How to Scaling Sales and CX Conversations with AI

Credit Unsplash

I recently heard a story of an AI CX chatbot that successfully handled over 2.3 million conversations, essentially doing the work of 700 full-time staff. This company spent a few months working with OpenAI to build their custom, ai-powered support chatbot. Now they are projecting this AI to bring $40 million in profit for 2024.

That’s amazing?!

Or is it scary?

Regardless… it’s 100% true.

Published on February 27, 2024. (link at the bottom) And it’s only the beginning!

Scaling Sales and CX Conversations with AI

If you want to scale sales or CX conversations with AI, then keep reading.

Today I want to camp on a topic near and dear to my heart.

If you don’t know my background, here’s the brief: I started my first (or second) digital marketing company around 2008. By late 2016, I heard of the global shift towards social media messaging and the advancements of chatbots, and in 2017 I started Chatbot Funnels. I mastered, consulted, and taught how to use chatbots to speed up sales and marketing funnels.

After COVID kicked my butt and I left the chatbot world, but then I returned with the rise of ChatGPT.

Now I help enterprises who want to add AI-powered calling and QA to their call centers. However, I’m still passionate about speaking, writing, and educating on integrating generative AI into our workflows and operations.

Within my new book, The AI Strategy Playbook, I highlight multiple strategies or plays that leverage AI for sales, marketing, and CX. Though AI-powered chat, calls, and messaging are powerful, you need to understand when to use it and how to set it up properly.

How to Set Up AI Sales / CX Chatbots Properly

When you have masses of sales or customer inquiries with predictable and recurring steps to resolve, you’ve got a perfect opportunity to scale with AI.

Unlike chatbots in the past where you had to define every single action, response, and option, AI chat can “catch” and respond to many questions without direct programming. As long as its core prompt is accurate and it has ingested enough supporting content, it should know to complete any assignment.

To set up a basic AI chatbot you need to focus on these 3 steps:

1: Feed / train the AI with the most common questions and basic facts about your company.

2: Build / write the AI’s prompts to engage and collect the most critical information, based on the current stage of the customer journey.

3: Test and tune internally with guardrails to avoid any hallucinations.

Soft launch in a limited and controlled environment and continue to iterate; never just set it and forget it.

These steps will benefit you regardless of the AI or chatbot platform you use. The work is in the prompting and training, less in the copywriting and sequencing of the conversation.

However, there are key differences between AI chatbots and agent platforms. Integrations, LLMs, channel integrations, and latency can vary greatly.

The company’s practices in support, compliance, and awareness of telephony, and SIP standards, can also be great assets or a huge risk. ( I just saw an ad on FB about an AI agent making robocalls and cringed. The company isn’t aware of the pending FCC regulations on AI calling. )

So yes… you need to align your strategy with a company you trust that understands how to balance performance ROI and compliance awareness.

Lastly, there is still conversational design and human behavior considerations to plan appropriately.

Even with AI doing much of the talking, I personally guide the AI with conversational design principles to maximize results.

Photo by CDC on Unsplash

*You might not be ready to automate sales convos with AI…

Though I spend all day/every day testing and teaching AI, I understand that it can’t solve every problem. One of my favorite quotes in adopting AI is from Eli Goldratt.

“Technology can only bring benefit, if, and only if it diminishes a limitation.” — Eli Goldratt

AI can only benefit you when you understand your business’s limitations and you apply AI directly to unlock that limitation. If you don’t have a limitation of say… speed to lead, or speed to respond to support requests, then an AI agent/chatbot isn’t going to really help you.

One type of AI could help increase awareness. While a different type of AI could help respond and engage in conversations. These are not the same.

4 things I look for before suggesting AI-powered chat:

1. You’ve got a steady flow of traffic
2. You’ve got a fully functional marketing and support process
3. You’ve got people to execute your process, but the costs are concerning
4. You’ve got requests, needs, and opportunities that are falling through the cracks because of resource limitations

Now, to put this a different way…

✓ I look for success

✓ I leverage the successful processes and turn them into repeatable solutions

✓ I scale the successful process with AI

I don’t take an “idea” and try to use AI to INVENT something that’s never been done before. That’s R&D. Expiriemting. I’m sure there are people out there who will take your money and experiment for 3–6 months on a project. But not me. Not now.

I’d rather meet with a successful business owner, who has something that’s working… take their 100k funnel, and use AI to add $20k. Not take someone’s dream and prayer and try to go from $5k to $10k.

I will say, that with my book, The AI Strategy Playbook, anyone can learn what’s possible and probably for them. But with my AI sales and CX chat/agent software solutions… those are more for mature companies.

To recap and close…

For AI Chat/AI Call Center Agents/ or Chatbots, these are best for someone who has:

1. A call center staffed with 20+ agents (or)
2. A flow of 100 to 1000 leads every DAY or WEEK and wants AI to help engage, close, support (or)
3. A database of 20,000 or more past customers and wants AI to automatically re-engage.

If you’d like to learn more and possible work together, please book a call here: https://cal.com/arvellcraig

Create a Great Day
Arvell Craig

P.S.

**Here’s the link to the story and company whose AI chatbot handled (in one month) 2.3 million requests, replaced 700 workers, and is projected to add $40 million to the bottom line in 2024.