Multiplayer by default
Bring teammates and agents into the same operational conversation.
- Roles and permissions
- Shared timeline
- Assignment
- Organization context
THE MULTIPLAYER AI PLATFORM FOR ENTERPRISE
Humans and AI agents.Weeks of work, done in hours.
Bring your team, company knowledge, and connected tools into one shared workspace. Specialized AI agents execute the work, people review the evidence and correct decisions, and approved workflows become reusable capabilities across the enterprise.
Build once. Improve together. Reproduce everywhere.
The Multiplayer Room
- Person acted
- Agent acted
- Waiting on a person
- Approved
- Reproduced
Felipe Matias opens an investigation and passes Contract CTR-2026-0417 to the Revenue Assurance agent. The agent queries CRM, billing, contracts, and email, collects the exports, and assembles an evidence bundle of 12 findings (3 critical), which travels to João Víctor. João opens the supporting files, corrects the renewal-date assumption, and the agent recalculates. An approval request goes to Felipe, who approves; Revenue_Assurance_Report.docx is generated and the workflow is reproduced across Finance, Legal, and Operations. Every document in the diagram opens into a full preview.
Felipe Matias
Operations Director
Approval request
Revenue Assurance Agent
Specialist agent · AI
- CRM
- Billing
- Contracts
Assumption corrected
João Víctor
Financial Analyst
Felipe Matias
Operations Director
Revenue Assurance Agent
Specialist agent · AI
- CRM
- Billing
- Contracts
João Víctor
Financial Analyst
Felipe Matias opens an investigation and passes Contract CTR-2026-0417 to the Revenue Assurance agent. The agent queries CRM, billing, contracts, and email, collects the exports, and assembles an evidence bundle of 12 findings (3 critical), which travels to João Víctor. João opens the supporting files, corrects the renewal-date assumption, and the agent recalculates. An approval request goes to Felipe, who approves; Revenue_Assurance_Report.docx is generated and the workflow is reproduced across Finance, Legal, and Operations. Every document in the diagram opens into a full preview.
Building with Speco
Adapt Imóveis
tday.comAdapt Imóveistday.comAdapt Imóveistday.comAdapt Imóveistday.comConversation is no longer the interface. It is the workflow.
In Speco, a conversation can contain people, agents, tools, evidence, approvals, generated files, and operational history. When the work succeeds, the whole process can be shared and reproduced — not buried inside one person's chat history.
A single-player chat
- One person, one model
- One answer, then it's over
- Context lives in a private history
- No shared review
- Nobody else can run it again
Speco Multiplayer AI
- Teammates and specialized agents together
- Tools, systems, and company context
- Evidence you can open and check
- Corrections and recorded approvals
- A reusable capability the organization owns
THE MULTIPLAYER ROOM
Give every team a room where humans and AI can work together.
Bring teammates, specialized agents, company data, and tools into one shared room. Review the evidence, correct assumptions, approve the outcome, and preserve the workflow for the next team.
Felipe Matias opens a new investigation. The Revenue Assurance agent and João Víctor join the room. The agent queries CRM, billing, contracts, and email evidence, then creates twelve findings and flags three critical. João opens the evidence and corrects an assumption — the renewal date is January 2027, not October 2026 — and the agent recalculates. Felipe approves the result, the room generates Revenue_Assurance_Report.docx, and the workflow is saved for the next team.
AI does not work alone here. People direct it, correct it, and approve it.
WHAT THE PRODUCT DOES
Five things a private chat cannot do
Built for tool-using work
Agents work with databases, company systems, documents, spreadsheets, and approved integrations.
- Tool calls
- Connection status
- Structured results
- Evidence you can open
Correctable and auditable
Review what happened, fix assumptions, approve outcomes, and preserve the operational record.
- Review state
- Comments
- Approvals and rejections
- Audit log
Reusable beyond one chat
Turn successful work into shareable and reproducible agent capabilities.
- Save and version
- Share across teams
- Reproduce
- Reconnect private integrations
Full outputs, not text blobs
Produce professional DOCX, PDF, XLSX, and CSV files, structured data, and system updates.
- DOCX
- XLSX
- CSV
FROM CONVERSATION TO COMPANY CAPABILITY
How work becomes a capability your organization owns
Every stage keeps a person in the loop. Nothing becomes a decision without someone approving it, and nothing worth repeating stays trapped in one person's history.
01
Bring the work
Start with the real task, files, context, and desired outcome.
- Request opened
- Contract store attached
- Contract CTR-2026-0417
- Objective set
02
Assemble the room
Invite teammates and the right specialized agents into one shared room.
- Teammates invited
- Specialist agent added
- Roles applied
- Room scoped
03
Agents use tools
Agents query systems, inspect data, transform files, and execute approved tools.
- Billing records queried
- CRM accounts cross-referenced
- Documents parsed
- Results structured
04
Humans review
See the evidence, assumptions, progress, and generated outputs before they become decisions.
- Evidence opened
- Formula shown
- Source records linked
- Review requested
05Preview
Correct together
Humans fix missing context or incorrect assumptions, and the agent applies the correction.
- Assumption flagged
- Comment added
- Correction applied
- Result updated
Opening evidence, commenting on findings, and validating or rejecting them ship today. In-thread correction that immediately re-runs the agent is in development.
06
Approve the outcome
Finalize reports, analyses, documents, and updates with human control.
- Reviewer approves
- Revenue_Assurance_Report.docx generated
- Audit entry written
- Outcome recorded
07
Reproduce the capability
Save the successful conversation as a reusable agent workflow for the next person, team, or organization.
- Configuration saved
- Reproduced to a team
- Credentials not copied
- New run starts approved
execution canvasrun_8fa2c1
Build once. Improve together. Reproduce everywhere.
When an agent workflow succeeds, your organization can share it, version it, and reproduce it for another team without rebuilding the process from scratch.
Travels with the agent
- Agent configuration is reproduced
- Provenance and versions stay visible
Never copied
- Private organization data stays isolated
- Credentials are never copied
Configuration travels. Credentials never do.
- v1Agent created
- v2Corrected by João Víctor
- v3Approved by Felipe Matias
- João VíctorFinancial AnalystFinance· reconnects own tools & data
SamuelLegal AssistantLegal· reconnects own tools & data
RafaelHR ManagerHR· reconnects own tools & data- Felipe MatiasOperations DirectorOperations· reconnects own tools & data
MaxSoftware EngineerEngineering· reconnects own tools & data
Destination teams reconnect their own integrations. Each branch runs independently — on its own data, under its own approvals.
PEOPLE
Agents do the work. A person still decides.
Speco is built for the roles that have to answer for a number after it ships. Every run ends the same way: a named person opens the evidence, corrects what is wrong, and approves what is right.
Felipe Matias
Operations Director
João Víctor
Financial Analyst
Max
Software Engineer
Rafael
HR Manager
Samuel
Legal Assistant
In every Speco run, the amber step is a person.
REAL FOOTAGE
Watch the product do the work
No mockups in this section — both players run recordings of Speco producing real files and a real specialist agent being built, end to end.
15 seconds inside a Multiplayer AI workflow
3:00Watch the complete Speco walkthroughThree minutes of actual product UI: create a specialist agent, shape its skills, watch the automated build, test it in the Playground, and take the keys to deploy it.
- 0:00Create a specialist agent
- 0:20Instructions, knowledge, and skills
- 1:10Automated build pipeline
- 1:40Test it in the Playground
- 2:10API keys and deployment
RUNNING IN PRODUCTION
Work you can inspect, correct, and run again
REVENUE ROOM
Findings with the evidence attached
Every figure is detector-computed and reconciled. Open a finding to see the formula, the inputs, and the source records it came from — then mark it valid or send it back.
- Findings
- 12
- 3 critical
- Confirmed
- $124,200
- Runtime
- ~60s
- Detectors completed12 / 12
- Awaiting review3
- Marked valid9
APPROVALS
Nothing leaves the room unreviewed
Generated files sit in a review state until a named person approves them. Rejections carry a note, and both outcomes are written to the audit record.
- Revenue_Assurance_Report.docxApproved
- leakage-model.xlsxAwaiting review
- case-package.pdfRejected — missing exhibit
REPRODUCTION
Copy the capability, not the credentials
Reproducing an agent clones its safe configuration only. Private organization data stays isolated and connection secrets are never copied — the destination team reconnects its own.
- Agent configurationReproduced
- Provenance and versionPreserved
- Organization dataNot copied
- Connection credentialsNot copied
BUILT AND RUNNING IN PRODUCTION
From millions of records to reviewable revenue findings.
Speco's Revenue Assurance Agent connects company data, runs deterministic detectors, persists findings, shows evidence, and brings analysts into the same review room.
- Confirmed leakage
- $124,200
- Detector-verified
- Estimated leakage
- $8,000
- Pending confirmation
- Annualized impact
- $132,200
- Projected across 12 months
- Findings
- 12
- 3 critical · 9 customers affected
Production-verified demonstration using a persisted synthetic dataset.
Largest findingCritical
Seat count billed below entitlement
Northwind Freight
$36,000
Formula
(active_seats - billed_seats) * unit_price_cents * months- active_seats
- 240
- billed_seats
- 190
- unit_price_cents
- 6000
- months
- 12
50 unbilled seats × $60.00 × 12 months
evidence · subscription:acc_seats~60-second investigation
ULTRADOC PIPELINES
From shared work to finished deliverables.
Speco agents do not stop at an answer. They assemble structured data, apply organization templates, run document pipelines, request human review, and generate professional files ready for real operations.
- 01Source dataConnected systems and uploaded files
- 02Structured extractionFacts, figures, and references
- 03AnalysisDetectors and domain logic
- 04Human reviewEvidence checked and approved
- 05Template appliedYour organization's format
- 06File assemblyPagination, tables, references
- 07Delivered fileDOCX · PDF · XLSX · CSV
Complete UltraDoc pipelines are included with Enterprise.
- PDF1.2 MB
Legal case package
case-package.pdf
MTR-2026-118
- XLSX88 KB
Financial analysis workbook
leakage-model.xlsx
ACC-1042
- DOCX412 KB
Executive revenue report
Revenue_Assurance_Report.docx
run_8fa2c1
- CSV24 KB
Insurance renewal priority
renewal-priority.csv
POL-4471-A
WORKFLOWS
Complex recurring work, turned into something repeatable
Example workflows, not customer deployments. Each one keeps the specialist in control of the judgment and gives the agent the assembly.
Legal teams
Build case material in hours, not weeks.
Specialized agents assemble, verify, and format case material while your lawyers keep every judgment call.
Lawyers remain in control while specialized agents assemble, verify, and format the work. Speco does not exercise legal judgment.
PDFcase-package.pdf
Example workflow
- 01Upload case records
- 02Extract facts and dates
- 03Cross-reference evidence
- 04Identify missing information
- 05Draft case material
- 06Generate structured DOCX and PDF
- 07Review with counsel
- 08Correct legal assumptions
- 09Approve the final package
Financial groups
Turn disconnected financial data into actionable decisions.
Connect billing and CRM records, reconcile inconsistencies, and produce reports analysts can actually check.
Analysts can inspect formulas, evidence, and assumptions before approving any result.
XLSXleakage-model.xlsx
Example workflow
- 01Connect billing and CRM records
- 02Calculate expenses
- 03Reconcile inconsistencies
- 04Detect revenue leakage
- 05Model return on investment
- 06Produce executive reports
- 07Assign findings
- 08Record approvals
Insurance companies
Find the best opportunities across fragmented client data.
Surface patterns across policy, claims, engagement, and commercial data so account teams can act on the strongest ones.
Agents surface patterns while humans validate the commercial context. Speco does not make underwriting decisions.
CSVrenewal-priority.csv
Example workflow
- 01Analyze customer profiles
- 02Cross policy, claims, and engagement data
- 03Identify high-potential leads
- 04Detect renewal opportunities
- 05Flag missing or conflicting information
- 06Generate prioritized outreach plans
- 07Review conclusions with account teams
Operations and consulting
Convert complex recurring work into a repeatable operating system.
Run the same high-value engagement again without rebuilding the process from scratch each time.
Every deliverable passes a named reviewer before it leaves the room.
DOCXRevenue_Assurance_Report.docx
Example workflow
- 01Collect client inputs
- 02Research the domain
- 03Analyze spreadsheets
- 04Generate recommendations
- 05Create documents and summaries
- 06Request approval
- 07Update internal systems
- 08Reproduce the workflow for the next project
CONTROL
Enterprise control, described precisely.
What follows is what Speco does today. We do not list certifications we do not hold.
Organization-isolated data
Each organization's data, agents, and connections are scoped to that organization.
Role-based permissions
Owner, admin, and member roles govern who can change what.
Auditable activity
Operational events are recorded and reviewable.
Human approvals
Generated outputs can require explicit sign-off before they are final.
Connection-level access
Agents reach systems through connections your organization configures.
Reproduction without copied credentials
Sharing an agent moves configuration, never secrets or organization data.
Reviewable evidence
Findings carry their formulas, inputs, and source references.
Organization-owned billing
The company owns the plan, the seats, and the usage.
Speco does not currently hold SOC 2, ISO 27001, HIPAA, or PCI certification, and does not claim to.
ENTERPRISE
The company buys Speco. The whole team can participate.
Employees use free Speco accounts and join through an organization invitation. Access, agents, knowledge, billing, permissions, and usage belong to the organization — not to each employee's personal subscription.
Who can participate
- Owner
Manages billing, the plan, and organization settings.
- Billing and plan
- Organization settings
- Full access
- Admin
Manages teams, members, and workflows where permitted.
- Invite members
- Manage agents
- Configure connections
- Member
Participates in authorized work — runs agents, reviews evidence, approves outcomes.
- Run authorized agents
- Review and comment
- Approve outcomes
- Invited employeeFree account
Creates a free Speco account and joins through the organization's invitation.
- Free account
- Access via the company plan
- No personal subscription
How billing and access work
- Employees do not need paid plans
- Invited employees create free Speco accounts. Their access comes from the organization's plan.
- The organization owns the capability
- Agents, knowledge, and connections belong to the organization, so they survive any individual leaving.
- Seats are counted per organization
- Starter includes 5 members, Growth 25, and Enterprise 100 by default, with custom capacity available.
- Organization boundaries stay isolated
- One organization's data and connections are never visible to another, including when an agent is reproduced.
PRICING
One price list. Every country.
Start free. Pay when your organization scales.
Free
Free/ forever
Explore the platform. Build your first AI agent.
- 1 AI agent
- 500 MB data ingestion
- 10 pipeline runs / month
- Community support
Starter
$20/ per month
Build and share unlimited AI agents with a small team.
- Up to 5 organization members
- Unlimited agents
- Shareable agents
- Core agent-building capabilities
- Standard knowledge and tool integrations
- Standard execution limits governed by usage policy
Most popular
Growth
$100/ per month
Scale reusable AI agents across a growing organization.
- Up to 25 organization members
- Unlimited agents
- Shareable and reproducible agents
- Higher operational usage allowances
- Expanded organization controls
- MCP integration
- Collaboration and advanced agent management
- Priority support
Enterprise
Starting at $500/month + usage
Deploy specialized AI workers and complete document and operational pipelines across your enterprise.
Contact sales
Need more than the listed capacity? Contact the Speco team
- Up to 100 organization members (custom limits via sales)
- Unlimited agents
- Infinitely shareable and reproducible agents
- Full Revenue Assurance Agent
- Complete UltraDoc document-generation pipelines
- Advanced build pipelines
- Enterprise integrations and MCP connections
- Organization-wide knowledge and context
- Roles, permissions, audit logs, and approvals
- Advanced evaluations
- Priority support
- Usage-based billing on top of the platform fee
The same prices worldwide — every plan is billed in USD.
Invited employees use free Speco accounts
Organization access is provided by the company's plan. Nobody on your team needs a personal subscription to participate.
Unlimited agents does not mean unlimited compute
Paid plans do not cap how many agents you create. Execution, model usage, storage, and premium tools remain governed by your plan's included capacity and applicable usage billing.
What's included
| Capability | Free | Starter | Growth | Enterprise |
|---|---|---|---|---|
| Organization members | 1 | 5 | 25 | 100 |
| Unlimited agents | Not included | Included | Included | Included |
| Shareable agents | Not included | Included | Included | Included |
| Reproducible agents | Included | Included | Included | Included |
| MCP endpoints | Not included | Not included | Included | Included |
| Human approvals | Included | Included | Included | Included |
| Audit logging | Not included | Not included | Not included | Included |
| Revenue Assurance Agent | Not included | Not included | Not included | Included |
| UltraDoc pipelines | Not included | Not included | Not included | Included |
| Advanced evaluations | Not included | Not included | Included | Included |
| SSO / SAML | Not included | Not included | Not included | Included |
| Priority support | Not included | Not included | Included | Included |
Questions worth asking before you buy
What is Multiplayer AI?
A shared environment where people and specialized agents work on the same task together. Agents use your systems and data, people review the evidence and approve outcomes, corrections improve the work, and a successful conversation can be saved as a capability the organization reuses. The alternative — everyone running their own private chats — leaves nothing behind.
Is Speco another chatbot?
No. A chatbot returns an answer to one person and the context disappears. In Speco the work is owned by the organization: agents call approved tools against your data, outputs are real files, a named person approves them, the activity is recorded, and the whole workflow can be reproduced for another team without rebuilding it.
Do employees need individual paid plans?
No. Employees create free Speco accounts and join paid organizations through invitations. Their access comes from the organization's plan.
Are agents unlimited?
Paid plans do not impose a count limit on agent objects. Execution, model usage, storage, and premium tools remain subject to your plan's included capacity or usage billing — unlimited agents is not unlimited compute.
Can agents use our company data?
Yes. You upload documents and configure connections to systems such as billing records, CRM accounts, contract stores, and data warehouses. Agents reach those systems through connections your organization sets up, scoped to your organization.
Can humans correct agent work?
Today you can open the evidence behind any finding, comment on it, validate or reject it, and approve or reject generated files with a note — all of which is recorded. Correcting an assumption in-thread and having the agent immediately recalculate is in development, and it is labelled as a preview wherever we show it.
What does reproducible mean?
An authorized team can recreate an approved agent's configuration and start from it. Reproduction copies configuration and provenance only — private organization data is not copied and connection credentials are never transferred. The destination team reconnects its own integrations.
What outputs can Speco generate?
DOCX, PDF, XLSX, CSV, PPTX, and structured JSON, plus updates written back to connected systems. Complete UltraDoc document-generation pipelines are included with Enterprise.
How is Enterprise priced?
Enterprise is priced "Starting at $500/month + usage" and includes 100 organization members by default. Custom capacity is available through sales.
What security certifications do you hold?
None yet, and we will not claim otherwise. What exists today is organization-isolated data, role-based permissions, auditable activity, human approvals, connection-level access control, and reproduction that never copies credentials. If a certification is a requirement for you, talk to us about timelines rather than assuming.
Weeks of work, done in hours. Owned by your organization.
Build specialized agents, connect company context and tools, review the work with your team, and reproduce what works across the organization.
Felipe Matias opens a revenue investigation. The Revenue Assurance agent queries billing and CRM records and returns twelve findings. João Víctor reviews the evidence and corrects an assumption; the agent recalculates. Felipe approves the result, the agent generates Revenue_Assurance_Report.docx, and the workflow is reproduced to three other teams.