#148 Why AI Could Expand the ESCO Market, Not Just Speed It Up
More work than
we can do.
Brodie Boland spent fifteen years at McKinsey helping building owners decarbonize. Now he runs Upgraid, which sizes up a building from space before an ESCO walks in the door — and he thinks the industry is fighting over a pie that's about to get a lot bigger.
#148Why AI Could Expand the ESCO Market, Not Just Speed It Up
Toronto, a PhD, then Washington, DC, where he rose to partner and led work on sustainability and climate risk for real estate and infrastructure investors.
He left halfway through his degree to lead AIESEC, ending as President of AIESEC International in 2005–06. "To my parents' great delight," he finished.
What an ESCO can get on one owner's buildings before the first meeting. "Extremely rich," Brodie says. "Maybe even too rich."
Projected to reach about $15B by 2027, per Lawrence Berkeley National Lab. Brodie's argument: the ceiling is much higher than that.
From the mountains to the boardroom to the boiler room.
Brodie's route to the ESCO world ran through Afghanistan, Sri Lanka, Kyiv, and fifteen years of consulting. Click any stop on the timeline.
Calgary Growing up in the mountains +
3 yrs Leaving school to lead AIESEC +
Colombo A feast in a tin-roof house, and then the tsunami +
Dec 2004 Kyiv, orange flags, and "we go to the streets" +
McKinsey Toronto, a PhD, and "over my skis" +
~10 yrs Real assets, decarbonization, boardroom to boiler room +
Today Co-founder and CEO of Upgraid +
There is going to be more work than we can do. So the question is: how quickly can we do it?Brodie Boland · on why the ESCO bottleneck is speed, not demand
"Having a zero-sum mentality is probably not in alignment with the facts."
Deferred maintenance is piling up across the MUSH sector. Energy constraints are raising the value of efficiency. Most of the big forces, Brodie says, are tailwinds. Once the market is growing, the bottleneck shifts from finding work to how fast you can move it, and that's where AI earns its place.
Four conversations about selling smarter.
About the building, the skeptic, the sales team, and the market itself. Tap any tab to switch.
Know the building before you walk in the door.
Upgraid starts with almost any building in North America and pipes in everything it can: multiple satellites, permitting databases, subsurface soil conductivity, local weather, solar radiance. From that it predicts what's inside: what kind of HVAC and how old, the thermal properties of the envelope, the EUI, the efficiency opportunities, and the payback on each.
The second layer is context: who owns the building, how they're financed, which incentives they qualify for, and what they care about, both as an organization and as individuals. The building energy models are Upgraid's own, built for this one job. That, Brodie says, is why they're more accurate than the large models.
What the platform infers
- HVAC type and age, envelope U-values, EUI
- Efficiency measures and their paybacks
- Owner, financial structure, and available incentives
- Rooftop units, spotted from satellite imagery
What it replaces
- Walking into the first meeting with a list of questions
- Waiting for site access before anyone knows anything
- Hand-keying utility bills into energy models
- Building proposal decks from scratch
A tale of two worlds.
Jim asked the obvious question: the ESCO community can be skeptical of new technology, so how is it really landing? Brodie answered directly. World one is a handful of the largest ESCOs who have used Upgraid for a while. They say it is transforming the way they sell and the economics of their business, and they use it more every week.
World two is everyone he's cold-calling, and they're skeptical. The early customers took a while to realize they could change their whole approach: deals moving faster, better conversion, more qualified leads, different conversations. The company's live challenge is getting that across to "someone who doesn't know me from Adam."
What long-time users report
- Deals moving along "way faster"
- Conversion rates improving
- Far more qualified leads
- Conversations that are totally different
Why newcomers hesitate
- "We've always done it this way"
- The value takes time to show up in the pipeline
- A cold call can't show months of results
- An industry that's wary of new tech
Sophisticated sales in an information-poor environment.
Jim's take as a former sales leader: data like this lets you ask your team why they're spending time on a prospect that can't buy, or whose systems don't fit. It's a more calculated way to sell. Brodie added a use he didn't expect: training. ESCO sales are highly sophisticated, but they happen with very little information, "and those two things don't really go together very well."
New reps who work through a portfolio report, sometimes 150 pages of it, figure out what veterans already know. They need to think about the financial structure, the debt position, the whole portfolio, and which buildings need which upgrades and what they'd pay back. That raises the level of the whole sales team.
Where it helps the team
- Qualifying out prospects that can't or won't buy
- Spending time on the right part of the Venn diagram
- Teaching new hires to think like owners' CFOs
- Raising the floor on sales effectiveness
Where it needs a caveat
- Heavy process loads, such as a car plant, throw off the EUI
- Financial data is thinner in the private sector
- 150 pages still needs filtering
- The model sees the building, not the factory inside it
Stop fighting over market share.
Putting his McKinsey hat back on, Brodie's first point is that the ESCO industry has massive room for top-line growth, not just a fight over share between firms. The deferred-maintenance backlog is growing, and energy constraints are raising the value of efficiency. "I would humbly suggest" that a zero-sum mindset doesn't fit the facts.
In a growing market the bottleneck becomes speed, and that changes the economics. A huge share of ESCO costs land before a contract is signed: pre-feasibility work, sales reps, and sales engineers. That hits margins and caps growth. He expects pre-sale costs to compress, sales cycles to shorten, and some segments to turn into a volume play.
What he expects to change
- Pre-sale costs compress
- Sales timelines shorten
- More efficient meetings between buyers and sellers
- A volume play in some segments
The tailwinds behind it
- A growing deferred-maintenance backlog in the MUSH market
- Energy constraints raising the value of efficiency
- Most macro forces pointing the same direction
- AI tools, "if they are fit for use"
We are literally addressing climate change and healthy schools at the same time.Brodie Boland · on what he loves about the ESCO industry
Six questions worth expanding.
The threads Jim and Brodie went deepest on, pulled from across the conversation. Click any question to expand.
01What did three years abroad teach him? +
That the world is much smaller than it looks. People live in fundamentally different ways and are still looking for the same things. When the tsunami hit communities he had visited in Sri Lanka, and when a protest he had talked about in a Kyiv apartment became the Orange Revolution, those events stopped being headlines.
It's a cliché, but we're far more connected than we realize.
02What made a McKinsey candidate memorable? +
McKinsey's interview process is deliberately structured: the same talent bar worldwide, and the same answer when a client asks for a favor for their son — apply on the web, same process, no shortcuts. Within that, two things stood out to Brodie. First, a candidate who was simply interesting, who had done something other than the consulting club and straight A's.
Second, and more important: the candidate who forgot about getting hired because the case got interesting. Three tourist railways in Alaska, one losing money — and someone who starts nerding out on it. That spark matters at 12:30 a.m. in a hotel far from home, with the team still preparing for tomorrow's meeting.
Someone who got distracted from their objective of being hired because they got really interested in the problem. You want to hire the latter.
03What does "boardroom to boiler room" mean? +
It was his McKinsey team's name for working at both ends of a decarbonization program. In the boardroom: capital allocation, portfolio strategy, investment returns, the alpha being created. In the boiler room: what actually changes in the building, how it makes the facility manager's life easier, and how the building runs better and healthier.
Jim's point for ESCO listeners: the people doing the installs and retrofits are the other half of that same picture. The strategy only matters if someone implements it, "down to the wiring."
04How do you model a building you've never been inside? +
Stack enough outside data and the inside becomes predictable. Satellites show the roof and the rooftop units. Permitting records show what was installed and when. Soil conductivity, local weather, and solar radiance shape the energy picture. Upgraid's models turn that into high-confidence predictions about systems, envelope, and energy use.
Once an ESCO has a utility bill or site photos, they upload them and every energy model updates automatically, with nothing to re-key. The output is proposal-ready documents for meetings and presentations.
05Where do the models fall short? +
Jim asked whether it works better in public- or private-sector buildings. Brodie hasn't heard a meaningful difference. Customers use it for both, though financial data is thinner on private owners. The real exception is buildings with heavy process loads.
When customers run major industrial portfolios through it, Brodie is careful to say the model covers the building, not what's made inside it. If a car plant is running inside, measures like EUI will be off.
06How does Upgraid keep AI from going off the rails? +
Mostly by not leaning on large language models. Upgraid uses them, but sparingly. The core work is done by purpose-built models: computer vision that reads a satellite image of a roof and counts rooftop units, and other machine-learning approaches that work out a building's energy dynamics.
Brodie lists three payoffs: more accurate estimates, fewer hallucinations because the models are built for the job, and much less compute. An LLM doing this work is "using the wrong tools for the job." Limiting their use also limits the strange failures they're known for.
Large language models do great things in other areas, but they're not fit for purpose for this job.
The four questions every BEP guest gets.
Jim closes every episode the same way. Tap any card to see how Brodie answered.
Concepts worth knowing before you listen.
The episode moves quickly through some specialized vocabulary. If any of these terms are new, tap for a 90-second primer. No industry background needed.
Brodie Boland
Brodie is co-founder and CEO of Upgraid, a building-intelligence platform based in Somerville, Massachusetts. Upgraid uses satellite and public data with purpose-built building energy models to help ESCOs find, qualify, and propose upgrade projects before they have site access.
Before Upgraid, Brodie spent about fifteen years at McKinsey & Company, starting in Toronto and becoming a partner in Washington, DC. There he helped lead the firm's work with real estate and infrastructure investors on decarbonization, climate risk, and the energy transition. He holds a PhD in organizational behavior from Case Western Reserve University.
He grew up in Calgary and, partway through university, spent three years with the global student organization AIESEC, finishing as President of AIESEC International in 2005–06.
Co-Founder & CEO · Upgraid (pronounced "upgrade"; the AI is the pun), Somerville, MA.
What an ESCO actually is.
An energy services company designs, builds, and often finances improvements to a building, such as new HVAC, lighting, controls, or envelope work. It gets paid, in whole or in part, from the energy those improvements save. It's a contractor, an engineer, and a financial partner rolled into one.
Under an energy savings performance contract, the ESCO puts a savings number in writing. If the building doesn't save what was promised, the ESCO covers the difference. That guarantee lets a school district or city pay for the work out of savings that haven't happened yet.
Brodie's McKinsey clients were "the folks that would be kind of ESCOs' customers." Now his customers are the ESCOs.
The MUSH market, by the numbers.
MUSH stands for municipalities, universities, schools, and hospitals, the public and institutional buildings that make up the core of the ESCO business. Their capital budgets are thin, their buildings are old, and they can sign long-term contracts, which suits performance contracting.
Lawrence Berkeley National Laboratory's latest industry report puts U.S. ESCO revenue at about $10.7 billion in 2024, with roughly 78% from public and institutional customers. ESCOs expect about $15 billion by 2027. LBNL has also estimated the remaining untapped market in the tens to hundreds of billions of dollars.
"Even if just in the MUSH sector, the deferred maintenance backlog is growing." That's the core of Brodie's growth, not zero-sum, argument.
Why pre-sale costs matter so much.
Before an ESCO signs a performance contract, it spends real money: sales reps and sales engineers chasing the opportunity, site visits, utility-bill analysis, and often a full feasibility study or investment-grade audit. All of that is at risk. If the customer walks away, the ESCO eats it.
That risk shapes the industry. It squeezes margins, stretches sales cycles to months or years, and pushes ESCOs toward large projects that justify the upfront cost. Make the early work cheaper and faster, and smaller projects start to pencil. That's the "volume play" Brodie describes.
"There's a huge share of costs that happen in advance of sale in the ESCO market." Brodie expects AI to compress them and shorten sales timelines.
EUI and building energy models.
Energy use intensity, or EUI, is a building's annual energy use divided by its floor area, usually in thousands of Btu per square foot. It's the miles-per-gallon number of buildings, a quick way to compare a school to a peer school or a building to its own past.
A building energy model simulates how a building uses energy from its geometry, envelope, systems, occupancy, and weather. Upgraid's approach is to estimate those inputs from outside data, such as satellite imagery and permits, rather than a site audit, then refine the model once real utility bills arrive.
The models are built for buildings, not the processes inside them. In a heavy industrial plant, Brodie warns, "things like EUIs will be off."
