The mid-market wasn’t forgotten. It was just never served correctly, writes Omar Akilah (pictured, below), SVP Product & Industry Strategy at Infios.
There’s a lot of talk in our industry about the mid-market being forgotten when it comes to enterprise technology and AI, but I don’t buy that argument. I don’t think these businesses were forgotten. I think they were never really served the right way in the first place. Frankly, I have a problem with how we define the mid-market to begin with.
We tend to put businesses into categories based on revenue. Maybe it’s $250 million, maybe it’s $1 billion. But at the end of the day, there are businesses that serve different markets, different industries, and different channels. You’ve got wholesalers, retailers, consumer goods companies, manufacturers and third-party logistics providers. They all have different needs. And those needs change as their businesses grow. As these businesses grow and shift their strategies, their technology ends up bottlenecking them.
Think about a small brand starting out in someone’s garage. Then it goes viral. Now it wants to add wholesale. It wants to add retail. Maybe it’s a consumer goods company that realizes it’s giving away too much margin and wants to go direct to consumer. the business starts shifting its strategy, and suddenly the technology foundation it built on can no longer support it.
The ERP that worked when the business started only solves so much. Maybe the company outsources warehousing to a 3PL, but that 3PL has its own software and its own way of doing things. Eventually, the company decides it wants more control. It brings warehousing in-house. Now it needs people to pick, pack, ship and receive. It needs software to manage labour and automation.
That’s how these mid-market businesses grow. They solve what they immediately need, but as they grow, the gaps start showing up. Then they get bigger, they want more, they want to grow, and suddenly what they have doesn’t work anymore. Now they’re scrambling.
Big companies typically have more established supply chains. They’ve figured out whether they’re going to fulfil on their own or use a 3PL. They’ve established their transportation networks, their channels and their operations. They may pivot along the way, but they’re not changing how they do business as frequently as these growing mid-market businesses. The smaller enterprise guys? The reality is they’re volatile. They shift.
Look at transport. The big guy is saying, hey, I need to hit my minimum commitments with UPS before I start rate shopping. The smaller guy is saying, I just don’t want to lose my shirt. Tell me who can get my product from point A to point B at the lowest cost with the highest reliability.
Both need transport technology. But they’re solving very different problems. And as that smaller company grows, even its transportation strategy changes. Now it has enough volume to negotiate its own carrier relationships. The solution that worked yesterday may not be the one it needs tomorrow.
Breadth, but not all at once
In my opinion, the most underserved market when it comes to comprehensive supply chain execution suites is the mid-market. They need the breadth, but at different points in time. Not all at once. When I think about what good looks like, I think about modules I can plug in and plug out as I need them. Think of it like LEGO bricks. I build something, and then as my business grows, I add more bricks and build something else. I don’t have to start over every time my business changes. If I already have an ERP and I want to improve how I promise orders across different sales channels, why should I need an entirely new order management solution? Why can’t I just plug in an order-promising component?
If I’m happy with my 3PL but not how they’re cartonising my shipments, I should be able to say, hey, I’m good with you, but I need you to use my own cartonisation algorithm. implementing a WMS can cost hundreds of thousands or even millions of dollars. And that’s before you get into everything that needs to connect.
Think about all the integration points, whether they’re carriers, 3PLs, marketplaces or wholesale channels. I’ve got to integrate, integrate, integrate, map, integrate. It’s money and time. And then there’s the cost to own it. I can’t staff it. Now I need a planning team. Now I need this team, that team. And supply chain is already such a big drain on margin. I can’t have the technology killing my margin, too.
These businesses need to be able to take what they need, when they need it, and connect it to what they already have. They can’t take on a multimillion-dollar implementation every time their business changes.
Making the connections work
A big part of what makes supply chain technology so expensive is integration. You’ve got to map data, build integrations, configure systems, and test everything to make sure it works together. Now think about what’s possible with AI.
Give me an integration document. I should be able to use AI to understand the requirements, map the data, and start testing an integration in a matter of days or hours instead of weeks, months, or years. It’s not just integration. Think about onboarding, configuration, and ongoing management. All the things that make technology expensive to implement and own in the first place.
Let’s go back to that consumer goods company. It’s growing. It’s selling through more channels. Maybe it’s brought warehousing in-house. Now it’s got orders coming in from everywhere, inventory sitting in different places, and multiple carriers moving its products. All those systems are doing their jobs. But who’s connecting the dots?
What happens when a big wholesale order comes in and suddenly there’s not enough inventory to fulfil the orders you’ve already promised? Now somebody has to figure out what inventory is actually available, which orders are affected, and what needs to happen next. Do I move inventory from another location? Do I change which orders get fulfilled first? What does that mean for the warehouse and transport?
That’s where I think AI can make a difference. It can recognize what’s happening across those systems, identify the problem, and recommend what to do about it. Then, within the rules the business has established, it can start taking action. These companies don’t necessarily need to replace everything they already have. They need to get more out of what they’ve got.
Now people are starting to ask, why do I need to buy a solution for every problem? If I understand my business requirements, why can’t I use AI to build something that solves what I need right now? Maybe you can and it works for the next six months or two years. But what happens when your business changes again? Now you’ve got another solution to manage, another integration to maintain and another thing that might not work when you need it to.
That’s the problem. Mid-size businesses don’t need more technology to manage. They need technology that can grow and change with them. At the end of the day, mid-size businesses aren’t trying to buy technology. They’re trying to grow. And the technology should help them do that, not get in their way.
Eliminate the noise
So where do you start? Understand your main pain points. Where are you spending the most time on menial tasks? Menial tasks, not the advanced stuff. We all want to start with the hard stuff. How many calls is your transportation team getting every day from somebody asking, where’s my order? When is it going to arrive? Where’s the driver? Maybe it’s 30 calls a day, and every one of them requires somebody to find information, call somebody else, and report back. That’s a very simple AI agent use case.
Chart out those processes. Stack rank the ones that create the most noise, and start getting rid of them. How do I eliminate the noise? There’s so much noise, and I’m buried in it. I can’t see around me. So I apply AI to get rid of those repetitive tasks.
And as AI takes on those repetitive tasks, I start to trust it more. Maybe initially it books my transportation appointments. Once it proves it can do that, I start asking it to recommend who I should book those appointments with based on reliability. It’s not just graduated autonomy. It’s graduated trust.
AI isn’t perfect. It’s going to make mistakes. Give it context. Show it the 10 steps you take to complete a process. Make sure it understands those steps and can follow the rules you’ve established. Then start automating and give it more to do as your confidence grows. That’s where I would start.