The answer
Largely. Capturing requests, acknowledging them, sending status updates and answering 'where is my request?' are highly automatable. In our model about 70% of intake time can go to software. What stays human is deciding priority between competing requests and the weekly conversation about what will not get done this week.
The tasks, and how much of each is automatable today
Six tasks make up request intake. Capturing requests and sending status updates are about 90% automatable. Answering chasers is about 80% automatable once status is visible. Asking for missing details is about 60% automatable with required fields. Prioritising and reviewing the backlog stay mostly human, because they are trade-offs between people.
| Task | Min / week | Automatable |
|---|---|---|
| Capture requests from Slack and email into one trackerA single form or Slack shortcut removes nearly all copying. The 10% left is requests made in person or on calls. | 150 | |
| Ask for missing detailsRequired fields per request type stop most back-and-forth. Unusual requests still need a conversation. | 100 | |
| Prioritise and assign each requestRules can assign by type and suggest urgency. Choosing between two managers' 'urgent' requests is a human call. | 75 | |
| Send status updates to the requesterTriggered by status changes in the tracker. A person only writes updates when something has gone wrong. | 100 | |
| Answer 'where is my request?' messagesVisible status and an AI agent that can look it up answer most of these without the ops team. | 60 | |
| Review the backlog each weekSoftware prepares the list and ages each item. Deciding what to drop is a management decision. | 45 | |
| Whole process | 530 |
Hours saved per week
For a 40-person company whose operations team receives about 50 internal requests a week, intake and updates take around 9 hours a week by hand. Automating the tasks above returns about 6 hours a week. Most of the saving comes from fewer clarification messages and fewer people asking for updates.
Modelled on: A 40-person company where a two-person operations team receives about 50 requests a week by Slack message, email and in passing: access, purchases, data pulls and fixes.
- 50 requests a week, about 3 minutes each to copy into a tracker from wherever they arrived.
- Around 2 minutes per request on average spent asking for missing details.
- About 2 minutes per request on status updates over its life.
- About an hour a week answering 'any update on this?' messages.
Tools that do the job
You need one front door for requests, a place that holds them with a status, and something that keeps requesters informed. A Slack workflow or a Typeform form feeds Airtable or Notion. n8n, Make or Zapier sends acknowledgements and status updates, and an AI agent answers status questions in Slack.
- Slack
- A workflow shortcut so people can raise a request without leaving where they already work.
- Typeform
- One form per request type with required fields, for people outside Slack.
- Airtable
- The tracker: request, type, owner, status, target date and requester.
- n8n
- Sends acknowledgements, status updates and the weekly backlog digest.
- AI agent
- Answers 'what's the status of my request?' in Slack by reading the tracker.
A simple workflow
The workflow starts with one form per request type, each with its own required fields. Submissions land in a tracker with an owner and a target date. The requester gets an instant acknowledgement. Every status change notifies them. A weekly digest shows the backlog to the operations lead for prioritising.
- TriggerSomeone submits a request through the Slack shortcut or form.
- ValidateRequired fields for that request type are enforced before submission, not after.
- AssignThe request is assigned by type to an owner with a target date based on type.
- AcknowledgeThe requester receives a message with the owner, the target date and a link to track it.
- UpdateEach status change notifies the requester; overdue items notify the owner.
- ReviewMonday digest of open, overdue and unassigned requests goes to the operations lead.
Keep these human
- Choosing between competing priorities. Software can sort by date, not by what matters to the business this week.
- Saying no. Telling a requester their request will not be done is a conversation, not a status.
Where it goes wrong
- People bypass the form. If the ops team still answers Slack DMs, the form dies. Reply to DMs with the form link, every time, for a month.
- One generic form. A single form with optional fields recreates the back-and-forth. Use one short form per request type.
- Target dates nobody meets. Automated promises that are always missed are worse than none. Set dates the team can actually hit.
What most guides miss
Most guides recommend a request form and stop. The form is not the saving. The saving comes from required fields per request type, which kill the clarification ping-pong, and automatic acknowledgements with a date, which kill the chasers. Measure clarification messages per request before and after. That number tells you whether intake actually improved.
The evidence
- Knowledge workers spend 58% of their day on 'work about work', including chasing status updates.Asana, Anatomy of Work Global Index 2023 (2023)
- 68% of people say they do not have enough uninterrupted focus time during the workday.Microsoft, Work Trend Index 2023: Will AI Fix Work? (2023)
- Managing others has a technical automation potential of only 9%, which is why backlog review stays human.McKinsey Global Institute, A future that works: Automation, employment, and productivity (2017)