Years ago, automation and artificial intelligence sounded more like science fiction than science. But today, automation and artificial intelligence (AI) are woven into our daily lives. We are automating tasks that were once performed manually – from smart vacuums to home thermostats that learn our patterns to our streaming services recommending what movies, shows, and songs we may want to enjoy next.
Like the AI we interact with throughout our personal lives, businesses also have immense opportunities to apply automation within business areas. However, AI is still relatively new for many companies, and the journey can be confusing to navigate. So, where do you begin?
Whether you are starting to consider how automation can impact your business or you are already on a path toward AI-led digital transformation, four big ideas are important to consider as a guide.
- Repetitive Manual Tasks: Identify low-hanging fruit opportunities where your team does manual, repetitive (often tedious) tasks. Automating these areas may allow you to upskill & reskill your workforce so they can add value in other higher-value areas of the business. It can also address issues of human error, talent shortages, and keeping your employees engaged.
- Return on Investment (ROI): Identity a desired high-impact gain and frame your business case around identifying areas where automation may provide the highest ROI. Start this identification process by thinking about the metrics that lead up to that ROI—for example, scalability, faster turnaround times, reliability, and accuracy.
- Predictive Insights: Predictive insights is about leveraging your data beyond its day-to-day operational use. Once applied, you’ll gain the ability to gather more real-time data, better data, and make more informed strategic decisions to be proactive vs. reactive. Some examples include the ability to provide customers a personalized experience, enhance current products or services, and leverage current capabilities to break into new markets.
- Competitive Advantage: Determine your company’s risk appetite. Are you looking for opportunities to disrupt your industry, looking for parody, or a fast follower position? If you want to differentiate your business from the competition, start by identifying areas to improve your customer or employee experience.
One approach for identifying automation and AI within your business is applying the Theory of Inventive Problem Solving (TRIZ). TRIZ is a Russian acronym loosely translated to the “Theory of Inventive Problem Solving.” It is a problem-solving method of understanding specific problems and their solutions, abstracting the core problem and solution components, and reapplying them to wholly new, often unrelated, problem domains. TRIZ is useful in learning from disruption in other industries and understanding how it may apply within your own domain.
Real-World Example: Applying TRIZ to work with a manufacturing client
The client had seen high growth over five years, and as they grew, they had to hire, train, and maintain more team members. Growth is wonderful; however, the challenge was that their project-based business model meant each salesperson reviewed hundreds of project bids via email each day, totaling 10,000 emails per month.
Their 12-person team would spend 7 hours each day reading emails, viewing attachments, making inferences, and keying in data into separate systems to find the 14% of relevant bids, thus wasting 86% of their time.
Additionally, due to time constraints, the company followed a “first-in, first-out” process, which led to spending time on projects that might have had a low probability of fit.
After working together to spend time evaluating the challenge, the SafeNet Consulting team implemented an automation solution that would sift through those emails to find the right keywords, key in the data, and score the incoming bids.
- Abstract Problem: poor results, unable to scale and focus on the most important and value work items, wasted effort on low-value or no-value items, employee engagement = fatigue & errors
- Abstract Solution: automate the process of interpreting info, making decisions, and prioritizing work
- Specific Solution: an Sales Automation Scorecard that prioritized the most valuable work and automated human workflows
- Benefits: Reduced waste, prioritized work based on potential value, reduced human error, and ability to engage employees with more rewarding work. Allowed the company to change customer/ employee experience, move on quick turnaround projects, score projects, & move forward with most likely to close.
As a Microsoft Gold Partner, SafeNet Consulting builds on Microsoft Azure to implement AI. There are two types of AI implementation. The first is Azure Cognitive Services, which is a pre-built solution with great “out of the box” tools. The second is Azure Machine Learning or Deep Learning, custom-built models.
Before starting your AI journey, business leaders need to consider these options and which fits their particular needs. First, think about whether you are doing things on-premises, in the cloud, or hybrid. If you’re in the cloud, which cloud? Does it make sense to use an “out of the box” solution, which can come with limitations. Even then, those limitations could change over time.
As you may be realizing, the AI journey is always evolving, and there are many considerations to make, not only before getting started but also in working toward implementation. It’s important to approach an AI project with an agile mindset and be prepared to learn and adapt along the way.
At SafeNet Consulting, we have developed our 7-Step model for designing and executing the project with a primary focus on data. That is what will set baselines, expectations, and goals for your project model.
SafeNet’s Seven Step AI Model
- Define the Business Problem: Working together to find out the problem & what is the business matrix you measure to (baseline).
- Understand & Identify the Data: Look at your incoming data and define it. Is it raw, clean, disorganized, data type (images, text, etc.)?
- Collect & Prepare: Normalize the data so you can work with it in your model
- Prepare & Train Model: Train the model with your collected data and watch if it’s overfitting or underfitting. Train multiple at once or ensemble models to create more powerful models
- Evaluate Performance: Is it hitting the results you want? Move forward or go back? Do we need to add more data? OR are we running the wrong model? May also realize you’re trying to solve the wrong problem.
- Operationalize & Measure: Deploy the model, but don’t only utilize 100% of your manufactured data into the model. Why? Because your bench data might not work.
- Refine & Optimize: Once the model is fully working in production, this is where you monitor & adapt your model. Data changes over time, and models may drift over time.
When it comes to implementing AI, you may not be doing anything yet. Or you might be doing some automation, but you’re missing strategy. You might even be doing it all really well, and you’ve reached the point where you’re wondering how to move to the next step when you can begin to scale, innovate or disrupt the industry.
No matter what stage you’re at, remember to think of the Big Four Ideas to help guide you. Look within your business for areas where your team is performing manual, repetitive, tedious tasks. Think about the metrics that lead to ROI— scalability, speed, reliability, and accuracy. Leverage data to gain predictive insights and gather more and better data in the future. And finally, look for ways to differentiate your business and gain a competitive advantage.
SafeNet Consulting is an IT consulting firm specializing in solving the day’s most complex business problems and guiding clients to the best possible outcome. So if you’re interested in learning more about data, automation, or artificial intelligence that may help you solve a business challenge, we’d love to connect, contact us at email@example.com.