Algorithmic management · for CEOs and COOs — Managers shouldn't find out about delays in the report. The rule chases, decides and acts right away.
Algorithmic management means running the operation through goals and rules that act on their own. The company goal cascades down to each person. When a goal goes off track, the rule opens the action plan; when a deadline passes, it chases the owner and then the manager; above a set amount, it requires approval. Execution stops depending on whoever does the chasing.
- 8 triggers, 4 conditions, 23 actions, no coding
- Rules chained up to 5 levels deep
- Every run recorded
Planning becomes each person's goal. A deviation becomes an action plan the same day.
Today, the path is long: the dashboard closes, the meeting analyzes, the action plan goes to a spreadsheet, email distributes it and the follow-up waits for the next meeting. At Olie, the company goal cascades into area and personal goals, and each goal can trigger an action. Leadership decides in advance what happens when the number deviates; the rule executes the moment it does.
- Cascading goals and OKRs6 goal types, per person or per team; the team goal adds up each person's.
- Goal that triggers an actionBelow pace, the automation creates the action plan as a project, with one task (child project) per owner.
- Owner, deadline and follow-upEach task is born with an owner and a deadline; if it stalls, the rule alerts and then escalates.
- The result goes back to the goalThe same goal shows whether the plan worked, without building another report.
Project idle past its deadline: the rule alerts the owner. If it stays idle, it escalates to the manager.
Today, someone opens the board every morning to see who is late and messages them one by one. At Olie, the platform checks idle projects every minute, alerts the person and then their manager. The manager also sees the postponement counter: how many times the date was pushed.
- Deadlines that chase on their ownIdle project trigger, checked every minute.
- Escalates to the managerThe second alert goes one level up the hierarchy.
- Postponement counterHow many times the forecast was pushed.
- Record per runWhat triggered, what ran, what failed.
Above a set amount, the decision goes to the board. Company policy becomes a rule, with no coding.
Today, the decision depends on who is on shift and how much they remember of the policy. At Olie, the policy becomes a rule: reaction to value, reaction to a field and goal that triggers an action. There are 8 entry triggers, 4 conditions and 23 actions, chained up to 5 levels deep, without coding. For calculations, there are advanced expressions.
- Reaction to valueAn order above the set amount requires board approval.
- Reaction to a fieldA field's answer changes the path: a score above the cutoff moves the lead (the interested contact) to Qualified.
- Approval via WhatsAppThe director replies on WhatsApp and the answer is recorded.
- Chaining up to 5 levelsOne automation calls the next.
And when the rule needs math?
Advanced expressions inside the automation, without leaving the visual setup. Rule-fed goals and an impact scale from 1 to 10.
Any field, at any time. And/Or conditions, one path for true and another for false.
Today, every exception lands back on the manager's desk. At Olie, any edit to any form field, at any time, can trigger an automation. The filter decider combines conditions with And and Or, looks up fields from other forms (from the step, the funnel, another area) and follows two paths: if true, one list of actions; if false, another. Automations chain up to 5 levels deep.
- Trigger on any fieldForm edited, field changed, tag, date, run, content received.
- Filters with And and OrAs many conditions as you need, including data from other forms.
- True and falseEach path with its own actions: email, tag, form, group, hierarchy, webhook.
- Optional lockBlocks the automation's whole flow when the condition is false.
The rule reads the data and acts before the delay. One agent reads, another decides, the automation executes.
Today, someone exports the spreadsheet, analyzes it and finds the problem weeks later. At Olie, the rule anticipates the risk. Every 10 minutes, an AI agent queries the forms (sales, finance, any data) and saves its reading; a second agent interprets it and decides; the automation executes the action. Every run is recorded.
- Agent reads the formAny data in a field can be read.
- Runs every 10 minutesIdle-time automation as a clock.
- Second agent decidesInterprets the output and chooses the action.
- Trend analysis through ChatGPTOfficial plugin: forms, logs and data, with a recurring task.
- Forecast weighted by step"Proposal sent" is worth 40%, "contract out for signature" is worth 90%: the month's forecast is what is actually close to closing.
And the board's tailored view?
The tailored view comes through MCP, the connector that lets the board's ChatGPT or Claude query Olie: "which funnel is slowest this month?", and the answer comes with the data.
The manager stops chasing. Merit shows up in the numbers.
When the rule does the chasing, the manager only steps in on escalation and gains time to develop the team. What each person delivers is measured: on time, with quality, without rework, with the goal met. The company defines what counts, and pay for performance is based on numbers.
- The manager only steps in on escalationThe first alert goes to whoever has the project; the manager shows up later, already with the history in hand.
- Rework countedHow many times the card went back a step, per person.
- Merit with numbersIndividual goals adding up to the team's and each person's time per step.
- The bonus rule is your company'sOlie measures timeliness, quality, rework and goals; you decide what counts and how much it weighs.
What is algorithmic management in Olie?
It means running the operation through goals and rules that act on their own inside the process. Goals cascade from strategy down to the person; a goal that goes off track opens the action plan with an owner and a deadline; a missed deadline chases the owner and escalates to the manager; above a set amount, the step requires approval. AI agents read the data to anticipate risk, and the sales forecast weights each step. There are 8 triggers, 4 conditions and 23 actions, chained up to 5 levels deep, without coding, and every run is recorded.
Updated on · Source: Olie technical base · Guide to the 77 features
Design the rule once. It chases every day.
Execution that doesn't depend on people: goals and rules decide and follow up, without anyone having to remember. In the next tab, see who does what, in what order, across people, AI and areas.
Next step: 30 minutes to see your process on screen; then, a diagnosis with your leadership.