Image Credit: Igor Kutyaev/Getty Images
Check out the on-demand sessions from the Low-Code/No-Code Summit to learn how to successfully innovate and achieve efficiency by upskilling and scaling citizen developers. Watch now.
Three things are certain in life: death, taxes and artificial intelligence (AI). Not only is AI significantly less depressing than the other two, it’s also produced a number of business innovations and is becoming more affordable as a mass solution. In the startup space, AI has helped predict Covid variant trends, power military tools, and prevent burnout among doctors.
Beyond the “sexier” applications of AI, the tech has boomed in a sector that quite literally drives daily life: logistics. Here, AI has been optimizing delivery routes, reducing times for last mile delivery, boosting sustainable measures and lowering operational costs. I know this first-hand because I designed the AI for my logistics startup, having initially created an algorithm for my master’s thesis that planned routes for firefighters. This algorithm saved 1,400 lives and reduced time-to-arrival in some of the world’s most congested cities by 40%.
The advantages of AI are undeniable. However, companies often shy away from fully committing to it because they believe that it’s too complex or costly to integrate. Of course, in light of the current unstable market, businesses need to double down on efficiency, but it is still possible to embrace AI and not send shockwaves to your accounting department. With that in mind, these are my tips for Small and Medium-sized Businesses (SMBs) hoping to move their last mile into the AI world, and to make a long-term home there.
SMBs: Make sure your foundations are right for AI
It sounds obvious, but every company needs to first identify that they have an actual need for AI before firmly making it part of their model. In the last mile, that means asking yourself if your customers want personalized delivery — for example, if they want to be able to select the times they receive goods — or if they’re content with more standardized processes. If there isn’t a demand for nuanced delivery, AI may not be the right route for you, as AI’s specialty lies in its ability to cater to multiple varied outcomes.
Intelligent Security Summit
Learn the critical role of AI & ML in cybersecurity and industry specific case studies on December 8. Register for your free pass today.
Next, take a look at your customer behaviors and expectations. Are they changing on a daily basis or are they generally consistent? If their preferences are fixed (for example, when and how they receive deliveries stays the same), AI won’t be as beneficial to your business. AI is valuable for detecting and understanding patterns in datasets, so if you already have a clear interpretation of your customers, AI won’t be able to tell you anything new.
For the final sense-check stage, turn to your existing tech. If you don’t have intelligence software in place to begin with, skipping ahead to AI could cause you problems. Ideally, you need some automated, smart processes taking place so that you can scale them up using AI. Remember, AI isn’t the end result, it should be an accelerator among your pre-fine-tuned practices.
Most SMBs will opt to leverage AI through third-party tools, which makes sense as building your own AI from scratch essentially means becoming a software company. That said, even if harnessing others’ AI, you’ll need to construct a team to manage the tech — that means data scientists, people who know what to do with AI output, how to measure it and how to seamlessly assimilate it in workflows. The more tech-oriented your team is, the quicker and more seamlessly you’ll be able to integrate AI.
Make a toolbox to cultivate your AI
AI isn’t a “set it and forget it” solution; you’ll need a comprehensive toolbox to power and measure its effectiveness from day one. Fortunately, because of AI’s prominence in business, there are a host of tools to keep your AI in check.
Let’s start with the basics. Over the past decade, the most common elements of AI have been packaged and made more accessible for a range of industries. One of the most popular AI tools is TensorFlow which is great for bundling and building AI — the core open source library aids in training machine learning models and can be run directly in your web browser. Meanwhile, Python is a common AI programming language, and R helps data scientists scale and align with different AI models.
Beyond these tools, you need to ensure that you are regularly gathering feedback from the real people using the AI. Take care to recalibrate the algorithms accordingly. It’s all well and good having the tools to fix a car, but if you don’t know how to use them to accommodate the driver, they’re of little value. At SimpliRoute, we ask all our delivery personnel to rate the routes our AI recommends them on a scale of 1 to 5. This quantitative information is then used in tandem with qualitative data (such as surveys) for us to more deeply recognize what does and doesn’t work with the AI.
Prep data to be your long-term AI energy supply
Becoming an AI company means entering into a long-term relationship. AI won’t serve your SMB or your users if it’s stagnant — it has to be dynamic, combining historical and real-time data to generate insights. That’s why roughly 80% of your last mile spending will shift to collecting, pulling and fixing the data that powers your AI and keeps those insights coming.
However, data needs maintenance. You should constantly be retrieving data from multiple sources to guarantee that you have the fullest picture possible of your last-mile operations. For example, we need a lot of GPS data, but we also need service information about the time needed to unload trucks and what drivers’ preferred routes are. You can’t cherry-pick the data that confirms what you already know (or want to know). Your data should be genuinely representative for your AI to be most effective.
Be conscious not only to invest in data resources but in data people too. You’re going to need training around emerging AI trends and models for current staff, as well as any new members you bring on to manage AI. If you are hoping to grow your AI team, partner with universities to attract cutting-edge talent or offer internships that detail why your AI application is unique – having a mix of the business and academic world can do wonders for your AI standing.
At the same time, data shouldn’t be siloed to just the departments where AI is at play – it should be influencing decisions across the whole company, in your marketing teams, sales funnels and more. If data isn’t put at the heart of all decision-making, you’ll never truly step into your end-users’ shoes and determine more accurately what to do with the conclusions your AI gives you.
Embracing AI doesn’t have to be an almighty hill to climb. With so many businesses that have successfully carved out their space in the AI landscape and so many resources to facilitate new players’ ventures in it, SMBs are better prepared than ever to become an AI authority.
Welcome to the VentureBeat community!
DataDecisionMakers is where experts, including the technical people doing data work, can share data-related insights and innovation.
If you want to read about cutting-edge ideas and up-to-date information, best practices, and the future of data and data tech, join us at DataDecisionMakers.
You might even consider contributing an article of your own!