It decides whether any of the rest gets used
Filters look minor next to the intelligence in the platform, and they determine whether it is ever reached. A queue that cannot be narrowed is watched rather than managed.
AI Inbox
Filters in the AI Inbox
A single queue across every channel is only useful if somebody can narrow it to the part they own. Filters are what turn one large inbox into the working view a support lead actually manages — by product, severity, customer tier, channel or SLA risk.
- Narrow by product, severity, tier, channel or SLA risk
- Build the view you manage rather than watching all of it
- Combine filters to isolate what is at risk right now
- The small feature that decides whether a shared queue works
A shared queue is unusable until it can be narrowed
Consolidating every channel into one inbox solves a real problem and immediately creates a smaller one. The total is now visible, which is what you wanted, and nobody actually works on the total.
Without a way to cut it, everybody stares at the same undifferentiated list and filters it in their head — slow, inconsistent between people, and precisely the work that quietly reintroduces the separate queues. That is why shared inboxes fail: the last step never got finished.
Building a view you can actually work
Four cuts, and the fourth is the one that changes how a day goes.
- Step 1
By product
Narrow to the product line you are responsible for, so the queue reflects your remit.
- Step 2
By severity
Separate what is broken from what is a question.
- Step 3
By customer tier
Isolate the accounts with commitments attached, so a strategic customer stops sitting behind routine volume.
- Step 4
By SLA risk
Surface what is approaching a breach rather than what arrived earliest. This is the view a support lead checks first thing.
Why a small feature decides a large one
This is the difference between a queue that is managed and one that is merely watched.
Everybody needs a different cut
The team lead, the engineer on rota, the account manager and the person on SLA duty are looking at one queue for four unrelated reasons. No default view serves them.
Combining them is where it earns its keep
Any single dimension is coarse. The useful views are intersections — high severity, one product line, tier-one customers, approaching an SLA — which no sort order can answer.
Why the usual approaches fall short
Three of these rebuild the problem consolidation was meant to solve.
Separate queues per team
Gives everyone their own view and takes the total away again.
Sorting by date
Treats arrival time as the most important fact about a request. It is rarely in the top three.
Manual assignment into buckets
Somebody spends their morning distributing work, and the buckets go stale.
Saved searches in a ticketing tool
Closest to the right idea, and usually built on fields the customer filled in.
What things are called
Age and risk in particular are not the same thing.
- AI Inbox
- The single queue every channel arrives into, categorised and prioritised together.
- Filter
- A cut across that queue — by product, severity, customer tier, channel or SLA risk.
- View
- A combination of filters somebody works from day to day.
- SLA risk
- How close a request is to breaching its commitment, as opposed to how long it has existed.
- Customer tier
- The commitment level attached to an account, used to weigh its requests against others.
Frequently Asked Questions
Because it is the feature that decides whether a shared inbox survives contact with a real team. Consolidating the channels is the impressive half; being able to narrow the result is the half that determines whether anyone is still using it in month three.
Where this fits
AI Inbox
Describe the view you wish you had
Most support leads already know the cut they want and cannot build it in the tools they have. Tell us what it is and we will show it to you against your own queue.
Talk to us