How it works
From a template or a sentence to a workflow you can trust
Six stages — start, plan, approve, run, verify and monitor. A person stays in control of the ones that matter, and Awish keeps checking the work after it goes live.
Start from a ready-made template for your team or industry, or describe the process in your own words. Awish plans it, asks for each connection, stops before sensitive steps and runs it across your apps. Then it checks that the expected result actually happened, and tells you when a workflow needs attention and why.
Free plan · Outcome verification included · No credit card required
Start from a template, or describe it
Pick a ready-made workflow built for your team or industry, or write what should happen in the words you would use with a colleague. Both become the same kind of workflow, with the same approvals and the same monitoring.
A template is a request already written for its job: its apps, its approval point and what Awish should check afterwards come with it. Describing it yourself works the same way — “when an invoice arrives, log it to my books sheet and ask me before anything over $1,000” carries a trigger, an action, a threshold and an approval point, and Awish reads all four.
- Templates for every team and industry
- Or one sentence, in your own words
- The same approvals and monitoring either way
Awish plans the workflow
Awish works out the trigger, the actions, the apps involved, the conditions, how to handle exceptions, and which steps should stop for a human. It shows you that plan before anything connects.
You can change it in conversation — move the approval point, raise a threshold, add a channel — and the plan updates. Nothing is committed while you are still reading it.
- Trigger, steps and apps, in order
- The step that waits for a person
- How Awish will check the outcome
You grant the connections
Each app is requested separately, and you can decline any of them. An automation granted your inbox cannot reach your files, because it was never given them.
Sensitive actions — sending externally, refunding, publishing — are approval-gated by default rather than as an option you have to remember to switch on.
- Each app requested on its own
- Any connection can be declined
- Sensitive actions wait for approval by default
It runs, and you stay in control
The workflow runs on its trigger or its schedule. When it needs a decision it asks you in WhatsApp, Slack or Telegram; when it finishes it reports what it did.
From the same conversation you can ask what ran, start a run, or pause one. If a step fails, the run stops and tells you why rather than skipping silently.
- Decisions asked in WhatsApp, Slack or Telegram
- Ask what ran, start or pause a run
- A failed step stops and says why
Awish verifies the outcome
A run that finishes is not the same as work that got done. After each run, Awish checks the result the workflow was meant to produce: the row in the sheet, the updated CRM record, the message that reached the customer.
When a run reports success but the expected result is missing, incomplete or wrong, Awish flags it as a silent failure and tells you straight away, instead of leaving it for someone to discover at month end.
- Checks the result in the destination app
- Flags runs that report success but did nothing
- Tells you straight away
It keeps monitoring, and shows what it gives back
Once live, every workflow is watched: connection health, failures, triggers that go quiet and behaviour that drifts from its normal pattern. When something needs attention, Awish explains the likely cause, shows which workflows are affected and recommends what to do next.
You decide what happens next: reconnect an app, replay the affected runs, or change the workflow. Awish also measures the hours your automations save, so you can see which ones are paying off. Anomaly detection, root-cause explanations and recommendations come with Advanced Monitoring on the Business plan.
- Connection health, failures and quiet triggers
- Likely cause and a recommended next step
- Hours saved, per workflow
Questions about the model
What if the plan Awish proposes is wrong?
You change it in conversation before it runs. The plan is shown to you precisely so it can be corrected while correcting it is cheap — nothing is connected and nothing has executed at that point.
Do I have to know which apps to use?
No. Describe the outcome and Awish identifies the apps it needs, then asks you to connect each one. If it picks the wrong tool for a step, say so and it will use the one you name.
What happens when a step fails?
The run stops at the failed step and reports the reason in your channel. Awish also catches the failures that do not stop a run — a step that reports success but leaves the expected result missing — and flags them as silent failures.
Does Awish fix problems on its own?
No. It explains the most likely cause, shows which workflows are affected and recommends the next step, such as reconnecting an app and replaying the affected runs. You decide, and nothing is replayed without you.
Which monitoring comes with each plan?
Every plan includes live status, failure alerts, outcome verification, silent-failure detection, connection health, the execution timeline and hours saved. Pro adds search and filters, manual replay and 30 days of history. Business adds anomaly detection, root-cause explanations, recommendations and cross-workflow health with 90 days. Free keeps 7 days.
Can I change a workflow after it is running?
Yes — by describing the change, the same way you described the workflow. You do not reopen a canvas to move an approval point or adjust a threshold.
Start with a template, or describe your own
See the plan before anything connects, and know afterwards that the work got done.
Free plan · No credit card required