AI Workflow in Oppora.ai: A Complete Step-by-Step Guide

What are AI Workflows? 

AI Workflows automate multi-step prospecting so your lists stay full without manual work.

Instead of running a company search, then a people search, then enrichment, then scoring by hand, you chain those steps into one workflow of connected agents that runs on demand or on a schedule. 

A typical workflow reads like a sentence: find companies that are hiring → find the right people at them → enrich their emails → score them against your ICP → (optionally) start a campaign.

Each step is an agent you drop onto a canvas and configure. 

 Workflows list each workflow shows a status, and how many companies, people, and emails it generated. 

The Workflows list: statuses (Complete, Ready to run, Running, Paused) with the counts each run produced. 

The Workflow Canvas and Actions Panel 

Open or create a workflow to reach the canvas — a visual editor where every workflow begins with a Workflow Start node and flows downward through the agents you add. 

• Canvas — drag, connect, and configure agent nodes. Use Tidy Layout to auto-arrange, and the zoom controls to navigate. 
• Actions panel (right) — the library you build from, with two tabs: Agents (the steps that do the work) and Triggers (events that can start a workflow). 
• AI Chat — a copilot that can build or edit the workflow from a chat instruction. 
• Run / Save — save your workflow, or run it. 

 Workflow canvas: Workflow Start, a Company Finder node, and a Job Finder node with the Agents panel. 

The canvas: nodes connect top-to-bottom from Workflow Start; the Actions panel lists Agents and Triggers. 

Agent input: Manual or Prompt 

Most finder agents accept input two ways, via a toggle in the node: Manual (you set the filters yourself, like Manual Filters in Finder) or Prompt (you describe what you want in plain language, like AI Search).

Some agents also let you pick a source platform for example, Job Finder can source from LinkedIn or Glassdoor. 

Agents the Building Blocks 

Agents are the steps that make up a workflow. The main ones: 

Agent	What it does
Company Finder	Finds companies matching your criteria or prompt (the workflow equivalent of Company Finder).
Job Finder	Finds job postings by title/location from LinkedIn or Glassdoor — great for hiring-signal prospecting.
People Finder	Finds people (decision-makers) at the companies in the workflow.
Enrichment	Fills in missing company or lead data.
Lead Scoring	Ranks leads by how well they match your ICP, so the best are contacted first.
Company Remover	Filters companies out of the set by rules you define.
Lead Remover	Filters leads out of the set by rules you define.
Web Search	Runs a web-wide search to gather additional context.
If / Else	Branches the workflow down different paths based on a condition.
Wait	Pauses the workflow for a set time before the next step.
Email Finder	Finds email addresses for the leads in the workflow.
Email Verifier	Verifies that the found addresses are deliverable.
Campaign Bot	Prepares campaign content/actions within the workflow.
Start Campaign	Kicks off an outreach campaign from the workflow's results.
CRM Writer	Writes results (contacts/opportunities) into the CRM.

Pattern: Chain finders → removers/scoring → enrichment → Start Campaign to run a full source-to-send motion hands-free. Put a Wait or If/Else in the middle when you want the workflow to pace itself or branch on lead quality. 

A Triggers tab also exists (events like Job Change, Funding, Job Posted, Score reached, Warm Up, and more) for starting workflows automatically. These are marked coming soon and are not yet available. 

Running a Workflow (When to Run) 

When you run a workflow, Oppora asks when it should execute. Click Run (or the Trigger node) to choose: 

Option	What it does
Run now	Starts immediately when you click Run.
Run once at…	Runs a single time at a specific date and time.
Run on a schedule	Runs repeatedly — every day or every week at a set time.

 "When to run" dialog Run now, Run once at a date/time, or Run on a schedule. 

Choose Run now, a one-time scheduled run, or a recurring schedule; change it anytime from the Trigger node. 

Workflow Statuses and Where Results Go 

The four statuses 

Status	Meaning
Ready to run	A draft workflow, built and waiting — not yet started.
Running	In progress right now.
Complete	Finished its run.
Paused	You manually paused a running workflow. (Nothing auto-pauses it — a pause is always a deliberate action.)

Where a workflow's results land 

It depends on where you started the workflow from: 

• Created inside a list results write back into that same list; no new list is created. 
• Created from the Workflows area you choose at setup whether to write into an existing list or create a new one. 

Reading the list: The Workflows list shows the companies, people, and emails each workflow generated.

If the dashboard says workflows are paused, that's because someone paused them open Workflows and resume the ones you want running. 

Workflow Credits 

Workflow billing has two parts that combine: a small AI charge for each agent step, plus data credits that scale with how much each step actually finds. 

• 1 AI credit per agent step: regardless of how many results that step returns. 
• Data credits scale with results:  e.g. if Company Finder returns 100 companies, that's 100 data credits; People Finder and enrichment steps charge per lead as usual. 

Worked example 

A workflow with Company Finder → People Finder → Email Finder: 

• Company Finder finds 100 companies → 1 AI credit (the step) + 100 data credits (the results). 
• People Finder runs on them → 1 AI credit + per-lead data credits for the people it returns. 
• Email Finder runs on those leads → per-lead data credits (1 each) for the emails it finds. 

So the AI cost stays tiny and predictable (one per step), while the data cost tracks the real volume you pull.

Cap results per step (for example, "find 133 companies") so an over-broad prompt doesn't burn a large batch of data credits. 

Tip: The canvas shows a running "credits consumed" counter as you build and run. Watch it on your first few workflows to calibrate how broad your prompts should be.