Approval workflows for IFS
Approval workflows for IFS put human-in-the-loop steps inside your automations: an AI agent assembles the context, the flow pauses, the right person decides in email, Slack or Teams, and the outcome is written back to IFS Cloud — with a full audit trail of who decided what, and when.
Approval by inbox
Most ERP approvals are an email with a question mark. The request sits in an inbox; the approver, when they get to it, opens IFS in one window, the request in another, maybe a spreadsheet in a third — then replies “OK” into a thread that no system records. Multiply by every requisition, exception and master-data change.
The result is a queue with no clock. Vacations stall it, context-hunting slows it, and the organization quietly settles for one of two bad defaults: rubber-stamp everything to keep moving, or approve everything formally and let work wait for days.
- Approval requests live in inboxes and wait for days
- Approvers hunt context across IFS screens, emails and spreadsheets
- No timeouts or delegation — one vacation stalls the queue
- Decisions are made in email threads no system records
- The only modes are rubber-stamp everything or wait on everything
Request to recorded decision, step by step
Approvals are native steps in NgageFlow, so any flow — invoice capture, order intake, supplier onboarding — can pause for a person and resume on their answer.
- 1
IFS Cloud — New or Updated Record trigger
The flow starts when a record needing sign-off appears — a purchase requisition, an exception from another flow, a master-data change — anywhere a decision gate belongs.
- 2
AI agent — assemble the decision package
The agent pulls the related IFS records with Read Records, summarizes what is being asked, attaches the relevant history, and — where useful — adds a recommendation with its reasoning.
- 3
Approval — the flow pauses for a person
NgageFlow sends approve and reject links by email, Slack or Microsoft Teams and waits. The approver sees the whole package in one place and decides in one click, from any device.
- 4
Branch — act on the outcome
Approved: the flow continues — an Update Record or Execute Action step applies the decision in IFS Cloud. Rejected: the requester is notified with the reason, and the record is routed accordingly.
- 5
Delay + branch — remind, then delegate
If the deadline passes, the flow reminds the approver, then routes to their delegate or escalates up — by rule, not by someone noticing the queue stalled.
- 6
Google Sheets or IFS Cloud — record the trail
Every decision is logged with who, when and the full package they saw — to an audit log, back onto the IFS record, or both.
Every step here is a standard piece of the platform: connectors from the 700+ integration library, the native IFS Cloud connector, and AI agents as workflow steps — assembled in the visual builder, no code required.
Agents prepare, people decide
The agent never gets blanket authority. You draw the line in the flow itself — and the flow enforces it.
What the AI agent does
- Gathers the IFS records and history relevant to the decision
- Summarizes the request and drafts a recommendation with reasoning
- Auto-clears the cases below thresholds you explicitly set — if you choose
- Executes the approved action in IFS and files the audit trail
What people approve
- Makes the actual decision — approve, reject, or send back with questions
- Owns everything above threshold: spend, risk, master-data changes
- Sets the policies: who approves what, deadlines, delegation chains
What changes
No magic, just removed friction — teams typically report:
Approval cycles shrink from days to hours, because context arrives with the request
Decisions are made on complete information instead of inbox fragments
Vacations stop stalling the queue — reminders and delegation run by rule
Every decision is auditable: who approved, when, and what they saw
Approvals are better with context. NgageChat lets an approver interrogate the underlying IFS record in plain language before deciding, and NgageEase brings structured frontline input into the same flows. Both ship in the Ngage Suite by EX10.
Approval workflows for IFS: questions, answered
What does human-in-the-loop mean in NgageFlow?
It means approvals, forms and pauses are first-class steps inside a flow. The automation runs until a decision is needed, sends the package to a person, waits for their answer, and resumes on the outcome. Agents prepare; people decide.
Can people approve from Slack, Teams or email?
Yes. The approval step delivers approve and reject links wherever the approver works — email, Slack or Microsoft Teams — and works on mobile. No IFS login is needed to decide, and the outcome is still written back to IFS Cloud.
What happens if an approver never responds?
The flow does not stall silently. You set a deadline; when it passes, the flow reminds the approver, then routes to a delegate or escalates up the chain — all by rules you define. Every step of that journey is logged.
Other flows IFS teams build first
See approvals running on your IFS
Bring one approval that hurts today to the demo, or reserve an early-access pilot for your team.