My CEO wants me to ‘do AI’. Where do I even start

by Brett White
VP Products, Clickatell

Your CEO just said, “We need to do something with AI.” You’ve got budget. Maybe even a vague deadline. But no roadmap. 

Sound familiar? You’re not alone. 

Before you dive into the hype and start integrating the latest LLM or productivity tool-of-the-week, take a step back. Here’s a simple, structured approach to help you think clearly, act strategically, and make AI work for your business. 

1. Pause🫸

Take a moment to decode the ask. 

“Do AI” can mean many things. Does your CEO want to integrate AI into your internal workflows to boost productivity? Or are they talking about enhancing your product or service with AI to improve the customer experience? Clarifying the objective upfront will save you time, budget, and unnecessary detours. 

2. Understand 🧠

Do your research. 

AI isn’t a single tool or platform—it’s a vast, rapidly evolving space. Understanding what’s possible (and what’s hype) will help you narrow in on realistic, high-impact opportunities. 

2.1 AI for Products 

We’re in the middle of an LLM (large language model) gold rush. Every week brings a new contender, recently, DeepSeek and Manus, and established names like OpenAI, Google, Microsoft, Meta, Anthropic, and X AI are pushing boundaries. These models can understand, generate, and interact using natural language, images, and increasingly audio and video. 

2.2 AI for Productivity 

AI-enhanced tools can make your team faster, sharper, and more effective. Think chat-based tools like ChatGPT, Gemini, Claude, Copilot, Meta AI, and others that offer expert-level assistance across disciplines: strategy, coding, legal, marketing, finance, and beyond. You’ll also find domain-specific AI tools like GitHub Copilot for development, Figma with AI design features, and countless more arriving daily. 

3. Assess 🔎

3.1 Product Opportunities 

Where can AI meaningfully improve your customer experience? Do you want to: 

• Reduce support overhead? 

• Save customers time? 

• Improve usability? 

• Increase conversion? 

• Lower churn? 

LLMs are excellent at interpreting human language (voice and text), analyzing images, and generating content, such as text, video, audio, and code. They can make your product smarter, more responsive, and more responsive and delightful to use.. 

3.2 Productivity Gains 

Where do you need to accelerate delivery and free up capacity? One obvious area is software development. Microsoft CEO Satya Nadella says AI now writes 20–30% of their code. CTO Kevin Scott expects that to reach 95% within five years. At Google, Sundar Pichai has stated AI now contributes to 30% of their code, up from 25% just a year ago. These are leading indicators that AI is transforming how work gets done, not just in engineering, but across medicine, science, design, and more. 

3.3 Budget Reality Check 

Don’t skip the math. AI tools can be expensive, especially at scale. Know what problem you’re solving, what it will cost, and what the return might be. Build a business case to avoid the trap of investing in “cool tech” without clear ROI. The hype is real, but so is the bill. Enterprise licensing and team-specific tiers vary widely. Some tools are free for individuals but scale up sharply for teams, so model your rollout accordingly.  

4. Implement ▶️

Start small. Be focused. 

4.1 AI for Products 

• You’ll need to choose how to implement AI. Do you: 

• Self-host an open-source LLM like Meta’s LLaMA? 

• Use a hosted API like OpenAI’s GPT or Google’s Gemini? 

Each option comes with trade-offs in complexity, flexibility, and cost. The landscape is changing fast, so focus on quick wins and keep your implementation adaptable so you can pivot when the next breakthrough arrives. 

4.2 AI for Productivity 

Identify the functions that will benefit most from AI acceleration and start there. Productivity tools often come with per-user licensing, so be mindful of how many team members need access. Some vendors offer enterprise pricing or discounts for larger teams, which is worth asking about before you commit. 

5. Learn 📚

Before deploying your shiny new AI-powered solution, define your success metrics. What are you trying to achieve? Set KPIs so you can measure whether your initiative is having the intended impact and then iterate. 

Whether you’re improving CX or boosting internal efficiency, success with AI isn’t just about launching a tool. It’s about learning quickly, adapting fast, and continuously improving. 

Final Thought 

AI isn’t a silver bullet, but it can be a serious advantage. You don’t need to boil the ocean. Find your leverage point, test fast, learn faster, and then make it real.