Why AECOM bought Consigli - the Autonomous Engineer
Unpacking AECOM's Acquisition and AI Strategy from their latest Earnings Call
Last week reports emerged that Consigli, the Autonomous Engineer, was acquired by AECOM for $390m.
These acquisition prices are almost unheard of in consulting and design (from an incumbent for software) and to understand, I watched AECOM’s Q4 Earnings Webcast where they unveiled their AI strategy. I highly recommend viewing or reading the transcript as it paints a vision for how engineering design will be disrupted by AI.
Here’s what they said (quotes are from the transcript) with my comments at the end on where I think value be captured:
Note: at times I changed the order of the quotes.
Acquisition Background
(CEO of AECOM)
18 months ago, we questioned whether AI was an existential risk for our industry. We were genuinely concerned that…one of our peers or a new entrant…[or] existing software providers…could [use] AI [to] find a way to effectively put us out of business.
To respond, AECOM began experimenting with emerging technology firms. They partnered with several new entrants to understand where the real capability gaps were and soon realised that one firm was significantly ahead.
We found an organization that had built math models, and they were going to profoundly change the industry. And they were going to put us out of business.
AECOM then had to determine how they could move ahead of the disruptor (it’s unclear when they identified the acquisition and the date exactly it occurred). They realised that they have to build models, build tools and “even think about this as building new [AI] teammates.”
Building a Competitive Moat
So over the last 18 months AECOM had quietly began to build a team, investing through their margins. They now believe they have a competitive edge which which is not present elsewhere in the market that allows them to leverage AI in client delivery in a way that’s difficult to copy. It’s predicated on:
An Experienced Team
They have globally experienced engineers, designers and architects who deliver design and are able “to certify those results and…stand in front of regulators.”
Note: In many regions you need a Professional Engineer to legally certify designs.Trusted Client Relationships
A client doesn’t pick you to deliver infrastructure worth billions of dollars unless they trust you and you’ve proven your ability to deliver at this scale.Balance Sheet
Engineering firms have to be able to stand behind their design and procure insurance to support the project.Technology
AECOM is building a team of AI professionals (as shown by the acquisition). That includes “AI PhDs, machine learning PhDs, data scientists.”
Note: Acquiring Consigli appears to be related to strengthening this pillar.
In their words: they are “only one that has that team.”
It means they no longer view AI as an existential risk. Instead they view it as “an existential opportunity…should be considered the disruptor.”
And they validated this opportunity by working closely with, whom I assume is the Consigli team (wasn’t clearly stated). They:
learned…using AI with our teams that we can deliver things at an incredibly fast pace. So in the design process, what might take months, AI can do almost instantaneously…And then for us, it profoundly changed the ways that our team will work.
What AI will allow them to do
It leads to two outcomes:
It allows them to deliver work with less people and do not have to add headcount to grow.
People can learn faster, shortening the time it takes to become an expert designer.
In the past…you spend years learning how to design and doing the same thing over and over…it takes a long time, maybe a decade…now because we’re able to take those professionals and give them those profound experiences in a much shorter period of time, we can accelerate their careers and their career growth…A 10-year season design professional, right? We can create it in a much shorter period of time.
Where They’ve Seen Success with AI
In the words of AECOM they’ve had:
Success building it [AI] out in segments and deploying it on projects and testing it for over a year. Now we have the confidence that we can extend this across our entire portfolio of engineering design
They then asked themselves:
What is the monetization opportunity when you have technology, which will deliver an asset earlier for your clients use of revenue generation, maybe to its constituents at a lower cost basis, delivering a higher ROI.
Where Value will Accrue
The answer for them was it would increase their operating leverage. Historically at AECOM:
For every dollar we generate, it takes somewhere between $0.80 to $0.90 of variable cost to deliver that revenue, resulting in operating leverage of 10% to 20%...generative technology that we have achieved and will be scaling, it will allow us to expand that operating leverage…It will continue to expand in the near future, in the midterm to 30%, 60%, 80%... It is for the entire revenue base
This is where, in my opinion (and feel free to correct me), the math gets shaky. Their assumption about operating leverage depends on maintaining the same market price for work while structurally lowering their cost basis.
And so we look at it this way is we sort of have the opportunity to price to market. And if the margins are better in pricing to market, we decide what we want to do with it. Our objective is to keep it…because we’re going to be delivering designs or delivering our work that’s such a profoundly improved outcome for our customers, I don’t think we need to change market price to actually capture more market share.
But the question is if this assumption will hold true?
Yes AECOM’s margin expansion per dollar of revenue is mathematically correct if costs fall. However, AI reduces the total number of billable hours required. That’s important as contracts often operate with a hourly or cost plus model.
If we go to do environmental work…You have 5 people that work on this project. Imagine if we show up with 1 teammate and 4 AI teammates
This could mean overall revenue declines even if the margin per dollar improves. The way to preserve total revenue (and market price) is to ensure they are paid for the outcome they provide.
And so there’s a market price for what we do. Put aside how it’s actually structured, whether it’s a fee, a fixed fee, a cost plus a unit price, there is a value that the industry pays for getting a design outcome or getting a program management outcome.
Their strategy is to first understand how AI reshapes their delivery model and then have conversations with clients about pricing structures that match this new reality.
we can deploy it on unit cost or fixed price…so we can test it and we can understand what that difference is going to look like. That will then enable us with that information to actually go have a client conversation…at the end of the day, if you can provide something that’s much more compelling in terms of value...you will have an ability in that conversation to shift, right, the pricing model and shift from cost plus to a fixed price where you gain more certainty around that outcome.
They go into further details as to how they believe they can have these conversations and the additional value they provide to customers from value engineering to redesign and reducing materials.
But looking at the market, the reaction to their AI strategy has been sharp. AECOM’s share price dropped from $131 on November 17th to $104 by November 21st.
While I have yet to read an analyst note, I’m curious about AECOM’s ability to maintain their revenue projections in this AI led future.
As they deploy AI and begin capturing more margin, competitive pressure will increase as rival firms will invest heavily to incorporate similar AI capabilities into their delivery models. In the short and possibly the medium term, the market price for design services may stay the same.
But in the long term, as AI assisted workflows become industry standard, cost structures will shift.
For example, a project previously priced at $100k might have cost AECOM $80k to deliver. With AI driven productivity, that cost might fall to $60k. Once firms like WSP or Jacobs reach similar AI enabled efficiency, they will begin to bid at lower prices, such as $80k instead of $100k.
As this competitive dynamic plays out, operating leverage will begin to normalize.
It is still unclear how value will ultimately accrue or what AI led delivery will look like. What is clear is that design delivery models and pricing structures are being fundamentally rethought.
Ultimately though I do not believe the engineering designer will be the primary winner. As AI reduces delivery costs across the industry, the largest beneficiaries will be the real estate developers and infrastructure authorities who can now justify the ROI on projects that previously appeared too expensive.
This shift in the cost structure means the ultimate winners will be us, taxpayers and communities, who benefit from infrastructure delivered faster, at lower cost, and with higher value.
If you have any thoughts about the acquisition, let me know! I’m curious to hear what people think.