LAYER 01 · Mission Design
Intent → Workflow
Describe what you need. Bixie builds it.
Tell Bixie what you want automated in plain English. The AI analyses your objective, matches connected tools, builds the execution plan, and creates the mission through conversation. No forms, no drag-and-drop, no YAML.
01
Natural Language Builder
Describe what you want automated in a conversation. Bixie generates the full workflow with the right agents, steps, and triggers.
02
Live Tool Access
Pull CRM data, check pipeline status, send messages. Chat queries can call your connected tools directly, not just answer questions.
03
Grounded in Your Documents
Answers are sourced from your uploaded documents and synced knowledge bases, with citations. No generic or hallucinated responses.
04
Voice Input & Output
Speak your requests and hear responses read back. Hands-free operation with speech recognition and text-to-speech.
Bixie Chat
builder.sessionU
When a new lead comes in from HubSpot, enrich it with Apollo, score it, and route hot leads to the right sales rep on Slack.
B
I'll create a mission with these steps:
- 1Pull new lead details from HubSpot
- 2Enrich with company data from Apollo
- 3Score lead and assign priority tier
- 4Notify assigned rep on Slack with briefing
Trigger: HubSpot webhook · 3 integrations detected
Describe what you want to automate…→
LAYER 02 · Agent Execution
AI Workforce in Motion
Multi-step automation with human oversight.
A mission is a multi-step workflow that chains specialist agents together, each with its own tools, permissions, and approval gates. Define once, trigger from anywhere, and monitor every step as it runs.
Mode 01
Step-by-step
Predictable workflows with deterministic ordering. Best for finance, HR, and compliance.
Mode 02
Adaptive
Sandboxed agents handle code, write tests, and self-correct. Best for engineering tasks.
Mode 03
Conversational
Domain experts answer questions in natural dialogue. Best for advisory and research.
01
Specialist Agent Roles
Assign agents as executors, reviewers, auditors, validators, or monitors. Each has scoped permissions: they can only access the tools you grant them.
02
Flexible Triggers
Run manually, on a schedule, from a webhook, from an inbound email (Bixie classifies which mission to run), or when a Studio document is ready.
03
Automatic Data Passing
Data flows between steps automatically. Map outputs from one agent into inputs for the next, and every step gets exactly the data it needs.
04
Human Approval Gates
Insert checkpoints where humans review and sign off. Pause execution until approved, with full audit trail and reasoning.
LAYER 03 · Content Studio
Professional Outputs
An AI workspace for structured deliverables.
Studio is where strategy becomes action. Generate research briefs, board decks, compliance audits, sales proposals, and 15+ other document types, then hand them off to missions for automated follow-through.
19 document templates
Finance
Financial Analysis
Quality
Data Quality Report
People
Onboarding Playbook
Procurement
Vendor Assessment
19 Document Types
Feature plans, research canvases, slide decks, sales proposals, incident postmortems, financial analyses, campaign briefs, and more, each with built-in structure.
Documents That Trigger Action
A compliance audit in Studio can automatically trigger a remediation mission. Strategy documents become operational inputs, not shelf-ware.
Custom Templates
Define document types for your organisation's specific needs. No code required — just describe the structure and outputs you need.
LAYER 04 · System Integrations
Enterprise Connectivity
Connect once. Use everywhere.
Every integration uses a universal connector standard. Connect your tools once and every agent, mission, and chat session can use them. Credentials refresh automatically, health is monitored, and adding new tools takes minutes, not sprints.
SSalesforceHHubSpotSSlackTMicrosoft TeamsGGmailOOutlookJJiraLLinearGGitHubXXeroQQuickBooksSStripeOOktaAAWSGGoogle CloudNNotionCConfluenceSSharePointPPagerDutyIIntercomSShopifySSAPDDocuSignWWorkdayAApolloSSnowflakeDdbtBBloomberg+ any custom integration
01
Universal Connector Standard
One connection model for every tool: SaaS apps, custom APIs, and internal systems. Add new integrations without platform changes.
02
Always-On Credentials
Login tokens refresh automatically in the background. No expired credentials, no broken workflows at 3 AM.
03
Team & Personal Accounts
Shared credentials for team-wide tools, personal credentials for individual accounts. All stored securely with strict access controls.
04
Automatic Tool Discovery
When you connect an integration, all its available actions are surfaced instantly. Agents pick tools by name with no manual configuration.
LAYER 05 · Knowledge Management
Cited, Screened, Trusted
Your documents become AI-ready context.
Upload files, connect your wikis and repos, and every chat response and agent run gets grounded, cited answers from your organisation's actual documents, not generic training data.
↑
Ingest
- Upload · Confluence
- GitHub · SharePoint
- Notion · Web URLs
→⚙
Process
- Parse 12+ file types
- Auto-tag & classify
- Screen sensitive content
→⌕
Index
- Smart section splitting
- Semantic search index
- Keyword search index
→≡
Retrieve
- Combined smart search
- AI-ranked results
- Source citations [1][2]
01
Sync Your Existing Sources
Connect Confluence spaces, Notion pages, GitHub repos, OneDrive/SharePoint libraries, and web docs. Choose exactly which folders to include. Stays current hourly, daily, or weekly.
02
Intelligent Search
Combines semantic and keyword search. Prioritises recent documents and latest versions. AI-powered deep search available for complex questions.
03
Sensitive Content Protection
Before anything is indexed, AI screens for personal data, HR records, medical information, legal documents, and financial details across 10+ categories.
04
Per-Expert Knowledge
Give an Expert exclusive access to specific sources. Domain isolation without separate systems. Documents only surface when that Expert is asked.
LAYER 06 · Expert Advisory
Specialized Intelligence
Domain specialists you can talk to.
Agents do the work. Experts know the domain. Build conversational specialists with custom personas, grounded knowledge, voice, and guardrails. 39+ ready-made templates across finance, legal, engineering, security, and more.
Multi-Expert Session
Debate ModeC
Contract Expert
Analyse non-standard clauses
C
Compliance Expert
Flag regulatory risks
T
Tax Expert
Evaluate withholding implications
Contract ExpertClause 7.2 contains a non-standard liability cap at 2x contract value. Industry standard is typically 1x. [1]
Compliance ExpertThe data residency clause doesn't specify GDPR adequacy requirements for cross-border transfers. Flagging as high risk.
01
Custom Personas & Voice
Configure tone, expertise depth, citation style, and refusal behaviour. Each Expert sounds like the specialist your team would hire.
02
Structured Debate
Multi-Expert sessions in parallel, sequential, or adversarial mode. Surface disagreements, weight by confidence, generate a synthesised summary.
03
Four Ways to Feed Knowledge
Sync from Confluence/GitHub/web, upload from media library, paste policies and runbooks directly, or point at URLs that Bixie crawls on a schedule.
04
39+ Ready-Made Templates
Finance, legal, HR, security, engineering — start from a template and tune the persona, knowledge, and citation strictness to match your standards.
LAYER 07 · Operational Governance
Audit + Control
Full visibility into every execution.
No black boxes. The Control Panel shows every step of every run: what happened, what data was used, how long it took, and what it cost. Watch in real time, search historical runs, and step in to pause or redirect workflows as they execute.
Customer Onboarding · Run #623
Duration 34.1sCost $0.031● Completed
Create WorkspaceSalesforce → new account setup
3.4s
$0.005
Provision AccessOkta → create user group
2.1s
$0.004
Generate Welcome PackStudio → onboarding guide
18.2s
$0.012
Schedule KickoffGoogle Calendar → book meeting
1.6s
$0.003
Send Welcome EmailGmail → personalised intro
2.8s
$0.007
01
Step-by-Step Trace
See exactly what each agent did, in order. Compare the planned execution against what actually happened, with full visibility into every input and output.
02
Cost Breakdown
See cost per tool call, per agent, and per mission. Know exactly what each automation costs to run, and where to optimise.
03
Live Intervention
Pause, resume, or stop any running workflow from the dashboard. Inject corrections mid-run when something needs adjustment.
04
Schedule Overview
See all upcoming automated runs, what will trigger them, and when they're expected to fire next.
