AI Agents & Automation

Keep commercial work moving.
Put agents and automation to work.

Respond to changing costs, investigate exceptions and route actions for approval, with workflows and AI agents operating within your business rules.

From signal to action

Defined steps. Reasoning where it helps.

Follow a cost update through checks, investigation and a proposed response, with approval at the required checkpoint.

Automations · Cost pass-through, Northeast Distribution
Illustrative workflow built from supported triggers and actions
What this run producedAwaiting approval
  • 1,842 items checked against the refreshed cost
  • 137 past the 2% threshold, stamped for review
  • 1 finding saved as a note, citing the records it used
  • 1 price list submitted, waiting on the people its policy assigns

Submitted is not approved, published, or recovered margin. Illustrative figures.

Read the workflow as text
  1. Receive the update. Triggers are manual, event or schedule. An event trigger answers every event of its kind in the workspace. The run starts and opens a run record.
  2. Identify affected items. A data-change condition watches a key column and asks what moved, with a threshold in absolute or percentage terms. It has two outputs, not one. Items past the threshold take the Yes path. The rest take the No path and are marked, so a checked item is distinguishable from one the run never reached.
  3. Mark the rest within tolerance. A condition has two outputs, so the items that fail it still go somewhere you chose rather than nowhere. Those items are stamped as checked, which is what lets the next run tell a cleared item from one it never reached.
  4. Investigate the response. The agent plans and calls tools. It chooses what to look at; its tools check the automation’s permissions on every call, so a step that only reads needs no write rights. A written finding, and a note saved through the agent’s own create_note tool. The run record keeps the prompt, the reply and the tools it used.
  5. Prepare the action. This is a step you draw, not something the agent’s answer becomes. Nothing converts a reply into a pending price change on its own. A set-field step stamps the records the finding names, so the next step has something definite to act on.
  6. Submit for approval. A transition names a verb rather than writing a state. The policy decides whether that verb is legal from where the record actually stands. The price list moves to Submitted and waits for whoever the policy assigns. It is not approved, published, or recovered margin.
Delegate

Choose how the work gets done.

Automation

Follow defined rules and steps.

Use for repeatable checks, updates and routing.

ExampleFlag a cost change and route a review.

Agent

Investigate a goal using permitted tools.

Use when the work requires examining information and deciding what to inspect next.

ExampleExplain a margin decline with supporting evidence.

Combined workflowalso called Hybrid

Add investigation inside a defined process.

Use when routine steps need an agent-assisted assessment before the next action.

ExampleAssess affected accounts before preparing a pricing response.

Compare them side by side
PropertyAutomationAgentCombined workflow
Who chooses the pathYou do, on the canvas, before it runsThe agent does, at runtime, a step at a timeYou draw the path and hand one step over
RepeatabilityFollows configured rules and branchesReasons toward the goal, so the route can differDefined either side of the reasoning step
Starts onA manual run, a system event, or a scheduleSomebody asking, or a step inside a workflowWhatever starts the workflow
Runs onPrice lists, rule sets, datasets, views and the newsfeed
GovernanceThe same for all three: the creator’s permissions, your lifecycle policy, one run record per execution
How record-scope processing works

A condition identifies a subset of records rather than testing them one at a time, and a supported action then operates on that subset, so a flag across many items is one write rather than one per item. Two limits are worth knowing. A step that works on each member separately, such as moving a family of records through a lifecycle, takes on at most 500 in a run, because each member is its own round of reads and writes inside a single worker. A step that does one piece of work over the whole set is not capped that way. Record counts still affect how long the underlying queries take.

Inspect

See what happened, and what needs attention.

An agent investigation ends in something you can read and check, not a wall of tool calls. The calls are still there when you want them.

Revomo Copilot · margin investigation, Northeast, Q2
Goal“Margin in the Northeast is down about two points this quarter. Find out what caused it and show me the customers behind it.”
FindingRun complete

Margin fell 2.2 points. Landed cost accounts for nearly all of it, concentrated in three suppliers, and eleven customers carry most of the exposure.

Contribution to the margin movement, in points
Cost−2.4Landed cost, freight surcharge
Mix+0.3Shift toward higher-margin lines
Price−0.1Realized price broadly held
Net−2.2Points of margin, Q1 to Q2
Sources
  • Northeast Price Book, Q1 and Q2
  • Landed Cost 2026, with its provenance trace
  • Margin by customer, 1,842 items

Saved as a note, citing the records above. A note is a draft artifact: nothing was published and no price moved.

  1. 01Locate the relevant recordsNortheast price book, Q1 and Q2
  2. 02Compare margin across periods1,842 items
  3. 03Analyze price, cost and mixCost −2.4 pts, mix +0.3, price −0.1
  4. 04Trace contributing changesLanded cost, freight surcharge, 3 suppliers
  5. 05Identify affected customersTop 11 carry 68% of the exposure
  6. 06Prepare a sourced findingNote saved, nothing published
Inspect execution details

Illustrative execution using supported tools. Every tool named below exists in the product; the run itself is a composition for this page, not a recording.

  1. 01search_pricing_modelsreadNortheast price book, Q1 and Q2
  2. 02engine_query_dataread1,842 items
  3. 03design_new_analysisreadCost −2.4 pts, mix +0.3, price −0.1
  4. 04check_data_provenancereadLanded cost, freight surcharge, 3 suppliers
  5. 05engine_query_datareadTop 11 carry 68% of the exposure
  6. 06create_notedraftNote saved, nothing published

Calculations are performed by the query and analysis tools, not by the language model. The model decides which tool to call and writes up what came back.

Delegated work. Defined authority.

Permissions

A workflow acts under its own name, with its creator’s authority. Roles are read fresh from the creator every run, so a permission you remove applies to the next run, and a workflow can never reach further than the person who built it. If that person loses their role, it refuses to run and records why.

Approvals

Your lifecycle policy decides, not the workflow. A step that is assigned to named people cannot be performed by a workflow at all: it is refused, and the work waits for them. Steps nobody is assigned to run on the creator’s authority like any other action.

Execution limits

A run is a job, and cancelling the job stops it before it takes its next step. Work already completed stands, because stopping is not undoing. A run that dies with its worker is closed out as failed rather than left open.

Run history

Every run keeps its trigger, each effect with the step and record it touched, start and finish, and on a failure the step that raised and its error. Notes record why a run did less than it looks like it should have, which is how a run that reached nothing is told apart from one with nothing to do.

Apply

Start with work your team repeats.

Four shapes that recur. Each is something you would build, not a template waiting in the product.

Cost-change review

Combined workflow
Trigger or goal
Trigger: the cost dataset refreshes.
Work performed
Items whose landed cost moved past your threshold are identified; the rest are marked within tolerance.
Result
Affected items flagged and a review submitted to the policy. Flagging and submitting are not recovered margin.

Margin investigation

Agent
Trigger or goal
Goal: explain a margin movement.
Work performed
The agent compares periods, decomposes price, cost and mix, traces the change to its source and ranks exposure.
Result
A written finding saved as a note, citing the records it used.

Approval routing

Automation
Trigger or goal
Trigger: a record enters a review state.
Work performed
The workflow alerts the right desk and moves the record along the verbs its policy allows.
Result
The record sits with the people the policy assigns. A step assigned to named people is theirs to take.

Renewal monitoring

Automation
Trigger or goal
Trigger: a schedule, on a cadence or a cron expression with a timezone.
Work performed
A dataset or view holding agreement and rebate dates is checked for milestones coming due.
Result
An alert naming what is approaching, with the records behind it.

Built on a dataset or view that carries those dates. Automations run on price lists, rule sets, datasets, views and the newsfeed; there is no agreement or rebate entity they attach to directly.

Built into your commercial platform.

Workflows and agents use Revomo’s shared records, supported actions and governance controls.

Bring the workflow your team keeps repeating.

Walk through how Revomo can combine defined steps, agent-assisted investigation and approval controls around your commercial work.

Plate B2 · Series 2026