We use cookies to provide you with a better experience. By continuing to browse the site you are agreeing to our use of cookies in accordance with our Cookie Policy.
Everyone knows at this point that artificial intelligence (AI) will change the world. That includes wholesale distribution. As a chief technology officer for a company offering AI automation for distributors, I’ve had many conversations with C-level executives and IT directors. Many feel the need to create an AI strategy but are a bit overwhelmed because there’s so much noise to sift through.
Vendors and “experts” understand they’ll sell better if there’s more buzz and if things seem complicated. What’s more, it’s common for people to think the way AI can help in business is too good to be true. It’s hard to get them to fully understand until they try it.
As an example of what AI can do, our company, Canals is AI for wholesale distributors. It currently has 2 products live in the market:
Order entry automation: Integrate with your sales reps’ email, identify orders & quotes, understand the content, and create the order automatically for the rep to review. Anything Canals isn’t sure about it asks the sales rep to fill in and learns from their actions for next time. It cuts down order entry time by >80% and learns your business and customers’ preferences.
Invoice entry automation: Integrate with your AP email, identify invoices, understand their content, match them up against your POs and confirm they’re valid and correct invoices, and automatically enter them into your ERP.
The big question is what can AI do for your company, and how do you jump in?
This article is a simple and straightforward primer on how to look for, evaluate and implement AI solutions. No buzz, no fluff.
Where you’ll likely be in two years
Nearly everyone has used ChatGPT at this point. We know what these large language models can do and can imagine the types of solutions that are likely coming. Several of the items below already exist or are being worked on:
1. Order entry automation. Monitors your incoming emails and voicemails, understands them, identifies the purchase orders (POs) and enters them into your enterprise resource program (ERP) automatically.
2. Accounts payable automation. Monitors your incoming invoices, understands them and enters them into your ERP automatically.
3. Customer service automation. Chatbots answer your customers’ questions, check on their orders, etc.
4. Knowledge base management. A chatbot ingests all your company’s documents and allows any employee to ask questions.
5. Automated regulation monitoring. Summarizes regulatory changes relevant to you.
6. Automated contract review. Flags any irregularities or sensitive clauses.
7. Much more. You can imagine possible products as well as me.
Distributors using these tools will be cheaper, faster and more accurate. Imagine a customer sends a 1,000-line PO to three distributors. Two of them reply with a quote the next day, while one responds in 5 minutes. Who’ll get the sale?
Employee retention is another often overlooked benefit of AI. While today people are scared of it, soon they’ll get used to it and no one will want to do the tedious data entry tasks that AI can do for them. It’ll be harder and harder to hire people for these roles.
Lastly, managing a large organization is hard. With AI, distributors will be able to do more with the same team. For example, one customer did an analysis after using Canals for a year and estimated that they could do 20% more sales with the same team. I think that’ll be typical for the most effective AI tools.
So that’s where the world is going. There’s a bit of urgency, but the good news is vendors will come to you. So as long as you can evaluate them effectively, you’ll be fine.
How to evaluate AI solutions
Evaluating AI solutions can be hard because so much noise and obfuscation is floating around. Every company now claims to have AI. How do you sift through it all?
Actually, you already know how to evaluate software; you’ve been doing it for years. Whether under the hood it uses AI is, in reality, irrelevant to you as long as it provides value.
That said, here’s my checklist for evaluating AI solutions. If you follow these steps, you can make good decisions.
1. Don’t trust demos
It’s very easy for a vendor to curate cases that its tool handles well or even to manually massage the results. Demos are nice, but you’ll want to test the tool live.
For example, you’ll want to watch a sales order entry tool process some real POs from your customers. That includes different formats and difficult cases: pdfs, Excel docs, upside-down images, handwritten notes, typos and industry jargon. The vendor should be happy to let you do that on a live call. If they need to process your data and get back to you later with the results, that’s a red flag.
2. Check references
A vendor’s website is a good place to start, looking at the strength of its testimonials. However, you can also ask vendors to introduce you to customers similar to you. You can learn a lot based on whether the references are only reservedly positive — or raving. You can ask about reliability, speed of iteration, challenges encountered and other things that are hard to discover before buying and investing a lot of time and money.
3. Consider integration costs
Most AI tools require an integration; you’ll want to understand how much time and effort that requires. Except for especially complex situations, top-tier tech companies should aim to have you live within two weeks after you sign up, with minimal time spent by your team. The customer references are great for verifying this.
4. Confirm customizability
Our experience is that more often than not, customers have some business process that’s different from other distributors and which we don’t yet support. What’s more, they often don’t discover the gap until they’re already using Canals.
You’ll want to confirm whether the vendor can customize the tool if needed. Again, the customer references are great for verifying this.
Implementation tips
Changing any business process can be a significant undertaking. Here are my top tips based on our experiences so far:
1. Start small
Starting with high-impact areas is obvious. Less obvious is prioritizing use cases that are low investment. If you can have results within one month that are noticeable to everyone in the organization, that’s a huge win over a six-month project. As mentioned previously, in most cases, top-tier tech companies should aim to integrate and launch you within two weeks.
2. Start with tools your users will love
Change is hard. People have their inertia. Even with Canals, only a subset of sales reps and finance people are enthusiastic while many resist a new tool. However, some get it and through them, it spreads. Within a few weeks, people say they wouldn’t want to work without it.
If you can give your organization that type of experience, it’ll be that much easier to introduce other AI solutions in the future.
3. Don’t emphasize the AI
People are scared of AI. You’ll want to communicate that you don’t plan to replace anyone or take away their commissions. It is only a tool so they can earn more while spending their time on higher-value work.
Where to go from here
Everyone will get there eventually because competition will demand it. For forward-thinking distributors, now is a good time to get a head start and build an advantage until everyone else catches up.
Industry peers are a great source — you can ask what tools they’ve heard of or found especially valuable. As long as you follow the evaluation steps above, you’ll be in a great position to identify the tools that will be most impactful for your company.
Erez Arnon is co-founder and CTO of Canals, a company specializing in AI-driven automation solutions for wholesale distributors. Before that, he co-founded Office Hours, a knowledge-sharing marketplace. He was a software engineer at Adap.tv before it was acquired by AOL.