A practical preparation guide for business owners and managers in Laos who feel that "AI adoption is something for the future, not for us yet."

Leveraging AI does not mean implementing a sophisticated system all at once. By systematically building five foundations—organizing business data, identifying repetitive tasks, explaining the initiative to staff, reviewing the budget, and establishing information management rules—companies of any size can reach a state where they can "start today."

This article clearly explains the five preparation steps while addressing challenges commonly faced by a wide range of Laotian businesses, from small and medium-sized enterprises in Vientiane to retail shops in rural areas. By the time you finish reading, the feeling of "it's too soon" should transform into the confidence of "we can actually do this." No special IT knowledge is required. Let's start together by taking stock of where things stand right now.

The desire to "try using AI" is steadily growing among small and medium-sized business owners and managers in Laos. However, companies that have actually taken action remain in the minority. The primary reason is not a lack of technical capability or funding—it is a lack of preparation.

There are three main patterns of insufficient preparation:

• Data is managed on paper or verbally, and is not in a format that AI can read
• It is unclear which business processes AI could be applied to, so there is no knowing where to start
• Staff feel that "AI will take their jobs," and internal consensus has not been reached

These are not technical problems—they are issues of organizational habits and mindset. Conversely, they are also challenges that can be resolved before purchasing any tools.

Another often-overlooked point is that the attitude of "waiting until everything is perfectly in place" is itself a risk. While waiting for conditions to improve, the gap with competitors and foreign capital quietly continues to widen.

This guide explains, in order, five foundations that can be acted upon starting this week without major investment. None of them require special IT skills, and the content is applicable across a wide range of industries—from small shops in Vientiane to manufacturing sites. Use this as a practical starting point for turning the feeling of "it's too soon" into the conviction that "we can start now."

To leverage AI, you first need "information it can read." No matter how advanced an AI tool you introduce, its power cannot be unlocked if data exists only on paper or in verbal exchanges. Foundation 1 introduces ways to gradually build the habit of keeping daily business records in digital form. No special systems are required—you can start with the tools you already have on hand.

In many Laotian companies, sales records and inventory management are still handled with notebooks and paper slips. If you try to introduce AI tools while things remain in that state, you will quickly run into the wall of "the data can't be read" or "there's no material to analyze." Replacing paper-based information with digital records is the starting point for everything.

The reason the transition often feels difficult is largely that people try to "do everything at once." In practice, it is sufficient to simply start small, as follows:

• Sales records: Enter only the daily sales total into a smartphone notes app
• Purchasing and inventory: Record just three items—product name, quantity, and purchase date
• Customer information: Create a list with three data points—name, contact details, and last purchase date

Using your smartphone camera, you can also save existing paper ledgers as images. Even imperfect data greatly expands the possibilities for AI use, simply by being stored in a searchable format.

One thing to be mindful of is standardizing the recording format. If date formats are mixed—such as "2025/1/5" alongside "5 Jan"—the effort required to organize the data later increases. Deciding on one simple rule at the outset tends to make subsequent steps much smoother.

As a rough timeline for the transition, it is easier to sustain if you begin recording new transactions digitally first, and take the approach of "entering past data little by little whenever there is time." Rather than aiming for perfection, prioritize building the habit of recording even just one entry digitally starting today. The next step will look specifically at where to consolidate the data you have collected.

As a first step toward digitization, expensive dedicated software is not necessary. Simply consolidating records into Google Sheets (free) or Excel—tools already at hand—is enough to begin building a data foundation that AI can work with.

The priority order for records to consolidate is as follows:

• Sales and cash flow: Enter date, amount, and counterparty name, one transaction per row
• Inventory: Standardize product name, quantity, purchase date, and unit price in consistent columns
• Customer information: Compile name, contact details, and purchase history into a single sheet
• Staff shifts and work logs: Record who did what and when

The key is standardizing the format. For example, if dates are entered inconsistently as "2025/1/5," "Jan 5," and "the 5th," AI and aggregation tools will later be unable to read them correctly. Simply summarizing the input rules on a single A4 sheet and sharing it with all staff tends to significantly improve data quality.

The advantage of choosing Google Sheets is that it can be updated from a smartphone with an internet connection, and multiple staff members can edit it simultaneously. Even at rural shops in Laos, as long as Wi-Fi is available, an environment for on-the-spot data entry is within reach.

The initial goal is "to enter one month's worth of data for just one business process first"—whether that is sales records or inventory management. Once a single sheet is filled in, you will be ready to move on to the next foundation. There is no need to aim for a perfect system; the goal is simply to reach a state where "there is at least one set of data in a format that can be handed to AI."

To feel the effects of AI adoption as quickly as possible, it is important to first narrow down where to use it. The simplest starting point for doing so is to write down three repetitive tasks within your organization.

Repetitive tasks are tasks performed using nearly the same procedure every day, every week, or every month. Some examples include:

• Drafting templated emails and quotations for clients
• Tallying and transcribing inventory counts and sales figures
• Writing captions for social media posts
• Sending templated replies to inquiries
• Filling in invoice and delivery note formats

All of these share the characteristic of being tasks where "the content changes slightly each time, but the process follows the same pattern." This is also the area where AI excels most.

Why "three"? If you list too many, priorities never get set, and in many cases nothing ends up being tried at all. By narrowing it down to three first, the discussion of "which one to start with" becomes concrete.

A useful tip when writing the list is to ask the staff members themselves: "How many times did you do the same task this week?" The repetitive tasks a manager assumes exist and the ones frontline staff actually feel are repetitive can differ. By gathering input from the frontline, you can build a list that more accurately reflects reality.

Once you have written down three tasks, note alongside each one "how many times per week it occurs" and "how long it takes each time." This information will serve as the basis for choosing an AI tool in the next step. Having a task list also makes discussions around addressing staff concerns (Foundation 3) and budget considerations (Foundation 4) significantly smoother.

In a five-person company, if one person says "I don't want to use AI," adoption stops right there. In a large corporation, a dedicated promotion team can lead the charge, but in a small or medium-sized enterprise, everything depends on the agreement between the owner and a handful of staff.

Even if the tools and budget are in place, nothing will change if the people on the ground don't move. Particularly in Laos's small and medium-sized enterprises, where the distance between staff and owner is close, "talking directly and getting buy-in" is the most reliable path to adoption.

Here, we introduce two approaches for easing staff resistance: explaining through logic, and getting buy-in through experience. Rather than choosing one or the other, doing both in sequence is the most effective approach.

When staff feel anxious, responding with a general statement like "AI won't take your job" doesn't resonate. In small and medium-sized enterprises, the most reliable approach is for the owner or manager to speak directly with each person individually.

The key to how you communicate is not to explain "the general benefits of AI," but to explain how that specific person's specific tasks will change.

Example conversation: Explaining to B-san, the accounting staff member

"B-san, it takes you a full two days every month to enter the invoices, right? I'd like to have AI read just the amounts and dates, so you only need to do the checking. With less data entry, you'll have more breathing room at the end of the month — I think you'd be able to use that time to sort out the outstanding receivables list you've been wanting to get to."

Framing it this way — focusing on "what you'll be able to do" rather than "what will be taken away" — tends to ease resistance.

Ways to avoid communicating

• ✕ "We'll streamline operations with AI" → Too abstract; the anxiety doesn't go away
• ✕ "People who can't use AI will be left behind" → Threats are counterproductive
• ✕ "We're introducing AI across the whole company" → The scale feels overwhelming

Effective ways to communicate

• ○ "I'd like AI to draft your [specific task] for you" → The scope is clear
• ○ "Let's just try one thing first, and if it's useful, we'll keep going" → They have a choice
• ○ "You'll do the final check on whatever AI produces" → The human stays in control

For a company of five or fewer people, ten minutes with each person is enough. Convey these three points — "why we're doing this," "what will change for you specifically," and "you can stop if you don't like it" — and in most cases the response will be "I guess I'm willing to give it a try."

Rather than explaining in words, actually trying it once makes anxiety disappear faster. However, if the first attempt involves something difficult, it tends to end with "I knew it wouldn't work." The key is to start with simple tasks that have a high chance of success.

What to do next Monday (time required: 15 minutes)

1. Open ChatGPT (free version) on your smartphone
2. Type in the content of a work email you regularly write (a thank-you to a client, a delivery notification, etc.) as if you were explaining it out loud
3. Read the output and copy only the parts that seem usable

That's all. You don't need to expect perfect output. The goal is to get a feel for "so this is what AI is like."

Examples of easy tasks to try (in order of difficulty)

Tasks to avoid for the first attempt

• Sales forecasting or demand analysis → Too difficult; leads to a failure experience
• Translating contracts → Concerns about accuracy make it easy to conclude "this is unusable"
• Analyzing customer data → Should not be done before information management rules (Foundation 5) are established

What to do after trying it once

Just ask the staff member "How did it go?" If the response is "It was easier than I expected," try one more different task the following week. If the response is "It was kind of underwhelming," ask what felt underwhelming and switch to a different task.

Keeping up this pace of "one task per week" for four weeks will naturally build the perception among staff that "AI is a useful tool." Preparing training materials and manuals can wait until after that.

When considering AI adoption, it is not uncommon for the uncertainty of "I have no idea how much this will cost" to stop people from taking action. In reality, however, options are available that allow for a gradual approach — from free tools to paid plans costing a few thousand yen per month. Deciding in advance on a budget ceiling in the form of "how much per month are we willing to try spending" makes the process of choosing a tool significantly smoother. In the next H3 section, we will introduce specific options broken down into the free tier and the paid tier of approximately $20–50 per month.

If you want to get a feel for AI without spending money, starting with free tools is a practical first step. Since you can begin with zero upfront investment, the reassurance of "nothing to lose if it doesn't work out" can also help lower resistance among staff.

What You Can Try with ChatGPT Free

ChatGPT's free plan covers a wide range of text-based tasks.

• Drafting emails and documents: Generate draft messages to business partners or thank-you letters in Japanese, English, or Lao
• Brainstorming: Quickly list out catchphrase ideas for new products or rough concepts for promotional campaigns
• Organizing operational manuals: Enter procedures that were previously communicated verbally, and the tool will format them into easy-to-read bullet points

It's worth noting that there are limits on usage frequency and response speed, but for a few uses per day as a work aid, it tends to be sufficient in most situations.

What You Can Try with Google Translate / Google Docs

Google Translate is a free AI translation tool that supports multiple languages including Lao, Thai, English, and Chinese. For companies that frequently handle import/export operations or serve foreign tourists, there are reported cases of it reducing working hours from day one.

• Google Docs "Smart Compose": An AI feature that suggests word completions as you type is built in as standard
• Google Sheets formula assistance: A feature that explains the meaning of formulas in natural language is also available for free

We recommend trying these free tools for one week, limited to specific tasks. Getting a hands-on sense of "what kinds of instructions are actually useful" will give you the information you need to decide on your next step. There's no rush to move to a paid plan — it's perfectly fine to wait until you've gotten a feel for what works with the free version first.

If you can secure a budget of around $20–$50 per month, a range of options opens up that can dramatically increase your sense of improved operational efficiency. As a next step after finding the free plan "seems useful," it's well worth considering upgrading to a paid plan.

ChatGPT Plus (approx. $20/month — reference value at time of writing; check the official page for current pricing)

Compared to the free version, response speeds tend to be faster and higher-accuracy models are available. The following are examples of use cases that are easy to envision for businesses in Laos:

• Drafting emails to business partners in English or Thai
• Creating the framework for product descriptions or proposals
• Summarizing and organizing internal manuals

In situations requiring multilingual support in particular, the difference from the free version tends to be noticeable. If there is a staff member who uses it many times a day, converting the monthly cost into time saved per task makes it easier to assess cost-effectiveness.

LINE Official Account (LINE OA)

LINE has a large user base in Southeast Asia, including Laos, and in many cases it is already established as a means of communication with customers. Using a paid LINE OA plan increases the message delivery limit, making it easier to manage customer announcements, reservation handling, and after-sales follow-up all in one place. Pricing varies by plan, so please check the official website for the latest information.

The Idea of Combining Both

• Use ChatGPT Plus to create the "raw material" for text and responses
• Use LINE OA to automate "delivery and responses" to customers

Keeping this division of roles in mind makes it easier to cover both the front and back end of operations even on a limited budget. Trying just one first, then adding the other once you're comfortable, is a realistic approach for getting it to stick while minimizing disruption on the ground.

Before starting to use AI, one thing many companies tend to overlook is establishing "rules for handling information." Because generative AI tools like ChatGPT generate responses based on the text you input, the level of risk varies depending on what you enter.

Start by writing down on a single sheet of paper what information is "acceptable to enter" and what is "not acceptable to enter." There's no need to overthink it. One rule is enough to start.

Examples of information acceptable to enter:

• Translation or summarization of general text (not containing proper nouns)
• Drafting internal manuals
• Creating templates for product descriptions or informational emails

Examples of information not acceptable to enter:

• Personal information such as customer names, phone numbers, or addresses
• Contract amounts with business partners or unpublished pricing information
• Employee salary or performance evaluation data

Simply deciding on a single statement — "do not enter customer personal information into AI" — enables all staff to act without hesitation. Complexity makes rules impossible to follow, so simplicity is key.

A common situation in the workplace is copying and pasting an email directly into the tool "because it's convenient." People may send it without noticing that customer names or transaction terms are included. To prevent these kinds of mistakes, it's a good idea to make it a habit to include a confirmation step of "check for personal information before entering."

Note that the data handling policies of each AI tool are subject to change, so it is advisable to check the official documentation periodically. Rules are not something you "set and forget" — building a habit of reviewing them every three months will steadily strengthen your organization's trustworthiness.

Let's confirm how well the five foundations covered so far are in place at your company. We recommend printing or copying the checklist below and reviewing it together as a team.

Foundation 1: Digitization of Business Data

• At least one of the following — sales, inventory, or customer information — is recorded digitally rather than on paper
• More than one month's worth of data has been accumulated in Google Sheets or Excel
• The person responsible for data entry and the update frequency have been decided

Foundation 2: Inventory of Repetitive Tasks

• Three or more repetitive tasks performed weekly or daily have been written down
• At least one of them feels like something that "could be automated or made more efficient with AI"

Foundation 3: Addressing Staff Concerns

• The understanding that "AI is not something that takes away jobs" has been shared across the team
• Staff have already experienced trying at least one small AI tool

Foundation 4: Setting a Budget

• A trial AI budget per month (anywhere in the $0–$50 range is fine) has been decided
• Based on the results of trying free tools, the next step has been considered

Foundation 5: Establishing Information Usage Rules

• At least one rule has been decided regarding "what information is and is not acceptable to enter into AI tools"
• A means of communicating that rule to staff (verbal, written note, chat, etc.) has been decided

Assessment Guide

• ✅ 9–11 items: You are ready to introduce AI. Move on to selecting specific tools.
• ✅ 5–8 items: More than halfway there. Aim to complete the unchecked items within one week.
• ✅ 4 or fewer items: Don't rush — the quickest path forward is to start with Foundation 1 and Foundation 2.

It's perfectly fine if not every item is checked. Finding "one item you can act on today" is your very first step.

Once the five foundations are in place, it's time to move on to the practical phase of "actually integrating AI into your operations."

Articles to Read First

The next thing for companies that have completed the preparation stage to tackle is designing a concrete implementation plan. We have prepared an article covering the practical implementation process in detail — from visualizing business processes and conducting a PoC (Proof of Concept), to building a cloud environment and operating a hybrid model of AI and human work.

→ AI Adoption Guide for Laos-Based Companies — 5 Steps to Achieving Operational Efficiency

This article (the preparation edition) covered laying the groundwork before implementation — things like organizing data, identifying repetitive tasks, and gaining staff buy-in. The article linked above (the practical edition) explains how to build AI on top of that foundation — covering infrastructure selection, multilingual support, and performance measurement across 5 steps.

If You Want to Know About Use Cases by Industry

If you have a sense of the direction for preparation but want to know "how this can be used in my specific industry," please also refer to the following industry-specific articles.

• Cross-industry comparison of tourism, logistics, agriculture, and manufacturing → How AI Can Be Useful Across Laos's Four Major Industries
• Concrete examples of getting started today with just a smartphone → AI Applications That SMEs in Laos Can Start Using Today

A good rule of thumb for when to move on to the next article is once you have checked off three or more items on the checklist. There's no need to wait until everything is perfectly in order.

Q1. Our staff don't have enough IT skills to use AI effectively. Can we still start preparing?

Yes, you can. Many of today's AI tools are designed to be simple enough to operate on a smartphone. The most practical first step is to try a chat-based tool where you simply "type text and receive a response." No advanced programming knowledge is required.

Q2. Most of our data is on paper. How long will digitization take?

It depends on the scale of your operations, but you can start on day one simply by switching to a policy of "recording only new transactions digitally from today." We recommend deprioritizing the back-entry of historical data and instead building habits around your current, ongoing work first. Rather than aiming for perfection, prioritize creating a system you can sustain.

Q3. Are there AI tools that support the Lao language?

ChatGPT and Google Translate both support Lao language input and output. However, accuracy may be lower compared to English or Japanese in some cases, so we recommend establishing an operational rule that requires a staff member to review the content of any important documents. Please check each service's official page for the latest information on supported languages.

Q4. Are there benefits to adopting AI even for a small company?

In fact, smaller businesses tend to feel the benefits of automating repetitive tasks more acutely, precisely because each individual carries a heavier workload. Many cases have been reported where room for improvement is found within day-to-day operations, such as drafting responses to inquiries or handling simple translation tasks.

Q5. I'm concerned about security. How much information is safe to enter?

As explained in "Foundation 5" of this article, it is important to establish rules from the outset that prohibit entering customers' personal information, contract amounts, or undisclosed business plans into AI tools. The safe approach is to start with low-confidentiality use cases, such as rephrasing general text or performing translation tasks.