How Lighting Agents Paved the Way for AI

Technology is changing many areas of the lighting industry – but providing wide access to information that previously was only provided by agents may be the biggest disruptor yet.

By Benjamin Rapkin

One of the first things I was told back in 2018, when I started signing lighting agencies for Lightglass, was to never, ever disclose pricing to anyone except your agents. At the same time, I was told to never schedule a meeting with a potential customer, always let the agent do it. I thought these agency rules were not only ridiculous, but could kill our company. Nonetheless, I played by the rules, for the most part.

I remember testing these boundaries in the early days of Lightglass in 2019. I had sales targets to hit, after all. We were based in Philadelphia and had a simulated window product perfect for healthcare, so I was excited to meet two designers from EwingCole in my IES Fundamentals of Lighting class. I asked for their help setting up a demonstration for EwingCole’s lighting team, and it worked. A huge opportunity for Lightglass that could launch us into our next phase of growth.

About a week later I got a call from our local rep, fuming. They scolded me for 10 minutes about how I crossed the line, made them look bad, and hurt our relationship. At the end of the call, they told me that they reached out to EwingCole and cancelled the meeting, and would try to reschedule.

They cancelled the demonstration. Let that sink in. A rep cancelled a meeting that would have made them money, to protect something they valued more than that money: control of the relationship.

I started to ask myself if I agreed with their point of view. What would cause someone to cancel a meeting that could make them a bunch of money? What happened to this person that caused them to become so defensive, to the point of harming one of the small independent manufacturers they represent? I realized that this is the behavior of a rep trying to defend their position in the market, but without a unique value proposition. In other words, they didn’t have any logical reason for a specifier like EwingCole to choose them over a competitive agency, so they were afraid. They want the specifiers to think of them when they need lighting help, not the manufacturer. They want to own the relationship, and keep everyone else at arm’s length.

That call was the moment I understood the business model I had actually signed up for. In the years since, I have watched the same instinct show up in seven or eight different forms, and every single one of them is an information problem. Here is a list of ways lighting agents have eroded their business model and paved the way for AI.

1. Intentionally concealing information from customers and partners

Lighting designers are always explaining to me: “If we just had good pricing information earlier in the project, it would solve so many problems!” Well, why don’t they?

Manufacturers usually give price lists with Distributor Net pricing to their agencies upon signing the contract. The pricing for most products is either already known or can be quickly calculated with predetermined numbers. So why can’t specifiers get it?

Ask the reps. They will tell you that if they reveal the true prices of things, they won’t be able to survive as a business. This is true! But that doesn’t make it any more ethical. Why will their business die if they don’t hide pricing? Because the commissions they make from manufacturers are too low to create a sustainable business.

This is where the conversation usually stops, but I have a follow-up question for reps. Why are you signing contracts at commission rates where you lose money? If an agency needs 17% to survive, why are you signing contracts at 10% or 5%?

When pricing is concealed, middlemen get rich, and everyone else gets hurt by a lower quality built environment.

2. Massive hidden markups

The way reps make their money from concealing pricing is through a financial mechanism called overage. Instead of quoting their customer at the price provided by the manufacturer (which already includes their commission), they provide a heavily marked up number instead. In their agreements with manufacturers there is a clause regarding how much of this markup agencies get to keep, usually between 85% and 100%. The largest overage markup I ever personally witnessed was 200%.

Some markets are known as “package” markets. This means that the reps provide one price for the entire project’s lighting schedule, instead of itemized pricing for each product. Package markets are great for reps because they can hide markups, and nobody knows which products are adding the extra expense to the project. That opacity is what makes the rest of it possible: coordinated positioning with the electrical contractor and distributor, budget cuts, value engineering, and more opportunities to swap out high-quality products for cheaper alternatives and pocket the difference.

3. No market feedback or growth KPIs for manufacturers

From the manufacturer’s perspective, it is very frustrating working with reps, because you put someone else between you and your customer and receive very little market feedback in return. I was proactive about asking for information on project opportunities, target design firms, and at-risk projects, and received information less than 40% of the time. We even created a process in the company to systematically gather information from agencies on key projects, and it barely returned anything useful.

I realized later that agents would conceal project information from us so that we couldn’t give them trouble for value engineering us out of a project. If we didn’t know about a project, we couldn’t influence it, and we couldn’t complain if we lost it.

As a response to this, we stopped providing price lists to agents and required them to request pricing from us every single time. This at least gave us a hint at how the market was interacting with our products. To me, coming from a software background with too much market feedback, the agency model felt like we put the fate of our business in someone else’s hands, and we were lucky if they provided any information at all before the purchase order arrived. Or didn’t.

4. Spec registration and territorial disputes

When a manufacturer signs with an agency, one of the first processes that gets triggered is checking agency territory maps. The manufacturer has to make sure that the new agency territory doesn’t encroach on the territory of an existing agency partner. The territories seem pretty arbitrary, like some strange gerrymandering process occurred amongst a council of agency principals. Some territories are the entire state, some are two states, some are half a state, some are a third of a state, some are just a handful of metropolitan areas. The territories expand and contract and change with every merger and acquisition. It is a pain to keep track.

Unfortunately, it gets more complicated. Agency contracts also have stipulations for commission splits within territories. There are three typical designations for a commission split. First is the Spec Rep, the agency that got the product specified on the project. Second is the Order Origination Rep, the agency that processed the Purchase Order from the distributor or customer on behalf of the manufacturer. Third is the Territory Rep, the agency in the location where the project is being built. Usually the lion’s share goes to the Spec Rep, and the remainder is split between the Order and Territory reps.

Since it is rare for agents to register specs or volunteer any type of project information with manufacturers, there are often disputes between reps over who earned the commission, and no paper trails or evidence to support either side. I can’t tell you how many awkward phone calls I have made to architects to ask who showed them the product or who helped them write the spec. Sometimes I was the one who helped the customer write the spec, but the agents need their piece of flesh regardless. This is a lose-lose situation for everyone. Manufacturers have to make decisions about which rep to upset, and sometimes the manufacturer just pays everyone to avoid any ill will.

From the specifier perspective, I always get frustrated that designers use billable hours trying to figure out which product is represented by which agency in which territory. Then a few weeks later the manufacturer changes agents, the spec rep got poached by a competitor, and the work needs to be redone. Designers, how much time do you spend in a year figuring out which reps should be contacted for your projects? Or do you just call the ones you like, regardless of the manufacturers they represent? That’s what I would do, because in a service business, time is money.

5. Offering lighting design services for “free”

Most reps compete with lighting designers directly by offering lighting design services for “free.” They poach lighting designers from top firms, pay them a higher salary or commissions, and offer “free” lighting design for architects and owners. It obviously is not really free, because the agent will make their money back with overage if they win the project.

Lighting designers are very vocal about how undervalued and underappreciated the art of lighting design is. Well, when someone is willing to do your job for free, it is going to decrease your value. And this “free” service isn’t independent lighting design. It is highly biased by the rep’s line card and target profit margins. The incentive is not to choose the right solution for the project, the incentive is to maximize profits and control over what gets ordered. This is yet another behavior that enriches the middlemen and hurts everyone else with a lower quality built environment.

This is also another reason why agents claim they “can’t survive on commission alone.” Part of their overhead is lighting design services. So, to sum this up, they are poaching lighting designers, undercutting lighting design firms to win projects, hiding markups in pricing, and then complaining that they can’t survive on commissions alone. It seems like it would help solve these problems if they stopped competing in lighting design services and let the lighting design firms handle that. It would lower their overhead by at least one salary per year.

6. Line cards too big to handle, with conflicts of interest

The other day I was talking to a spec rep in DC about a manufacturer, and he cut me off to ask, “Is this one of my lines?” I was unfortunately unsurprised to hear this and said, “Well, I don’t know, did you sign a contract with them?”

If you are a manufacturer reading this, and you aren’t seeing the agency activity you hoped for, they may have forgotten they represent you. I am serious. Some agencies represent over 150 unique manufacturers, each with their own product catalogs. Agents aren’t stupid or lazy, they simply put themselves in a situation they are not equipped to handle. Spec reps, who manufacturers count on as a front line in each territory to generate specs, usually just pick a handful of favorites from their line card and focus on them, leaving the vast majority of products gathering dust in the sample storage room. Agents represent so many manufacturers that there aren’t enough business days in the year to host each one in the territory to build relationships and generate sales. Agents must choose which manufacturers get fed and which don’t on each project.

If the agents don’t have time to work with you, have so many products that you get lost in the noise, and don’t let you book your own meetings or do your own sales, how are you expected to grow?

7. Artificial demand for “me too” products

How did we end up with “equivalents” to every single product? Don’t rush to blaming China. First, look closer to home at the reps.

In order for them to compete on a project, they need to substitute products represented by their competitors with matching products of their own. What happens if a rep wants to win the project, but doesn’t have a product to propose as an equivalent? If the project is important enough, they will go to the manufacturers on their line card and ask them to make an equivalent for a cheaper price. This creates a race to the bottom, where any innovative manufacturer is all but guaranteed to have a knock-off within a few months or years. Knowing this is how the reps operate, instead of innovating, manufacturers just watch each other to see what is getting specified and try to make cheaper versions. Believe it or not, this happened to us at Lightglass, a product that is difficult to copy well and has a niche market. We were not spared.

To be clear, I think it is totally normal and fair for manufacturers to compete on price with similar products. The difference is the disconnect between the manufacturers and the end decision-makers. Architects and real estate developers may not always understand the important differences between the originally specified product and the “equivalent.” The manufacturer is separated from their customer by a middleman, and the middleman doesn’t have the time to be an expert in the products, so the customer is not educated properly and won’t be able to discern between the basis of design and the equivalent.

If you go to LEDucation every year and have trouble finding anything innovative or new, this is a big part of the reason why.

8. AI crossing and substitution tools

Reps and distributors are already adopting AI tools to help recommend crosses and substitutions from their line cards automatically. Tools like ParSpec, and Elari are being built specifically for rep agencies and distributors, and they are already being used to recommend substitutions across an entire fixture schedule in seconds. These rep tools have pricing built in, so the rep gets a complete picture of how a substitution will impact their margin, alongside the differences between the specified product and the proposed alternative.

Meanwhile, specifiers and manufacturers have barely started adopting AI in a meaningful way. Reps have AI propose equivalents for entire fixture schedules in seconds, send them to the specifier to approve, and then the specifier spends hours manually redoing work they already did, trying to figure out if the proposed substitutions are acceptable. Reps spend seconds, specifiers spend hours, on two halves of the same workflow, except the specifier has access to less information.

Think about what that does to behavior. These tools encourage reps to propose more substitutions on more projects, because it is fast and cheap to do so. Specifiers are going to have trouble keeping up, and most likely wish their original design decisions could be protected and preserved.

Before I go further, here is what I am not saying

I am not saying every rep behaves this way, and I am not saying the function reps perform has no value. I have worked with agents who protected our spec when it would have been easier and more profitable to swap us out, who taught me things about my own market I did not know, and who earned every point of commission. Local presence, quotation support, and relationships built over decades are real.

What I am saying is that the model itself rewards the opposite behavior. When the structure pays more for concealing information than for sharing it, you don’t need bad people to get bad outcomes. You just need normal people responding rationally to their incentives. The good reps I know succeed in spite of the model, not because of it, and they are the ones most likely to thrive when this changes.

So why do I say lighting agents paved the way for AI?

If you look at the problems I described above, you will see that they are all connected. Each of them is an information or communication problem.

Concealing pricing is concealing information. Withholding project info from manufacturers is concealing information. Hiding overage in a lighting package is concealing information. Carving out agency territory maps is about organizing, storing, and distributing information. Spec registration is about sharing information at the right time. Line cards being too big is about storing, organizing, and distributing information. Crossing and substituting products is about retrieving and comparing information. Offering lighting design services for free is about controlling the project and the flow of project information.

The industry has spent decades building an entire business model on friction in the movement of information. And we have all just been given the greatest information technology of our generation, if not human history.

AI is finally capable enough to address these information problems:

• No line card is too big for AI. It will learn and remember every detail of every manufacturer and product it has access to.

• AI can calculate pricing for highly complex products faster and with fewer errors than human quoters, and it has no reason to withhold that information from you.

• AI can generate configurators and chat assistants for every product, and then provide market feedback directly to the manufacturer in real time.

•AI can learn territory maps in seconds and simply tell you who represents the product you are interested in, although there may not be much need for territory maps in the coming years.

• AI can find suitable equivalents and compare products faster and more comprehensively than any human. That is exactly why reps are already investing in AI crossing tools.

• AI can facilitate communication directly between manufacturers and everyone who needs information from them, 24/7.

• AI is the right technology to solve our industry’s problems. But it is not the only way, and it never was.

Tesla saw the same problem in a different industry

Before AI became mainstream, there were plenty of examples of other industries successfully solving similar problems. One of the most well-known is Tesla, followed by Rivian, Lucid, and other electric vehicle companies.

Cars are much more technologically complicated than light fixtures and come with much higher risk. Tesla looked at the traditional car dealership model and knew that if they launched through car reps, they would never get off the ground. Why? Just like the lighting industry, the incentives were mismatched. Dealerships make a large share of their profit from service, and gas cars have far more parts that break and fail. Electric vehicles have fewer parts and a much lower risk of needing maintenance. Selling EVs meant cannibalizing the service business.

Lighting agencies make their money from substituting products for cheaper equivalents and hiding markups in the price. So any manufacturer who makes something better, instead of cheaper, is in the same boat Tesla was in.

So what did Tesla do? They built a direct-to-customer sales model. They opened retail locations across the country, hired full-time staff, and interacted directly with customers. Similar to lighting reps, car dealerships don’t really care which car you buy as long as you buy it from them. Tesla needed to educate the market on why electric vehicles were worth the extra money, and they knew a dealer would never truly be an expert in every car on the lot. Full-time Tesla staff have to become experts in their products, or they don’t sell and they don’t eat.

Going direct also meant Tesla could control the pricing of their vehicles. Similar to lighting reps, dealerships add arbitrary markups based on whatever factors they deem important, which means the customer never really understands the real value of the car they just bought. By going direct, Tesla was able to build trust that the price they stated is the price the customer pays. No surprises.

Tesla foresaw the information obstacles in launching through traditional dealerships. So not only did they innovate the world’s best electric vehicle at the time, they innovated how cars are bought and sold, because they had no other path to success.

The takeaway

The answer to these information problems, in seemingly any industry, is to go direct to your customer. Close the gap between you and the decision-maker. Don’t let middlemen bottleneck your information or your growth. And now it is easier than ever since you can use AI to build a digital bridge between you and your customers. Your competition already is.

Manufacturers: If you disappeared from your agents’ line cards tomorrow, how many of your specifiers would notice?

Designers: How many hours did you lose last year trying to get information that somebody else already had?

About the author

Ben Rapkin is co-founder and CIO of chatLED, an AI-powered platform that brings lighting manufacturers and specifiers closer together. His career spans lighting manufacturing and software: he co-founded the lighting companies Lightglass and Lightly after exiting the NYC tech scene in 2017. He can be reached by emailing benr@chatled.com