START WITH THE PROBLEM; NOT THE SOLUTION: A FRAMEWORK FOR EVALUATING AI IN YOUR BUSINESS
Written by Jessica Murray
A continuation from last week
Last week, I set the stage for AI, why it’s generating so much buzz and approachable ways to get started with the technology. Today, we’ll continue the topic but shift gears to talk about how you can consider whether applications of AI are well-suited for your business today. This process might also uncover ways AI is already built into the fabric of your operations without you even knowing it.
Asking the right question
I often hear the question: “What AI tool should I use?” or “How can I automate this thing?”
While those are fair questions, they’re not the place to start.
Why?
It comes down to context. New technology brought into your business should enhance your systems and solve a specific problem. If you don’t take a step back to understand why you’re considering AI in the first place, it will be challenging to find the right solution to your business challenge.
Even though AI can be exciting and, as I spoke about last week, there’s a sense of FOMO if you don’t immediately act, I’m a firm advocate of the following: before jumping to the conclusion that AI must be the solution, deeply understand the business problem first.
We’re talking about decision-making around AI applications here, but this sentiment and framework I’m about to outline applies to any decision about new solutions for your business.
Let’s get to it.
A 10-step framework to help you better evaluate AI for your business
Step 1 - Define the problem(s):
Instead of starting with the question: “How do I use AI in my business tomorrow?”
Start with “What’s the biggest pain point in my business right now and why is it causing so much pain?”
Business Problem Example: Let’s say you’re running a rapidly expanding coffee business and one of your recent challenges has been struggling to keep up with delivering fast, effective and personalized customer service. This is becoming more problematic as the customer base expands, product offerings have increased (e.g., online coffee subscriptions) and you’re getting more questions from customers. Falling behind on service has also caused customers to churn at a much higher rate in recent months.
Pain Point: The business is losing customers.
Why? Not implementing a scalable system and processes to support the customer base as it grew.
Step 2 - Prioritize:
If you have a long list of problems (and let’s face it, we all do 🙂), be sure to prioritize and focus your efforts. Think about the relative impact each has on your business and how much effort might be required to fix it. Level of urgency and dependencies should be considered when thinking about impact.
Once you’ve prioritized, start to tackle #1 on the list.
Business Problem Example: In our coffee business example, even though it’s been expanding, the business is struggling to get the adoption it hoped for on new product lines like coffee subscriptions. So, customer acquisition is another challenge on the list of items the owner wants to solve. The owner decides, however, that the churn issue is a bigger threat to preserving the business’s revenue base and solving it needs to remain #1 on the list.
Step 3 - Analyze the current state:
Analyze and map out the existing process so you have a fresh picture of how things operate today, including where the inefficiencies live.
Business Problem Example: For the coffee business, it will be important for the business to outline all of its existing customer interactions and touchpoints. Mapping a clear customer journey will be highly beneficial. Additionally, this step requires looking at existing processes associated with the customer experience, including standard operating procedures, service level agreements or technology.
Step 4 - Define outcomes and success criteria:
Know what success looks like. Set clear, quantifiable success criteria.
Business Problem Example: For the coffee business, churn is top of mind so a primary metric will be monthly improvement in customer retention.
Step 5 - Research solution options:
Now, you can start evaluating different approaches to solving the challenge. Analyze different solutions. AI may be one option, but it’s not the only one.
In doing so, think about:
The cost vs. benefit of each solution.
How each solution fits into your broader business workflows.
Scalability and long-term potential.
Internal capabilities and potential challenges to adoption.
Business Problem Example: There are many potential solutions the coffee business could implement to better serve its customers. A couple of examples include:
1. Integrating a customer service platform on the website with automated chatbots to answer frequently asked questions about the physical location, subscriptions and other products. There’s probably AI mixed in there.
2. Update marketing automation strategies to optimize when and how you message customers via email, text or other mediums.
3. Clean, centralize and improve data available internally about customers to better glean insights about trends and behavior.
Step 6 - Evaluate AI as a solution:
If an AI-powered solution surfaces as a potential option, I would consider the following questions:
Are we dealing with repetitive tasks or a need for pattern recognition?
What data is required and what’s the level of data quality that exists?
Does my team have the skills to adopt this today?
Is this a complete AI solution or an AI and human collaboration?
Are there ways to isolate a small test of an AI implementation?
What resources are required?
Step 7 - Risk assessment:
After weighing options and narrowing down the approach you want to take, it’s important to do a risk assessment. For our purposes, we’ll assume that an AI-enabled solution may be best suited for your needs.
This risk assessment will help avoid several of the common pitfalls I mentioned in last week’s post.
Considerations here include, but are not limited to:
Data privacy: Understand how data fed into AI models is stored, used and managed. Also, be sure to understand how you can control access to data.
Security and compliance: Understand these protocols and make sure they meet your needs. You can also investigate if there have been any prior data breaches and leakage of sensitive information.
Company background: Review the company, its team, reputation and funding. Additionally, evaluate their AI expertise, level of transparency and commitment to supporting the platform.
Ethical and governance risks: Test for bias in systems, and understand the company’s AI ethics principles. Conduct an internal assessment to determine whether implementing the AI solution could negatively impact stakeholders.
Strategic and execution risks: How does the solution align with your business’s strategy and objectives? What could impact your ability to implement and launch the solution?
Step 8 - Make a decision:
You’re now clear on the business challenge and have thoroughly analyzed your options, benefits and risks. It’s time to make a final decision on your solution approach.
Step 9 - Rollout and manage change:
As we all know, introducing new technology and processes into a business isn’t a “build it and they shall come exercise.” You need careful consideration to effectively navigate the change with your team and realize the ROI on your new investment. This includes thinking about a rollout plan, training, how best to drive adoption, etc.
Step 10 - Continuously learn and iterate:
You’ve successfully launched a solution. Congratulations! The work doesn’t stop there. Create a process to oversee progress and periodically assess performance.
Closing thoughts
The above is not exhaustive and this type of assessment can vary greatly depending on the type of business, complexity, regulation, jurisdiction, etc. You should, however, have more confidence in conducting an evaluation and making a decision the next time AI surfaces as a potential solution for your business.
If you’re time-crunched but want to evaluate how to streamline your business, Empower can help you navigate this process. Reply if you want to learn more. You can also get more information about our offerings here.