Home Use Cases Work Insights About Contact
Back to Our Work

How a Growth-Stage Educational Institution Built an Inquiries Agent That Answers Every Question It Can Prove

Operations2
Open the walkthrough

Five-Step AI Workflow with Human Oversight

Each step shows the workspace, what the AI completed, and what still needs human review.

Steps 3 and 5 need a person. Step 3 is waiting for approval right now. Click it to open the full walkthrough.

Estimates are directional and based on stated assumptions. All names, organizations, and identifying details have been anonymized in accordance with our confidentiality agreements.

The Transformation

How a Growth-Stage Educational Institution Built an Inquiries Agent That Answers Every Question It Can Prove

Before
A person used to read every message and file it by hand.
A person used to look up and type every answer.
A person used to close every ticket and chase the survey by hand.
After
every message labelled and tied to the person who sent it
every answer carrying the page it came from
the short queue only a staff member can settle
the ticket record, and the conversations that came back
the week's misses turned into next week's answers
Executive Summary

Answers every student question it can prove, on the channel it arrived on, at any hour, and hands a person the ones it cannot. Written for an org where the institution's word matters: every answer carries the page or the staff reply it came from, a claim with no source never sends, and the only thing left for a person is the short queue the knowledge base genuinely cannot settle. Five stations, run as a loop: the articles a person approves at station 5 are the answers station 2 sends next week.

What Was Broken

A person used to read every message and file it by hand
A person used to look up and type every answer
The real cost
A person used to close every ticket and chase the survey by hand.

What We Built

Five stations and 11 subagents. Each subagent carries its own tasks and its own refusal.

A4
Channel Watch
Poll the chat widget, the support mailbox and the SMS line, and stamp each message with where it came from
A3
Intent Classifier
Classify each inquiry: registration, financial aid, records, housing, IT access
A3
Answer Retrieval
Search the knowledge base and closed tickets for passages that answer this question
A2
Answer Drafting
Draft the reply from the retrieved passages, in the shape of the channel it goes back on
A3
Escalation Routing
Route each handoff to the office that owns that kind of question: Registrar, Financial Aid, IT Service Desk
A2
Case Brief
Summarize each handoff in one line the staff member can reply from
A3
Resolution Check
Confirm an answer landed on the thread, the bot's or a staff member's, before anything closes
A4
Record Writing
Write the ticket record: the question, the answer, its sources, who settled it
A2
Gap Finding
Group the week's handoffs by what the bot was missing when it held each one
A1
Article Drafting
Draft each proposed article from the staff member's own replies
A4
Week Reporting
Assemble the week's counts: answered, handed off, reopened, survey results

How It Runs

1

Inquiry Intake

Nobody is asked anything here

Agents watch every channel questions arrive on, tie each message to the person who sent it, and label what it is asking before anything tries to answer it. Nothing waits on a person here. A student reaches somebody only when a later station cannot settle their question.

2

Question Answering

Nobody is asked anything here

Agents search the knowledge base and past staff replies, draft the answer with a source on every claim, send it back on the channel it arrived on, and hold the conversation across turns. A person sees only the conversations the bot holds back, and station 3 carries those to them.

3

Staff Handoff

A person answers here

Agents route each held conversation to the office that can actually answer it, with the whole thread, what the bot already said, and a drafted reply to start from. This is the station where a person works. Everything else exists to keep this queue short.

4

Ticket Closing

Nobody is asked anything here

Agents confirm the student got an answer, send the one-question survey, write the ticket record with its sources, and reopen a ticket when the same student comes back on the same question. Nobody signs a closing. The answer was either a staff reply from station 3, or a bot reply that cited its source and drew no comeback.

5

Weekly Learning

A person answers here

Agents group the week's handoffs by what the bot was missing, draft the knowledge base articles that would have settled the recurring ones from the staff's own replies, and assemble the week's counts. A person approves every article before the bot is allowed to repeat it, because the bot will say it all year.

Where a Person Decides

Step 3, Staff Handoff. Read one line and send a reply, or send the draft as it stands. Open the full thread only when the line is not enough. what the bot cannot settle, a person must.
Step 5, Weekly Learning. Read the two drafted articles and approve or reject each one. Settle the late add fee: the catalog or the Bursar page, and fix the loser. the bot repeats only what you approve.

Operating Model

This changes how work flows through the team.

Role
Responsibility
Staff Handoff owner
Read one line and send a reply, or send the draft as it stands. Open the full thread only when the line is not enough. what the bot cannot settle, a person must.
Weekly Learning owner
Read the two drafted articles and approve or reject each one. Settle the late add fee: the catalog or the Bursar page, and fix the loser. the bot repeats only what you approve.

What Transfers, What Must Be True

What transfers
A person is in the loop wherever the college's word is given to a student the bot could not serve, and wherever a new answer enters what the bot is allowed to say. Two of the five stations refuse to proceed on their own, Staff Handoff and Weekly Learning, and those are the two a person carries. Everywhere else the agents run at volume and reach you only when they cannot cite something.
Staff Handoff stops for a person, and what the bot cannot settle, a person must.
Weekly Learning stops for a person, and the bot repeats only what you approve.
Every subagent says what it will not do. 11 of them do.
What must be true in your environment
The agent can read the systems your records already live in. This one reads 14.
Somebody owns Staff Handoff and has time for it.
Somebody owns Weekly Learning and has time for it.

Failure Modes

What breaks this pattern:

✗ The nearest paragraph becomes the answer

A student asks about a deadline the knowledge base does not cover. The bot sends the closest paragraph it found, and the student acts on a date that was written for a different program.

✗ The handoff lands in a queue

The bot could not help, so it drops the case into a shared queue with nobody named on it. Every staff member assumes someone else owns it, and the student who most needed a person gets nobody.

✗ Closed tickets with no answer

A conversation stalls, the student stops replying, and the ticket closes as resolved. The dashboard shows the question was handled, and nobody ever finds out it was not.

✗ Weekly numbers nobody can check

The week report says 94 percent of inquiries were resolved, and nobody can trace that number back to a ticket. You make staffing decisions on a figure the system invented, and you find out the season it mattered.

Directional Outcomes

What the agent counts, and the station that counts it.

These counts are the tallies from one monitored run of the agents. They are not monthly or annual totals.

Messages labelled
Counted at Inquiry Intake
3,118
Filed under the wrong question
Counted at Inquiry Intake
0
Held unreadable
Counted at Inquiry Intake
6
Questions answered
Counted at Question Answering
2,606
Claims sent without a source
Counted at Question Answering
0
Held for staff instead
Counted at Question Answering
292
Tickets closed
Counted at Ticket Closing
2,887
Closed with no answer on the thread
Counted at Ticket Closing
0
Reopened and routed back
Counted at Ticket Closing
57
Our measurement policy: We do not publish precise ROI without baseline methodology. Every figure above carries its basis.

What Runs Where

Every step names the subagent that does the work, the record it writes, the thing that raises a question for a person, and what it is allowed to touch. This is drawn from the source, not from a diagram somebody kept in sync by hand.

1Inquiry Intake
subagentinquiry-intake
writesqueue/inbound/<conv-id>.json
raiseshold-unreadable
may touchqueue/**, records/** read-only
2Question Answering
subagentanswer-draft
writesconversations/<conv-id>/reply.json
raisesclaim-without-source
may touchconversations/**, kb/** read-only
3Staff Handoff
GATE
subagenthandoff-brief
writesqueue/handoffs/<conv-id>.json
raisesawait-staff-reply
may touchqueue/handoffs/**, everything else read-only
4Ticket Closing
subagentticket-close
writestickets/closed/<ticket-id>.json
raisesclose-without-answer
may touchtickets/**, conversations/** read-only
5Weekly Learning
GATE
subagentweekly-learning
writeskb/drafts/<article-id>.md, reports/week-<n>.json
raisesapprove-article
may touchkb/**, reports/**, tickets/** read-only

Stack

Every system this agent reads or writes.

System
Read at
Stations
the SMS line
Inquiry Intake
1 of 5
the chat widget on the college website
Inquiry Intake
1 of 5
the closed-ticket archive
Question Answering
1 of 5
the course catalog
Question Answering
1 of 5
the handoff log
Weekly Learning
1 of 5
the knowledge base
Question Answering
2 of 5
the same three channels the question arrived on
Question Answering
1 of 5
the satisfaction survey
Ticket Closing
1 of 5
the staff directory
Staff Handoff
1 of 5
the student information system
Inquiry Intake
2 of 5
the support mailbox
Inquiry Intake
1 of 5
the support team's shared workspace
Staff Handoff
1 of 5
the ticket system
Staff Handoff
2 of 5
the week's counts on the support dashboard
Weekly Learning
1 of 5
Next Step

Want to see if this pattern fits your customer support bots?

No build commitment·Real samples, not a demo·Estimate in writing