Building a Value Equation for AI Agents

Just when I think I’m done with articles about artificial intelligence another one grabs my attention. This time I dove into the concept of AI agents and the role they stand to play in negotiating deals between suppliers and buyers in the retail space. Is the art of negotiation truly that, an art no machine will ever be equipped to perform? Or, is it a series of mathematical equations best handed over to fool proof technology? Of course, it's not quite that simplistic.

AI Negotiation Agents

Automating supplier negotiations is something big companies like Walmart have leaned into in an effort to handle deals with their suppliers more efficiently. For a large company like this, it is virtually impossible for negotiation teams to handle the thousands of contracts they deal with every year. In fact, Walmart has admitted that 80% of their supplier relationships lacked the type of management practices that they know have the potential to save them money or foster regular communication.

For the most part, artificial intelligence agents handle lower level deals and accumulate small savings, making large companies a lot of money as they begin to compound. An AI agent that saves a few hundred dollars several thousands times over can create big savings for a business. And in many cases, this is work that is not being done at all which makes the cost of a monthly subscription alluring.

Can AI Agents Renegotiate Deals to Save Money?

If an AI agent works to save an average of $2,000 on 1,000 deals with suppliers over the course of a year, the company has just saved $2 million. A savings of $2,000 doesn't seem like much to bother with if you are a billion dollar company looking at a single deal where the total value is $50,000 or less. But, it would be hard to find a CEO who would say no to saving millions annually when they look at the bigger picture of accumulated savings across multiple deals.

It sounds like a mathematical no brainer, even for a small company. Give the AI agent a goal of shaving an average of $200 dollars off of 100 deals valuing $5,000 or less and you’ve saved the company $20,000 in a year. A midsized company, averaging 800 suppliers could likely find enough contracts to renegotiate to make it with the addition of AI agents. And again, 80% of contracts with small time suppliers typically only account for 20% of a retailer’s profits. Can humans be reserved the most valuable 20% of negotiations?

Calculating value could be about more than just the money saved on deals. Companies swap employee salaries for monthly subscriptions to artificial intelligence technology. If you pay an employee $100 per hour but only spend $200 a month for an AI agent, it's an instant savings. Or, so it seems.

Can We Truly Swap Out Employees for AI Agents?

AI agents are primarily data driven. One agent can often perform tasks it would take multiple humans to do. The more advanced AI systems become, the easier it is for them to sync with multiple platforms as well as accomplish tasks faster. 

But, using an AI agent still takes work. A person has to set parameters the artificial intelligence will use to make decisions, organize information, and calculate risk and saving potential. Mistakes can still happen if guidelines lack detail. And, even with clear parameters, an AI agent isn’t equipped with technology to finesse its way through conversations which are part of the relationship building process between retailers and their suppliers. 

Finding a price that works for both the supplier and the buyer typically involves a procurement team made up of strategic minds, risk management experts and regulatory or quality control professionals. The team members check in with each other and also have to communicate with a supplier’s team of professionals. An AI agent could do the initial legwork required to map out various deal options. They are great at sifting data. But, should they assume complete decision making control?

I think about tasks performed by a single person, or even a team of people, and wonder if AI could ever take on all of these jobs. Should an AI agent monitor workflow, approve spending for a company, search and compile vendor options, handle misunderstandings or disputes, take preventative steps to ensure there is no conflict of agreement in the first place or monitor how long it takes to secure goods from negotiated deals? 

Some of these tasks require human interaction; especially when a company is making multi-million dollar deals. The overarching debate centers on whether a company can really swap out a human employee for an AI agent or in the end, are they just adding another kind of employee to payroll? Will the addition of subscription costs be offset by new deals struck by AI agents and suppliers? For a bigger company like Walmart, the answer appears to be, yes. But a similar value formula could prove too expensive for small companies.

How Much Should We Depend on AI Agents?

If we can prove that AI agents make a business more profitable then we don’t need to debate whether or not a company should leverage the tool. The better question might be, how much AI should a business rely on? 

There is a growing concern over the subscription model. While it may be affordable now, will increased reliance on AI agents make them just as expensive to “hire” as an expert team of negotiators? If the majority of work is assimilated into artificial intelligence technology, what happens if a system glitches? Will the technology be backed by human knowledge that can step in and handle hiccups manually or have productive conversational interactions? The lack of predictability when it comes to the future of these areas of artificial intelligence has made smaller and midsized companies more hesitant.

As of right now, it is tricky to calculate the value of AI agents. A CEO might be incredibly willing to invest in AI agents, even at a high price point if the agent performed the work of three employees. But, AI agents currently in existence can’t do all of the work one employee would do. They can only do parts of it. Companies still need a blend of analytical minds, ethical communicators and adaptable thinkers who can spot new opportunities on the fly and bring them through negotiation into contract if they want to have a successful procurement team. It’s what makes the value of AI agents difficult to calculate on a broad scale and it's why it is important to analyze on an individual company level.