Roadmap
What’s live, and what’s next.
Brigham.ai ships as one platform. Here is the capability map, and the extensions we are building on top of it.
The platform
Ten capabilities, live. Extensibility is next.
Everything below marked Live is part of Brigham.ai today, built and used inside our own operating company. What is marked Next extends the platform so you can shape it yourself.
The capability map.
The foundation
Sign in through erp.io and talk to a chief of staff that answers from live data, with every step audited and costed.
- Persistent conversations that survive restarts
- Read-only tools for CRM pipeline and system health
- Every tool call audited, every run priced
- Protected-domain mechanism proven by tests
Your day
Gmail, Calendar and the whole CRM, with a morning briefing for each person.
- Connect your own accounts, read-only first
- Meeting prep and follow-ups owed
- Replies saved as drafts until you approve sending
- Your inbox stays private to you, admins included
Business memory
Documents, decisions and relationships Brigham.ai can cite and keep current.
- Answers that name their sources
- Nothing becomes fact without a person or verified source
- Conflicts flagged, facts carry a last-verified date
- A review queue for what Brigham.ai thinks it learned
Integrations
Drive, Docs, Sheets, GitHub, hosting, your CMS, search and analytics, and ad data.
- Built on one connection framework
- Any service with an MCP server can connect
- Each connection scoped and revocable
Automation and approvals
Schedules, triggers and workflows that pause for a person. Sending email is switched on per action, behind approval.
- Cron schedules and event triggers
- Multi-step workflows with approval pauses
- Sensitive actions only behind approval
- Proactive work items that recommend, not act
Specialist agents
The chief of staff delegates to sales, marketing, dev, ops and research agents.
- Each agent has its own tools and permissions
- Delegation never widens permission
- Cost shown per agent and per task
Portfolios
Portfolios link a parent company to its subsidiaries’ separate workspaces.
- Consolidated reports that cite their source company
- Cross-company tasks authorized in each company
- Adversarial tests for leakage between unrelated orgs
Compute and software
Sandboxed execution for building files, running code and shipping changes with approval.
- Isolated per-workspace execution with egress controls
- Reports, spreadsheets and charts as files
- Branch, code, test, pull request, deploy on approval
Proactive intelligence
Anomaly spotting and recommendations that show their evidence.
- Traffic, pipeline and deploy anomalies
- Each recommendation shows evidence, impact, cost and confidence
- Prompts, skills and policies versioned, never changed silently
Teams and administration
Seats, roles and customer admin, with data export and deletion under your control.
- Branded brigham.ai deployment
- Admin controls for agents, integrations and protected domains
- Data export and deletion on request
Build your own agents
A custom agent builder, plus skills and workflow templates you can reuse across workspaces.
- Define an agent’s tools, permissions and budget
- Reusable skills and workflow templates
Open platform
An integration catalogue and a public API with webhooks and scoped keys.
- A browsable integration catalogue
- Public API and webhooks with scoped keys
Enterprise scale
Department agent teams, enterprise approval hierarchies and private-model deployments.
- Agent teams organized by department
- Approval hierarchies for larger companies
- Private-model deployments
How we ship
The same five steps, every release.
Audit
Look at what exists. Write the checklist down before touching anything.
Build
Build only what the checklist covers. No reaching ahead.
Test
Tests are written so that bringing a bug back fails them. A skipped suite counts as a failure.
Document
The docs say what is done and how it behaves.
Then stop
We report, and wait for the go-ahead before the next release begins.
Questions about the roadmap
Is Brigham.ai available today?
Yes. The full platform is available through onboarding, and it runs inside our own operating company every day.
What does Next mean?
Next marks the extensibility work: custom agents, templates, the public API and larger-company controls. The capabilities marked Live are in the product now.
Can Brigham.ai send email?
Yes. Sending email and the sensitive CRM actions are switched on per action, each behind approval.
Can I ask for something?
Yes. Customer feedback decides what we build next. Use the get-started form and tell us what you need.
Shape what comes next.
Customer feedback decides what we build and in what order.