“Which campaigns wasted spend last week?”
A ranked finding with the reporting period, spend and revenue definitions, supporting rows, and a proposed follow-up Task.
Agent
Ask the question. Let the Agent prepare the work.
Use Vendo’s Agent or connect Claude and Codex through subscriptions your team already has. Agents investigate, create reports, and run Automations. People review important Approvals.
Inside the workspace
From a question to reviewable work.
The Agent keeps the answer, evidence, proposed next step, and human review together instead of ending at a block of generated text.
Ask a question in plain language
No SQL, no waiting on an analyst. The Agent answers from your Vendo data, respects your access rules, and asks when a business definition is unclear.
Questions about your data
Ask about attribution, spend, revenue, LTV, jobs, and data quality in plain language — get answers, not exports.
Recommendations with context
The Agent prepares options grounded in your data and documents, with the evidence attached so you can judge them fast.
Product workflows
Refresh data, build reports, investigate results, and run Tasks through conversation. Each Agent only sees the data and tools you allow.
Plain-language requests
Describe the result you need and the Agent drafts the query, model, or report. You review it before it ships.
Practical workflows
Ask for the work—not just an answer.
Vendo maps a business request to the underlying data, definitions, tools, and controls required to complete it.
“Why did paid revenue fall while platform conversions rose?”
A source-aware investigation covering freshness, attribution, conversion delivery, and the assumptions behind the comparison.
“Which customer group should receive a win-back campaign?”
A reusable group definition, estimated size, exclusions, destination requirements, and an Approval before delivery.
“Prepare the weekly growth review.”
Refreshed models and charts, a drafted narrative, visible caveats, and the activity trail for every completed step.
People and agents work from the same Tasks
Assign work to a person, the Vendo Agent, Claude, or Codex. Every handoff, status change, review, and result stays in one activity trail.
People and agent assignees
Assign a Task to a teammate, the Vendo Agent, or a connected client such as Claude or Codex.
Shared workflow
Move work from planned to in progress, review, and complete without losing the brief or evidence.
Complete activity trail
See who created, assigned, changed, reviewed, and completed the Task — person or agent.
Recurring Tasks
Schedule repeatable checks, report refreshes, and customer-group updates with every run recorded.
Approvals keep a human in the loop
Agents prepare the recommendation, evidence, and next step. A person approves, rejects, or asks for revisions before anything becomes an outbound change.
1
Detect
A Task or question surfaces a finding from your data.
2
Prepare
The item lands in your queue with the evidence and context attached.
3
Review
A person approves, rejects, or requests changes. Approved items can continue into execution.
4
Execute
The work runs and records its status, so you always know what happened and when.
Approval queue
Recommendations land in one queue with their evidence, owner, and status — ready for a person to review.
Approve, reject, or revise
Choose what happens next, reject what does not fit, or send the work back to the Agent with feedback.
Memory
Save context once — your goals, naming rules, quirks — and the Agent remembers it in every later conversation.
Human-owned execution
A recommendation cannot become an outbound change until a person approves it.
Built for trust
Every answer shows its work
Check any answer against your workspace definitions, source freshness, permissions, and job history — the evidence is one click away. The same Foundry tools and workspace controls provide the evidence — and you make the call.
Workspace definitions
Answers use your shared metric definitions, and every answer shows which definition and time range it used.
Visible data checks
See the recency, volume, and structural checks behind the data — so you know the inputs are solid.
Your workspace context
The Agent works from your tools, prepared tables, documents, and saved Memory — always within your permissions.
Approvals are human-owned
Agents prepare and recommend. A person decides whether to approve, reject, revise, or act.
Prefer your own AI? Bring it
Vendo exposes its tools over MCP, so Claude, Codex, and other MCP clients can use the same Insights, Automations, and Approvals — always scoped to the permissions you grant, never more.
FAQ