Building Business Systems with AI

Preface

The SME management team was buzzing with excitement. They had just gotten the go-ahead to invest in a new sales management system. They imagined sleek new dashboards, powerful features, and a big change in how they handled the sales process. But six months later, things looked very different. The system was barely used and sat idle. Employees had gone back to using Excel for reporting, causing double work for each client. What once seemed like a great system had become a source of frustration, filled with data errors and inefficiencies.

Introduction

"Systems" are crucial tools often overlooked in an SME. If poorly implemented, the expensive business systems purchased can turn into cumbersome tools for employees and be seen as wasted investments by bosses. However, it is undeniable that business systems are the foundation for scaling up a company.

With AI nowadays, I see this as an opportunity for employees without technical backgrounds to participate in the design and optimization of these systems. Many small and medium-sized companies, due to a lack of resources to hire or purchase professional systems, can now rely on experienced employees to build or optimize customized systems without spending big bucks.

Be Part of "Business System" Construction

As part of the company's operations, your advantage is that you understand the company's operational problems better than programmers or sales consultants. In the beginning, you don't need to build a complex system like CRM, WSM, or POS yourself. As long as your solution has the following characteristics, you can have your hand-crafted version 0.1 business system:

1. Standardization: Classify all foreseeable situations and unify them to simplify complex workflows into a set of repeatable and predictable steps. The purpose of standardization is to ensure consistent results every time the same task is performed.
2. Scalability: After standardization, ensure that these standards can be extended to a larger scope. In other words, even when dealing with more customers, orders, or projects, the same standards can be maintained.
3. Automation: Delegate repetitive and highly regular tasks to machines and focus your efforts on tasks that cannot be standardized and scaled. Tools like Google Apps Script in Google Workspace can be used to automate common tasks like email sending and data processing. Tools like Zapier can also help achieve cross-platform workflow automation.

Employees often experience burnout from prolonged "firefighting" work. When a problem ("fire") occurs, if you only solve the immediate issue ("extinguish the fire") instead of preventing and systematically solving problems through "standardization" and "scalability" ("fire prevention" and "best firefighting path"), you will end up firefighting every day. In the long run, you and your team will just be spinning in circles at the same stage.

How AI Can Help

Using ChatGPT for Problem Statement Review

1. Draft an initial problem statement describing the inefficiency in your process.
2. Train ChatGPT by providing it with this statement and asking it to challenge the assumptions and suggest alternative viewpoints.
3. Refine the problem statement based on ChatGPT's feedback, ensuring it is clear and precise.

Drafting Process Flows with AI Insight

1. Develop a draft of your ideal process flow, considering current challenges and desired outcomes.
2. Feed this draft into ChatGPT along with detailed context and objectives.
3. Use ChatGPT's feedback to identify potential weaknesses or overlooked aspects in the process flow.

Coding and Debugging Assistance

1. Outline the tasks you want to automate using Google Apps Script or Zapier Webhook Action.
2. Ask ChatGPT to help write the necessary JavaScript code, providing as much detail as possible about the required functionality.
3. Test the provided code and use ChatGPT's guidance to debug and fine-tune it, ensuring it meets your needs.
4. Expand your code to do more!

Any Ideas to Kickstart the Journey?

Based on my experience in sales management, many low-hanging fruits in a business system focus on the following three major issues:

1. Conversion rate of sales channels: How high is the conversion rate of your sales channels? At which stage do you see the most drop-offs? Which channel performs best?
2. Sales cycle: How long is your sales cycle? Which stage should be optimized or cut off?
3. Sales performance contribution: Are your salespeople maximizing their performance contribution? Why 20K and not 25K? Is it a problem of the quantity or quality of leads?

A super simplified example (data is not real):

1. Conversion rate: We found that the lead conversion rate recorded on weekdays is 150% of that on weekends. After analysis, we discovered that leads on weekends usually wait until Monday to be assigned to sales. Using Zapier, we automated the lead assignment process, so even after office hours, leads are automatically assigned to the sales team. This change improved the conversion rate of weekend leads by 40%.
2. Sales cycle: We found the sales cycle to be too long, especially between initial contact and signing. By standardizing the initial contact process and using Google Apps Script and Zapier to automatically send follow-up emails and WhatsApp reminders, we shortened the sales cycle by 30%.

Afterword

This is my first long article about work, inspired by Sunny Cheng from Satire Entertainment who shared his experience of creating an automated system for managing company contracts. With AI, you are not required to have a computer science degree to start building a business system. As I quote Jensen Huang:

We have democratized computing. It doesn't matter if you're a farmer, a doctor, a nurse, a frontline worker, travel agent, doesn't matter, a small business, a restaurant owner, it doesn't matter. Everybody is now a programmer, you just have to prompt this thing to write a program for you, to do something for you, automate something for you.

Start writing down your problems, grouping, breaking down, and questioning every requirement, then use AI tools to standardize processes and achieve scalability and automation. It's really fun!

Originally published on LinkedIn. Want to discuss this further? Get in touch.