Site Selection Services: What They Include and How AI Is Changing the Category
Traditional site selection is a weeks-long manual process delivered by generalist consultants. AI-native firms are compressing it to days and screening markets that consultants would not have the bandwidth to touch.
Site selection is one of the highest-stakes decisions in real estate development and industrial investment. Get it right, and the project benefits from the tailwind for its entire operational life. Get it wrong, and no amount of execution quality covers for a site with inadequate power, constrained labor access, or entitlement exposure that was visible in the data from the start.
For decades, site selection has been the domain of a small set of specialist consulting firms and the advisory arms of large commercial real estate brokerages. That model is being restructured by AI, and the gap in what AI-native delivery can screen, analyze, and shortlist versus what a traditional team can produce is widening quickly.
What Site Selection Services Traditionally Cover
A traditional site selection engagement typically runs 8-16 weeks for a complex industrial or digital infrastructure project and covers:
Market screening. Identifying which geographies meet the client's macro criteria: available labor pool, utility reliability, access to transportation networks, regulatory environment, and economic development incentive landscape.
Parcel identification. Finding specific sites within target markets that meet physical and zoning requirements. This involves broker outreach, GIS analysis, and local market knowledge.
Criteria scoring. Evaluating candidate sites against a weighted scorecard covering power availability, land cost, environmental risk, proximity to inputs or customers, and other project-specific factors.
Utility pre-application. Engaging with utilities to confirm available power capacity, estimate upgrade costs, and understand interconnection timelines.
Incentive analysis. Quantifying available economic development incentives -- tax abatements, grants, workforce development programs, infrastructure commitments -- from state and local governments.
Shortlisting and recommendation. Producing a ranked shortlist of sites with a recommendation rationale for the client's decision team.
Specialist firms include McCallum Sweeney Consulting, Civitas Advisors, and KPMG's site selection practice. The large brokerages (CBRE, JLL, Cushman) have site selection groups that operate within their industrial and logistics advisory practices.
The Core Limitation of the Traditional Model
Traditional site selection is linear. A consulting team takes one client engagement at a time, applies human bandwidth to each research step, and produces a recommendation over the course of weeks or months. The pace is determined by what two or three consultants can research and analyze, not by the speed of the available data.
This creates three structural constraints:
Market coverage is limited by bandwidth. A traditional engagement typically screens 8-15 candidate markets. Markets that seem less obvious are not screened because there is not budget or time. Counterintuitively, this means traditional site selection optimizes for well-known markets rather than for the best available site.
Criteria scoring is not reproducible. Consultant scoring models are judgment-based. Two consultants applying the same scorecard to the same site will often score it differently. Clients cannot audit the methodology or run sensitivity analysis on the weights.
Data is point-in-time and aging. Research conducted at the start of an engagement reflects conditions at that moment. Power availability, zoning status, and available parcels change. A recommendation delivered 12 weeks after the research was conducted may reflect a market that has moved.
What AI-Native Site Selection Delivers
AI does not replace the site selection judgment. It eliminates the research bottleneck and expands the coverage frontier.
Multi-market parallel screening. AI-powered site selection can screen 50 or more target markets simultaneously, applying consistent quantitative criteria across all of them. Markets that would never have made a traditional consultant's short list because they were not "obvious" get evaluated on the same basis as established markets. In practice, this often surfaces tier-2 and tier-3 markets with compelling criteria combinations that are being overlooked because conventional wisdom has not caught up to the data.
Real-time utility data. Power availability analysis used to require utility pre-application calls, which take 2-6 weeks to get a response. AI systems can infer preliminary power availability from publicly available data sources -- FERC interconnection filings, utility IRP documents, EIA capacity data, and substation proximity analysis -- to pre-screen sites before any utility contact is made. This filters out power-constrained sites before significant time is invested in them.
Zoning overlay analysis. Zoning research on 50 candidate sites would take a traditional team months. AI can layer zoning data, flood plain overlays, environmental risk databases, and historical permitting records across all candidate sites and produce an initial risk ranking in hours.
Continuous monitoring. Once a client has a target market list, AI can monitor it on an ongoing basis -- tracking new parcel listings, rezoning actions, utility capacity announcements, and competitive development activity -- and alert the team when conditions change. Traditional site selection is a one-time engagement. AI-native delivery turns it into a standing intelligence feed.
Reproducible scoring. The criteria model is explicit, auditable, and adjustable. When a client changes a weight (prioritizing water access over labor cost, for example), the entire candidate set is re-ranked automatically.
Where Human Judgment Remains Essential
AI-native site selection does not eliminate the consultant. It changes what the consultant does.
Incentive negotiation. Economic development incentives are relationship-dependent. A consulting firm that has placed 15 projects in a state has a fundamentally different negotiating position than a first-time applicant. AI can quantify the available programs; it cannot build the relationship with the economic development authority or structure a negotiation strategy.
Community engagement. Opposition to industrial and data center development is increasingly common. Managing the community engagement process -- understanding local political dynamics, presenting to planning boards, engaging with neighbor groups -- requires human presence and judgment.
Final site determination. The decision between two shortlisted sites with similar scores almost always comes down to qualitative factors that are hard to quantify: the quality of the utility relationship, the tractability of the permitting authority, the developer's read on the local business climate. AI provides the analytical foundation; the human makes the call.
Thin data markets. In emerging markets or unusual property types where the data layer is sparse, AI advantage narrows. The systems perform best where there is data to work with.
What to Evaluate Before Hiring a Site Selection Partner
For institutional developers and industrial investors comparing traditional and AI-native site selection options, the relevant questions are:
Coverage. How many markets will be screened? What is the methodology for expanding coverage if the initial set does not yield a compelling shortlist?
Data recency. How current is the power availability data? What is the lag between real-world conditions and what the system reflects?
Criteria transparency. Can you see the scoring model? Can you adjust the weights? Can you audit a site's score against the underlying data?
Timeline to shortlist. What is the realistic timeline from engagement start to a ranked shortlist of 3-5 sites?
Asset class experience. Site selection criteria differ materially between data centers, logistics, manufacturing, and life sciences. Has the team executed in your specific asset class?
Incentive track record. What incentive programs has the team successfully negotiated? In which states?
The AI-native firms compete most strongly on coverage, data recency, and timeline. Traditional consultants compete on incentive relationships, community engagement experience, and institutional trust. The strongest engagements pair AI-native screening with relationship-led execution at the shortlist stage.