10 critical questions to ask ahead of AI implementation

The year 2021 saw a significant growth in interest and implementation of AI in business. Data-driven and machine-learning technologies are continuing to push this field forward and accelerate the efficiency of the modern workplace. 

AI has already made its way into our personal lives as virtual assistants, smart homes, and partially autonomous cars (just to name a few examples). Because this technology has done so much to improve our personal productivity, businesses have grown curious about how this exciting new tech can ramp up productivity.

However, without the correct organizational readiness and preparation, a number of details can hold back AI implementation. From lack of business data to company culture apprehension, internal factors need to be addressed and considered carefully before and during the AI solution exploration process.

Being properly prepared means asking the right questions. You’ll need to have an understanding of the product, processes, ROI, and implementation/installation steps as you begin to narrow down the potential solutions. Here are 10 questions to ask every provider on your short list when considering implementing intelligent automation tools.

1. What use cases does your product offer?

Before you get too far along in your decision making, you’ll need to know if the intelligent automation tools you’re considering solve your specific business problems. This market is vast, with an expected size of more than $360 billion by 2028, so the glut of options can feel overwhelming. Pinpoint your largest problems and ask the shortlisted vendors if they can solve it. Don’t waste your time trying to fit a square peg into a round hole.

2. What types of PoC/trials do you offer?

You need to ensure that any potential product can deliver before you decide to implement it, whether that means using a free trial, a PoC or a short-term contract. Make sure the provider allows you to see the ROI for yourself before committing to a long-term contract. Don’t implement blindly without metrics.

3. How will intelligent automation tools affect my current business processes?

Does this add a new step to your day-to-day workflows or does it eliminate one? The goal with AI implementation is to streamline and make your development team’s jobs easier. If intelligent automation tools will add a step to your process, be careful to evaluate whether they’re worth changing your current processes and will provide value in other areas of your business.

4. How will this product integrate with my current API?

Will this product work with the tools that you already have? If you need a certain CRM or external tool in order for it to work, you’ll need to know that up front. Implementing just one new tool can greatly affect your development team, but if the rest of your staff will need to change how they work with multiple tools in order to use an AI solution, consider overall ROI and if it’s valuable enough to adjust your entire tech stack.

5. What is the installation process?

Is it easy to install? Does IT need to be involved? Who needs to service the product? Does each unit need to be set up individually and/or independently? It’s important to know what it will take to get the product up and running. Identify the stakeholders and teams who will need to assist in installation and get a rough timeline on how long this step will take.

6. What is the learning curve for the end user?

Will your teams need to be formally trained or is the solution simple and straightforward to use? It’s important that you know the complexity, how user-friendly the UI is, and how much effort and time it will take for your team to be productive. Also, ask if they offer any training or onboarding programs.

7. What types of data or tools will your product need to operate successfully?

As you narrow down your shortlist to one or two solutions, it’s important to ask your potential provider what they’ll need from you in order to get the trial or PoC up and running. Most AI tools require data sets to learn from, and to demonstrate results. Here are some follow-up questions that will give you a better idea of what you might need:

• Which types of data does this solution need to have access to?
• Which existing tools need to be integrated with this one?
• Do I need to perform a data audit/clean up ahead of AI implementation?

Asking these critical questions before your AI system is installed will save a lot of trouble later.

8. What security measures are built into your product?

Ensure that both your confidential business data and the data of your customers are under tight lock and key. There are all kinds of security measures that a company could take, whether it be unique encryption keys, blockchain or blind data. Make sure you fully understand their security measures, that no human will have access to your information (unless explicitly granted by you), and that the AI system will not pull data into external systems or databases unless requested.

9. What are the next steps for implementation?

Every solution has a different implementation process and steps that need to be taken to get the ball rolling. For example, at Automation Hero we ask for a point of contact, host a Use Case Discovery workshop, and then agree to PoC terms based on that workshop. Having an understanding of what comes next helps your organization quickly adjust and adopt this technology, and you start seeing results sooner rather than later.

10. Can the solution be customized to my company’s needs?

If your organization has specific needs, it’s important to ask if the solution provider can meet them. Do you need particular workflow models or system integration? Who is responsible for setting up the custom function or feature? Make sure you see the full depth of possibilities before making a final decision.

No two intelligent automation tools are the same, but asking the right questions will ensure you choose the best one to fit your needs.