a migration agent with scoped MCP tools to read the source spreadsheets (via your own loader), then `create_lead`, `create_contact`, `create_company`, or `create_entry` against Slab5. Use the Activity Log to leave audit trail for the migration itself.
API workflow
Application code now calls Slab5 APIs instead of the Sheets API. Webhook endpoints replace `onEdit` triggers. Schema is enforced server-side instead of by convention.
Outcome
Operational data finally lives in a versioned, typed, audited workspace with both REST and MCP access. Agents and apps consume the same records.
