Skip to main content
Wisest Digital

Six automations, taken apart.

Builds

Working automations we built to show how everyday business tasks can run on their own. Each one is explained step by step, with the route the work travels drawn out from start to finish.

Scroll into it

Read this first

A build is not a client project, and no result is claimed for it.

It is an automation we made to show how a task can run on its own. Real work for clients is under Work, and engagements with measured results are in Case studies.

The six builds

01 / 06

An enquiry form that screens itself

Every enquiry lands in the same inbox, and someone has to open each one just to find out whether the budget fits. Good leads wait while the rest are sorted by hand.

Sales enquiries

See how it works

02 / 06

An assistant that reads your calendar, inbox and sheets

A simple question, such as what is on today or who asked about sponsorship this week, means opening three apps and searching each one.

Everyday productivity

See how it works

03 / 06

A phone line that qualifies callers and flags the hot ones

Inbound sales calls are answered whenever someone is free, notes are patchy, and the most interested caller waits in the same queue as everyone else.

Sales calls

See how it works

04 / 06

Stock that counts itself down and warns before it runs out

Stock levels live in a spreadsheet someone has to update after every sale, so the first sign of a shortage is often a customer asking for something you no longer have.

Stock and ordering

See how it works

05 / 06

Sales and spending, reported every night without a spreadsheet

Sales arrive through several channels, receipts pile up, and knowing whether today was actually profitable costs an evening with a spreadsheet.

Money in and out

See how it works

06 / 06

A WhatsApp assistant that answers the menu and takes orders

A restaurant’s WhatsApp fills with the same questions. Are you open? Where do I park? Is this dish vegetarian? Can I order? Staff answer between everything else, and messages after hours wait until morning.

Customer messages

See how it works

01 · Sales enquiries

Every enquiry lands in the same inbox, and someone has to open each one just to find out whether the budget fits. Good leads wait while the rest are sorted by hand.

An enquiry form that screens itself

The moment someone submits the enquiry form, the budget they entered is checked against a threshold you set. Enquiries that fit receive an email inviting them to book a meeting. Those that do not receive a polite, personal decline. Every enquiry is written to a Google Sheet with the decision beside it, so the team can see everything in one place.

How the work travels
  1. StartsEnquiry form submitted
  2. RuleBudget checked against your threshold

  3. Only one of these happens

    ActionMeeting invitation sent
    orActionPolite decline sent
  4. RecordLogged with the decision

Step by step

  1. 01

    The form collects the name, company, email address, budget and a short note about the project.

  2. 02

    A plain rule compares the budget with the threshold you have set.

  3. 03

    The right email goes out straight away, addressed to the person and their company.

  4. 04

    The enquiry and the decision are added as a new row in Google Sheets for the team to review.

Careful by design. No AI makes this decision. The rule is a threshold you can read and change, so nobody has to guess why an enquiry was declined.

To decide before it goes live. What happens to an enquiry with no budget, or a budget written in words. The rule needs an answer for those before it goes live.

Built with n8n, Form trigger, Gmail, Google Sheets

What arrives · Email sent automaticallyAn illustration, with sample names

To Alex, Example Co

Thanks for your enquiry, Alex

Thank you for telling us about the project at Example Co. It sounds like a good fit for what we do.

Pick a time that suits you and we will talk it through.

Book a meeting

02 · Everyday productivity

A simple question, such as what is on today or who asked about sponsorship this week, means opening three apps and searching each one.

An assistant that reads your calendar, inbox and sheets

A chat assistant you talk to in plain English. It works out which source holds the answer, reads your calendar, your inbox or your sponsorship tracker, and replies with a short summary. It keeps the last ten messages of the conversation in mind, so a follow-up question works without repeating yourself.

How the work travels
  1. StartsYou ask a question
  2. AIWorks out where to look

  3. Only one of these happens

    ActionReads the calendar
    orActionReads the inbox
    orActionReads the sponsor sheet
  4. AIWrites a short answer
  5. RecordReplies in the chat

Step by step

  1. 01

    You type a question, for example "What is on my calendar today?"

  2. 02

    The AI decides which tool can answer it: Google Calendar, Gmail or the sponsorship sheet.

  3. 03

    It reads only what that question needs.

  4. 04

    It summarises what it found into a short, readable reply.

  5. 05

    The conversation is remembered for the next ten messages, so you can ask "and tomorrow?"

Careful by design. It is read-only by design. It can find and summarise, but it cannot send, change or delete anything in your calendar, inbox or sheets.

To decide before it goes live. Which calendars, inboxes and sheets it is allowed to read, and who is allowed to talk to it.

Built with n8n AI agent, a language model via OpenRouter, Google Calendar, Gmail, Google Sheets, Conversation memory

What arrives · Chat with the assistantAn illustration, with sample names
Customer
What is on my calendar today?
Assistant
Two things: a supplier call at 10:00 and a team review at 15:30.
Customer
And tomorrow?
Assistant
Tomorrow is clear until 14:00.

03 · Sales calls

Inbound sales calls are answered whenever someone is free, notes are patchy, and the most interested caller waits in the same queue as everyone else.

A phone line that qualifies callers and flags the hot ones

A voice agent answers the call and holds a natural qualifying conversation. When the call ends, the transcript is analysed and the caller is marked hot or cold. A hot lead sends a Slack message to the sales team with the caller’s details and a summary of what they want. Every caller is saved to a Google Sheet, with hot and cold leads kept apart.

How the work travels
  1. StartsA customer calls
  2. AIVoice agent holds the conversation
  3. RecordTranscript saved when the call ends
  4. AICall marked hot or cold

  5. Only one of these happens

    A personHot: sales team alerted in Slack
    orActionCold: no alert
  6. RecordSaved to the hot-leads sheet
    orRecordSaved to the cold-leads sheet

Step by step

  1. 01

    Vapi’s voice agent answers the call and asks the qualifying questions in a normal conversation.

  2. 02

    When the caller hangs up, the transcript is sent on automatically.

  3. 03

    OpenAI reads the transcript for six things: buying signals, urgency, budget, the problem the caller needs solved, whether they can make the decision, and whether they are ready for a next step.

  4. 04

    A hot lead triggers a Slack message with who called, what they want and the suggested next step.

  5. 05

    Hot and cold leads are saved to separate Google Sheets with the transcript and the call details.

Careful by design. The six signals are written down, and the alert carries a summary of the call, so the salesperson calling back can see why this caller was marked hot.

To decide before it goes live. What callers are told at the start about the AI and the recording, and where transcripts are stored and for how long.

Built with Vapi, OpenAI GPT-4, n8n, Slack, Google Sheets, Webhooks

What arrives · Slack message to the sales teamAn illustration, with sample names

Flagged

Hot lead from an inbound call

Caller
Alex, Example Co
Wants
A quote for their team
Timeline
This month
Can decide
Yes
Next step
Call back today

04 · Stock and ordering

Stock levels live in a spreadsheet someone has to update after every sale, so the first sign of a shortage is often a customer asking for something you no longer have.

Stock that counts itself down and warns before it runs out

Each sale is recorded through a simple form, or a connection to the till, and the item is deducted from the inventory sheet straight away. When an item drops below its own reorder point, the purchasing manager is alerted by email or Slack with the supplier’s details and how much to reorder. Once a week, a report lists the fastest-moving items, the slow ones and the value of the stock on hand.

How the work travels
  1. StartsA sale is recorded
  2. ActionItem deducted from stock
  3. RuleBelow its reorder point?

  4. Only one of these happens

    A personPurchasing alerted, with supplier details
    orActionNothing to do yet

Step by step

  1. 01

    A sale is entered through Google Forms, or arrives from a point-of-sale system.

  2. 02

    The matching product is found in the inventory sheet and its quantity is reduced.

  3. 03

    The remaining quantity is compared with that product’s reorder point.

  4. 04

    If it is too low, an alert goes to the purchasing manager with the supplier contact and a reorder quantity.

  5. 05

    Every week, a scheduled report summarises top movers, slow items and total stock value.

Careful by design. Each product has its own reorder point, so a fast seller and a slow one are not held to the same line.

To decide before it goes live. Whether the system only alerts a person, or also sends the reorder request to the supplier directly.

Built with n8n, Google Sheets, Google Forms, Gmail or Slack, Scheduled workflows

What arrives · Alert to the purchasing managerAn illustration, with sample names

Flagged

Low stock: glass carafe

Left in stock
Below its reorder point
Supplier
Example Supplies Ltd
Contact
[email protected]
Suggested order
The usual reorder quantity

05 · Money in and out

Sales arrive through several channels, receipts pile up, and knowing whether today was actually profitable costs an evening with a spreadsheet.

Sales and spending, reported every night without a spreadsheet

Sales from every channel, in store, online and through delivery platforms, are collected in one place and tagged by product, source and payment method. Expenses go in through a short form with a category, the vendor and a photo of the receipt, which is saved to Google Drive and linked automatically. From that, revenue, cost of goods, margin, net profit and average order value are worked out for you.

How the work travels

  1. All of these happen

    StartsA sale, from any channel
    andStartsAn expense, with its receipt
  2. ActionTagged and filed
  3. RuleTotals and margins worked out

  4. All of these happen

    RecordSummary email at 11 PM
    andRecordLive dashboard updated
    andA personAlert if a budget is exceeded

Step by step

  1. 01

    Sales are gathered from each channel and tagged by product or service, order source and payment method.

  2. 02

    Expenses are added through a quick form, tagged by category and vendor, with the receipt photo stored in Google Drive.

  3. 03

    The key numbers are calculated: revenue, cost of goods, gross margin, net profit and average order value.

  4. 04

    Every night at 11 PM, a summary email shows the day’s sales, expenses and net profit. A weekly summary compares products and channels.

  5. 05

    A Looker Studio dashboard shows the numbers live, and a monthly package includes the profit and loss statement and the expense breakdown.

Careful by design. Receipts are linked to the expense they belong to, so any number on the dashboard can be checked against the paper behind it.

To decide before it goes live. Which spending categories get a budget alert, and who receives the nightly email.

Built with Expense form, Google Drive, Looker Studio

What arrives · Nightly email at 11 PMAn illustration, with sample names

To The owner

Today at a glance

Sales from all three channels are in, and today’s expenses are filed with their receipts.

Supplies went over its budget this week. The full breakdown is on the dashboard.

Open the dashboard

06 · Customer messages

A restaurant’s WhatsApp fills with the same questions. Are you open? Where do I park? Is this dish vegetarian? Can I order? Staff answer between everything else, and messages after hours wait until morning.

A WhatsApp assistant that answers the menu and takes orders

An AI assistant on the business’s WhatsApp number handles the first layer of every conversation. It sends the menu, gives opening hours and directions, recommends popular dishes, filters for dietary needs, quotes prices, and takes an order with the items, quantities, special requests and contact details. When a request needs a person, it hands the chat to staff with the conversation attached.

How the work travels
  1. StartsA customer messages on WhatsApp
  2. AIUnderstands what they need

  3. Only one of these happens

    ActionAnswers from the menu and FAQs
    orRecordTakes the order and confirms it
    orA personHands the chat to staff

Step by step

  1. 01

    A customer writes to the business’s WhatsApp number, at any hour.

  2. 02

    The assistant replies in the language the customer uses. This build was set up for English and Spanish.

  3. 03

    Common questions, such as the menu, hours, location, parking, reservations, prices and dietary needs, are answered from the information it was given.

  4. 04

    For an order, it collects each item, the quantity, any special requests and a contact number, then reads the order back to confirm.

  5. 05

    Anything it should not handle alone is passed to a member of staff, with the conversation so far.

Careful by design. Prices and dishes come only from the menu it was given, and it remembers the conversation, so a customer never has to repeat their order.

To decide before it goes live. Which requests always go straight to a person, such as allergies, complaints and refunds, and who on the team picks them up.

Built with WhatsApp Business, an AI conversation agent

What arrives · WhatsApp conversationAn illustration, with sample names
Customer
Hi, are you open tonight? Is the mushroom risotto vegetarian?
Assistant
Yes, we are open until 22:00, and the mushroom risotto is vegetarian. Would you like to order?
Customer
Two please, to collect at 19:30.
Assistant
Two mushroom risottos for 19:30. Can I take a name and a phone number?

Your move

Doing a task like these by hand?

Describe it in a few sentences. We will tell you whether it is worth automating, and whether it needs AI at all. Often a plain rule does the job.

Describe your taskOur services

Builds are shown to explain how the automations work. They are not client engagements, and no time saving, revenue or other result is claimed for them. Named tools belong to their owners; using them implies no partnership or endorsement.