Shackleton AI Raises $5M Seed for Construction Agents

At 2 a.m. on an Orlando construction job, an inverter burned out, leaving questions about repair, replacement and who would absorb the cost. In Shackleton AI's account of its contractor origins, reconstructing the night shift the following morning required calls across people and systems, even though the information already existed. The difficulty was getting a usable account of the job to the person who had to answer for it.

Shackleton AI is now selling the software that grew out of that experience to other builders and owners. On November 4, 2026, the construction technology company announced a $5M Seed round led by Battery Ventures, with the capital intended to expand AI-agent workflows and accelerate go-to-market. The investment puts fresh capital behind the work of connecting a site event to the schedule, cost record and approval that determine what a project team can do about it.

The funding follows operating experience inside Ernest

Shackleton's founding team includes CEO Michael Ragheb, CPO Hassan Yasin and CFO Olexandr Kaltsev. The Florida corporate record confirms those officer titles, while the company describes its beginnings inside Ernest, a commercial general contractor it says the team grew to $35M in revenue across the Southeastern United States. That history gives the software a practical starting point: the job that produced the data was also the job the founders were responsible for delivering.

The Seed syndicate includes Brick & Mortar Ventures, Trust Ventures, Correlation Ventures, Motivate Partners, Thayer Investment Partners, Act One Ventures, Brainstorm Ventures and Silence VC. Shackleton identifies Battery Ventures as the lead and says other investors also participated. The announcement does not specify a valuation or a complete prior-financing history, and the dollar denomination is not explicitly identified in its published text.

The company's contractor history describes work ranging from warehouse lighting and restaurant construction to luxury retail and healthcare renovations. Its account emphasizes the repeated process beneath those different assignments: estimating, budgeting, bidding, scheduling, procurement and turnover. That breadth helps explain the choice to connect several stages of a project inside the software. A field report can carry information needed by people further along the same job, long after the person who recorded it has moved to another task.

What Shackleton's AI agents can do with a shared project record

Shackleton describes a common project model containing bids, budgets, requests for information, submittals, change orders and field reports. Its AI agents draft RFIs, follow up on overdue submittals, identify budget variances before payment requests reach approval and turn field notes into daily reports. Review and approval remain with people, an important boundary around the company's longer-term ambition for projects that manage more of their own administrative work.

The connection between those records is the product's operating argument. An unanswered question matters differently when it is attached to the trade waiting on an answer, the schedule that trade affects and the cost associated with delay. Giving an agent those relationships can make its output more useful to a project manager. That is an interpretation of Shackleton's design, rather than a claim that its software has independently demonstrated every such outcome.

The company also maintains a bridge into accounting. Its integration directory lists QuickBooks Online, QuickBooks Desktop, Sage 300 Construction and Real Estate, and Sage Intacct, with different connectors mapping records such as projects, subcontractors, cost codes, budgets, invoices and bills. DevCuration has also covered construction operating software at Miter and construction accounting AI at Adaptive. This is a consequential accommodation to the buyer. A builder can consider replacing the coordination work around a job while retaining an established financial ledger, leaving the implementation team to make sure those two accounts of the project stay aligned.

An owner's view across locations gives the rollout another audience

Shackleton says its platform manages more than $1B in construction project value across more than 15 states. It names Red Lobster, Dog Haus and Pepper Lunch among the multi-location brands it serves, alongside midsize general contractors working in multifamily, industrial and commercial construction. The project-value figure describes the work managed through the software; it gives no measure of Shackleton's own sales or enterprise value.

For an owner commissioning repeated locations, the attraction extends beyond the convenience of a single contractor. The owner needs a coherent picture of progress across builds, while each project still has its own people, questions and financial decisions. Software that keeps those records connected could make the relationship with the builder easier to manage. It also brings another stakeholder into a deployment that has to work for both the field team recording events and the owner relying on the resulting account.

The company's reported buildout results reach up to 35% ahead of schedule and as much as 18% under budget for multi-location customers. Those are company-reported maxima, with no published sample or independent analysis establishing how much of the outcome came from AI. They are useful context for the sales proposition and require care when applied to a new customer's projects. For a brand planning several openings, each site still has a distinct sequence of work and a budget attached to it. As that portfolio grows, the information reaching the owner's desk has to remain specific enough to direct attention to the location that needs it.