pull down to refresh

TL:DR
By tweaking the scheduling outputs from legacy software, the company cut 28 days off a $149 million highway project’s timeline.
When contractors bid on a project, they’re often on a short, often unforgiving timer. Zachry Construction, headquartered in San Antonio, feels this crunch on infrastructure projects, where the estimation process can lead to a race against the clock.
Ranjeet Gadhoke, the builder’s vice president of project controls, said that on a $200 million infrastructure project for a public entity, for example, he could expect three weeks to turn a submittable bid around.
Within those three weeks, two are occupied by the task of building a schedule, which leaves one week for any questions or considerations in the estimate — for example, what if the construction team had double the number of workers or resources? Each of those estimates takes Gadhoke about a day or two to work through with his schedule.
“You have to put your best price here, put your best foot forward,” Gadhoke said. “That’s not enough time.”

Speeding things up

Zachry was looking to move the estimating process along in a quicker fashion. The company turned to Menlo Park, California-based Alice Technologies, which develops scheduling software for the construction industry. Zachry uses the company’s Alice Core product, which builders can use to generate scenarios on a pre-existing schedule and determine financial and time-based outcomes, per a news release.
The tech allows for builders to upload previous schedules that they’ve created in other software, such as Microsoft Project or Oracle P6, and adjust them. Within Alice, questions that would’ve taken days to answer — for example, how would a project change if crew size or resources shifted — now take minutes, Gadhoke said.
Zachry and Alice declined to share the cost of the service, which is customized for each project or client.
One project that Gadhoke and Alice’s founder, René Morkos, highlighted was a $149 million highway project in Texas.The construction team had some questions — how could they move construction crews around to be the most efficient, and how would that reduce the number of crews necessary on the jobsite?
Within three minutes, the team ran about 60 scenarios. With that information, the contractor lowered the number of flat work crews from nine to seven. Between resequencing the project and moving people around, Zachry saved 28 total days of work.
While the Texas highway project has since been impacted by delays, Gadhoke said that Alice was able to help Zachry figure out what the subsequent best course of action was, given those setbacks.

Artificial — and human — intelligence

Alice uses generative AI in its software, which contractors are racing to adopt in order to stay ahead of the curve and remain competitive.
Regardless of the power of the AI technology within the Alice Core system, it’s reliant on the user to fully function. In order to generate simulations for its highway job, Zachry still had to input a pre-made schedule, which Gadhoke cautioned can make or break the outcome.
“If you put garbage into Alice, it’s going to give you garbage. It’s really not the software’s issue there,” Gadhoke said.
He said that making sure that everyone was on the same page with what he called the master data was imperative to getting good use out of the tech.
“If you want to see true value, make sure you have a schedule process that you know represents reality,” Gadhoke said.

My Thoughts 💭

A three week window to respond to a bid solicitation that is about $200M seems way too quick to me. Plus once again we have no clue how much this construction company paid to have the AI software developed and how long it took before it way ready. Then we are told it saved 28 days of labor. Also cut the crew from 9 to 7 people. These production gains are modest at this time. It’s not disruptive. Trimming a few workers here and being a bit more efficient with material allocation is definitely goals to be celebrated but if this company is paying for an expensive yearly license fee and other costs to develop and maintain this AI system I just don’t see it affecting the bottom line much. With more bidding time these marginal gains could have been captured by a motivated human.
I think this is just the beginning...from here the actual development of the software will begin, which can bring further benefits... In addition, if this software "ensures" the success of the various contracts for the company (gets the given project or not), then I think the AI ​​software is worth every penny
reply