Best Practices For Effective AI Augmentation

Hi there,

The more we force people to do something, the more resistance we get.

This is also true in business - or should I say, especially in business? I saw a company struggle to "convince" their employees use the new CRM correctly - for more than 5 years.

Pushing hard in one direction just because leadership wants it, is a great way to kill innovation. Don't let that happen with your AI initiative. The key is to bring the technology to people, not the other way around.

Let's see how this works!

Augmentation is the easiest path to reap AI benefits

When most people hear AI, the movie in their head shows robots and automation.

(That's also the main reason why works councils are typically not very enthusiastic about the new AI tool you're proposing.)

However, as I explained here last week, effective AI augmentation beats AI automation by a wide margin - especially if you're just starting out and looking for internal use cases to boost productivity.

In fact, I believe that successful AI augmentation is the easiest way to reap the proclaimed benefits of AI, as described in a recent Havard Business School study:

• 12.2% higher work output

• 25.1% faster task completion

• 40% higher quality work

But AI augmentation doesn't mean buying a Microsoft Copilot license or a ChatGPT Enterprise subscription and calling it a day.

There’s a bit more to it.

It requires identifying the right workflows and augmenting them with AI in the most effective way.

Best Practices for AI Augmentation

Here are 5 best practices I've seen successful companies follow:

1) Start With What You Know

In essence, this is about integrating AI into your existing workflows. Focus on familiar processes that inherently have high business value (such as sales, marketing, or customer support) and are facing a critical pain point (current problem) or bottleneck (future problem).

Take your marketing planning, for example. As the saying goes, half of your ad spend is wasted, but you don't know which. Now imagine that when you are planning your marketing budget, for example in Microsoft Excel, you could access a little helper that gives you the estimated reach for different channels like social media, email, or display ads, given a proposed budget. This would help you quickly gauge marketing ROI and create a data-driven campaign plan - without even switching your primary tool.

Keep in mind that this type of workflow augmentation doesn't necessarily involve a lot of integration from the get-go. A simple copilot in Excel, a chatbot in teams, or a custom GPT in your web browser could be enough.

2) Enhance, Don't Replace

Once you have identified a workflow, think about improving rather than replacing first.

I know, if you use AI in an inherently poor bad process, you end up having an AI-powered poor process, but hear me out! Typically, this type of workflow improvement is a journey. Most companies fail by trying to "tear it all down" and starting from scratch. These initiatives often die on the vine because they are simply not feasible, too expensive, or lack the necessary stakeholder support.

Take sales calls, for example. In a perfect world, your sales reps would use a phone platform that works directly from your CRM, automating call recording, note taking, and real-time insertion of relevant conversations during the call.

However, many companies still use separate call center software alongside their CRM. So instead of starting all over from scratch, you can begin with the call center software's workflow. Transcribe and save the sales calls, and include a brief summary at the end of each call that pre-fills some required CRM fields. After the call, the sales rep can easily review and send it to the CRM with a single click.

Whatever workflow improvement you make, test it on a small scale. Start with a small number of reps, customers, or other cohorts and measure the difference-not just in hard terms like time or money saved, but also in qualitative factors like employee or customer satisfaction.

3) Quality Over Quantity

More is not always better: More sales calls, more emails answered, or more support tickets resolved. In most situations, it's actually better to prioritize quality over quantity.

Imagine a customer service scenario where an AI chatbot is trained to handle the most common inquiries with precision. This setup ensures that customers receive quick, accurate answers, and your team can focus on the more complex cases that require a human touch.

Overall, this approach wouldn't answer more customer questions or dismiss your support agents. Instead, your support agents would have more time to do higher-quality work-solving the tough cases.

In fact, a recent study by Salesforce discovered that 64% of agents with AI chatbots were able to spend their time solving complex problems, versus 50% of agents without AI chatbots.

And Hubspot found that 78% of customer service professionals agree that AI helps them spend more time on high-quality aspects of their job.

What can you do today? Take a look at your customers' most frequently asked questions and see if an AI chatbot can automatically answer them with high reliability. If so, set up a feedback loop between your support team and the chatbot to review and improve its answers, which can then be used to help customers on a larger scale.

4) Human-Centric AI Design

AI should feel like a natural extension of your team, not another entity they have to wrestle with. The goal is to design AI tools that fit smoothly into your existing workflows, enhancing rather than complicating them.

The most sophisticated AI tool is only as good as the team using it. Investing time in training and adapting your processes to incorporate AI is crucial.

For example, you want to make sure that workers see AI as a helper, not the one in charge. Avoiding the misconception of blindly following AI's instructions is critical. The goal is to create a bridge between human intelligence and artificial intelligence, where both collaborate to achieve optimal results – and where the human is not just in the loop, but in the lead.

Before implementing a new AI solution, talk to your team about what an effective workflow improvement would look like and which workflow they believe would benefit the most. Can this task be made easier, take less time, or provide more value to related processes?

5) Iterative Improvement

As I've mentioned here many times before, the secret to starting with AI is to start small. Pick a process that could benefit from a touch of AI, build a prototype rapidly, get feedback, and create a roadmap accordingly. Rinse and repeat. This iterative approach ensures that AI truly meets your needs and grows with you.

Getting on an AI journey is a shift that requires more than just adopting new technology; it demands a mindset open to experimentation and learning from each outcome.

Conclusion

Each of these steps is a building block towards a more efficient, AI-augmented workflow.

Remember, integrating AI into your business isn't about taking a giant leap into the unknown; it's about taking thoughtful, measured steps that enhance the work you're already doing.

Start small, focus on quality, and always keep the human element at the heart of your AI strategy.

If you need help figuring this out, just reply to this email and let's see if I can help.

Otherwise - see you next Friday!

Tobias