Back to Pratik
Nova Education Foundation Practical recommendations

Business Automation Report

Nova Education Foundation

Prepared by: Pratik Kuikel Purpose: Show what may be worth automating, how it could work, the likely benefit, priority, and practical next steps Sprint time: 10 total hours — 2 hours discovery, 6 hours research and report preparation, 1 hour walkthrough, and 1 hour revision and clarification support Regional pricing: NPR 1 lakh for Nepal-based businesses; USD 1,000 for international businesses Version: 1.0

Sample notice: This is an illustrative report showing what a client would receive from the research sprint. The workflows, problems, and benefits below are examples and are not verified claims about Nova Education Foundation.


The Short Answer

The best place to start is customer inquiry follow-up.

In this example, inquiries arrive through different channels and staff must remember who needs a reply, what was discussed, and what should happen next. A simple shared inquiry list with automatic assignment and reminders would solve an immediate customer-facing problem without requiring a complicated AI system.

This should come before an AI agent because it creates the reliable information an AI feature would need later: one inquiry record, a clear owner, current status, and the next action.

Recommended order

  1. Centralize inquiry capture and follow-up.
  2. Automate admission document reminders.
  3. Automate fee follow-up using verified payment status.
  4. Automate timetable-change notifications.
  5. Create a simple weekly operations summary.
  6. Test AI-assisted reply drafting after the underlying information is reliable.

What I Learned About the Business

The example focuses on the work between a student's first inquiry and their ongoing class experience.

How the work appears to happen today

  • Inquiries can arrive by phone, website form, social media, messaging apps, walk-in visits, or referrals.
  • Reception or counselors record information in messages, notes, forms, or spreadsheets.
  • A counselor follows up and gathers the student's requirements.
  • Admissions collects forms and supporting documents.
  • Finance confirms fees and payment status.
  • Coordinators share schedules and changes with students and instructors.
  • Management asks different teams for weekly updates.

Where the work may slow down

  • An inquiry may not have one visible owner.
  • Follow-up depends on a person's memory or private reminder.
  • The same student information may be copied into several places.
  • Staff may ask for documents that were already sent through another channel.
  • Fee reminder lists may become outdated before messages are sent.
  • Schedule changes may be shared differently by different staff members.
  • Weekly reporting may require manually combining several spreadsheets.

These are not all “AI problems.” Most should first be solved with clearer information, ownership, and simple automation.


What Can Be Automated

1. Customer inquiry follow-up

Priority: Start here

What is happening?

Inquiries arrive through several channels. Staff may record them differently, and follow-up can depend on someone remembering the next action. Management cannot easily see which inquiries are waiting or who owns them.

What can be automated?

  • Capture every new inquiry in one shared place.
  • Check for likely duplicate inquiries.
  • Assign an owner using simple rules such as program, location, or availability.
  • Remind the owner when a follow-up is due.
  • Escalate an inquiry when it has been waiting too long.
  • Stop reminders when the inquiry is completed, lost, or postponed.

How would it work?

Website forms and supported messaging channels would send the student's name, contact details, source, and area of interest to a shared inquiry list. The system would assign the inquiry, notify the counselor, and create a next-action date. If no action is recorded by that date, the counselor would receive a reminder. Older inquiries could appear in a manager's attention list.

AI is not required for the first version. Once the process is working, AI could help classify open-ended messages or draft a response for staff approval.

What is the benefit?

  • Fewer inquiries are forgotten or lost.
  • Students receive faster and more consistent replies.
  • Staff spend less time checking messages and personal notes.
  • Every inquiry has a clear owner and next action.
  • Management can see demand, response time, and reasons inquiries do not convert.

Why is this first?

It addresses a visible customer problem and creates clean information for later improvements. Other automation becomes safer when every inquiry has a reliable record and status.

What are the first steps?

  1. List every channel where inquiries arrive.
  2. Agree on the minimum information required for a new inquiry.
  3. Decide who owns each type of inquiry.
  4. Define simple statuses such as New, Contacted, Qualified, Enrolled, and Closed.
  5. Decide how long an inquiry can wait before a reminder or escalation.
  6. Confirm which existing tool should hold the shared list or whether a new one is needed.

What should we watch out for?

Duplicate records must not produce duplicate messages. Staff should be able to correct an assignment, pause reminders, and see when an automated step fails.


2. Admission document reminders

Priority: Next

What is happening?

Students may send documents through email, forms, or messaging apps. Admissions staff must check several places, update a list, and repeatedly tell students what is missing.

What can be automated?

  • Create the correct checklist for each program.
  • Mark documents as received, missing, invalid, or verified.
  • Remind students only about genuinely missing items.
  • Notify staff when a complete application is ready for review.

How would it work?

When an application starts, the system would create a checklist based on the chosen program. Staff would review submitted files and update their status. At an agreed time, the system would send an approved reminder listing only the missing items. A person would still verify documents and handle exceptions.

What is the benefit?

  • Less time spent checking and writing repeat reminders.
  • Fewer repeated requests for documents already received.
  • Students can understand what remains incomplete.
  • Admissions can focus on review and exceptions instead of list maintenance.

Why is this next?

It is relatively simple once student records and ownership are clear. It should follow the inquiry foundation so the same person is not represented differently across several lists.

What are the first steps?

  1. Create the checklist for each program and applicant type.
  2. Decide which document states staff need.
  3. Choose where files and checklist status will live.
  4. Approve reminder timing and message wording.
  5. Define who handles invalid, sensitive, or unusual documents.

3. Fee follow-up

Priority: Next, after payment data is confirmed

What is happening?

Finance may prepare a list of unpaid or overdue fees and pass it to staff for follow-up. By the time a message is sent, a payment, extension, scholarship, or dispute may have changed the situation.

What can be automated?

  • Build the follow-up list from current payment status.
  • Send approved reminders for straightforward overdue cases.
  • Stop reminders immediately after payment or a status change.
  • Send sensitive or repeated cases to a person instead of continuing automatically.

How would it work?

The accounting or payment system would remain the trusted source. At a scheduled time, the automation would select only eligible overdue records, exclude special cases, and prepare or send the approved reminder. Delivery and response status would be recorded for finance to review.

What is the benefit?

  • Less time spent building and checking reminder lists.
  • Fewer incorrect reminders after payment.
  • More consistent follow-up.
  • Staff can concentrate on exceptions and sensitive conversations.

Why is this not first?

Incorrect financial communication can damage trust. This should begin only after the payment source, exclusions, approval rules, and message history are reliable.

What are the first steps?

  1. Confirm the trusted source for payment status.
  2. List every case that must be excluded from automatic reminders.
  3. Agree on timing, tone, channel, and escalation.
  4. Test the list without sending messages.
  5. Review the test results with finance before enabling delivery.

4. Timetable-change notifications

Priority: Next

What is happening?

When a timetable changes, staff may update a spreadsheet and then separately message instructors and students. Different versions can remain in circulation.

What can be automated?

  • Publish an approved schedule change from one source.
  • Notify the affected instructors and students.
  • Record whether the notification was delivered.
  • Send failed deliveries to staff for follow-up.

How would it work?

A coordinator would update and approve the schedule in the chosen source. The automation would identify the affected class, prepare the approved message, and notify the correct people. The message would link back to the current schedule rather than copying a version that may later become outdated.

What is the benefit?

  • Faster and more consistent communication.
  • Fewer conflicting schedule versions.
  • Less manual recipient selection and repeated messaging.
  • Visibility when a notification fails.

What are the first steps?

  1. Choose the one trusted timetable source.
  2. Define who can approve and publish a change.
  3. Confirm student and instructor contact preferences.
  4. Create approved message templates.
  5. Define what happens when delivery fails.

5. Weekly operations summary

Priority: Build after the main workflow is reliable

What is happening?

Management may ask several teams for inquiry, enrollment, payment, class, and attendance numbers. Someone then reconciles spreadsheets before the meeting.

What can be automated?

  • Produce a weekly summary from the agreed sources.
  • Show new inquiries, waiting follow-ups, enrollment progress, and important exceptions.
  • Highlight records that need attention instead of only showing totals.

How would it work?

The report would read defined figures from the operational tools and refresh on a schedule. Each number would use an agreed definition and allow staff to trace it back to the underlying records.

What is the benefit?

  • Less time spent preparing the same report.
  • More confidence that teams are discussing the same numbers.
  • Problems become visible before the end of the week.
  • Managers can focus on decisions rather than reconciliation.

Why should it wait?

A dashboard will only repeat unreliable information if the underlying records are incomplete. Build the workflow first, then report from it.

What are the first steps?

  1. Agree on the few numbers management actually uses.
  2. Define exactly how each number is calculated.
  3. Confirm its trusted source.
  4. Measure how often the current data is missing or late.
  5. Build the smallest useful weekly view first.

6. AI-assisted reply drafting

Priority: Later, as a controlled test

What is happening?

Counselors repeatedly answer common questions about programs, schedules, documents, and next steps. Drafting similar replies takes time, but the correct answer can depend on current policies and student context.

What can be automated?

AI can prepare a suggested reply using approved information. A staff member would review, edit, and approve it before anything is sent.

How would it work?

The system would provide the student's message and relevant approved information to the AI. The AI would return a draft and show which source it used for important details. If information is missing, conflicting, or sensitive, it would ask the staff member to handle the reply instead.

What is the benefit?

  • Less time spent writing routine responses from scratch.
  • More consistent use of approved information.
  • Staff still control tone, accuracy, and sensitive decisions.

Why is this later?

AI drafts are only useful when program, fee, schedule, and policy information is current and trusted. The inquiry workflow and approved information should be established first.

What are the first steps?

  1. Collect approved answers and their source documents.
  2. Choose a small group of common, low-risk questions.
  3. Prepare examples that include ambiguous and sensitive requests.
  4. Require staff approval for every draft.
  5. Measure whether review is faster than writing from scratch and track factual errors.

What should we watch out for?

The AI must not invent fees, dates, policies, or eligibility. Unsupported and sensitive requests must go directly to a person.


Recommended Order

Start here

Centralize inquiry capture, ownership, and reminders. It solves the clearest immediate problem and creates the reliable records needed by everything else.

Do next

  1. Add admission document checklists and reminders.
  2. Validate payment data, then test fee follow-up without sending.
  3. Publish timetable changes from one trusted source.

Add after the foundation works

  1. Create the weekly operations summary.
  2. Test AI-assisted reply drafting with human approval.

Do not start with

A fully autonomous student-facing AI agent. In this example, the business first needs clear ownership, trusted information, approved answers, human escalation, and a way to detect mistakes.


Before Implementation

The following must be confirmed before pricing or building the first automation:

  • Which inquiry channels are in scope?
  • Approximately how many inquiries arrive through each channel?
  • Where should the shared inquiry record live?
  • Which current tools provide an API, webhook, or reliable export?
  • Who owns each workflow and exception?
  • Which messages require human approval?
  • What information is sensitive, and who may access it?
  • How long should inquiry and communication records be retained?
  • What manual fallback should staff use if an automation is unavailable?
  • What current response time and follow-up effort will be used as the baseline?

These answers may change the recommended tool, effort, price, and delivery timeline.


How to Know It Helped

Do not promise a saving before measuring the current situation. Record a baseline first, then compare:

Measure What to check
Inquiry response time How long students wait for the first meaningful response
Follow-up coverage How many open inquiries have an owner and next action
Missed inquiries How many inquiries pass the agreed follow-up time
Admission completion time Time from application start to complete documents
Incorrect reminders Messages sent using outdated or incorrect status
Reporting effort Staff time spent preparing the weekly summary
AI draft quality Factual errors, approval rate, and time saved during the controlled test

The exact target should be agreed only after the current baseline is known.


Practical Next Step

Validate the inquiry workflow with the people who receive, follow up, and manage inquiries. Confirm the channels, required information, statuses, ownership rules, reminder timing, exceptions, and existing tools.

After that validation, the first implementation can be scoped separately. Its proposal should state exactly what will be built, what is excluded, timeline, price, responsibilities, testing, support, and SLA.

The regional research-sprint price is NPR 1 lakh for Nepal-based businesses and USD 1,000 for international businesses. These are separate regional prices, not exchange-rate conversions.

The 10-hour sprint covers two hours of discovery, six hours of research and report preparation, a one-hour walkthrough, and one hour reserved for report revisions and clarification during the 48-hour window. It does not include building the recommended automation.


Walkthrough and Questions

The report includes a one-hour live walkthrough:

  1. The business problems observed.
  2. The recommended first automation.
  3. How each recommendation could work.
  4. The likely benefits and important cautions.
  5. The recommended order and immediate decisions.
  6. Questions and agreed next steps.

For 48 hours after the walkthrough, the client may ask for clarification about the report. New research, additional workflows, detailed technical design, and implementation estimates may require a separate scope.


Prepared by Pratik Kuikel Work directly with a tech founder to understand what is actually worth automating.