← Back to SCAIO
Procurement Workforce Economy May 2026 · SCAIO Journal

The Quiet Reset in South Carolina's Consulting Bill

AI is starting to absorb the document-heavy work governments pay outside consultants to do — federal advisory contracts already fell 19% in a year. South Carolina is next. But the bill splits across two separate budgets: the state's own agencies, and the counties, cities, and housing authorities of local government.

The short version
What this is about

Governments pay outside firms billions a year to write reports, run analyses, and process paperwork — work billed by the hour. AI can now do a meaningful share of it, and at the federal level that spending has already started to fall.

Why it matters for South Carolina

SC spends an estimated $130–270 million a year on this kind of outside help. That total sits in two separate pots — state-government agencies and local governments (counties, cities, and housing authorities) — each with its own budget, its own procurement, and its own timeline for change.

What we found

A realistic middle estimate is $25–80 million a year in eventual savings across South Carolina, arriving more slowly than at the federal level because of how SC's contracts are written. For the first few years the visible effect is less a smaller budget line than faster turnaround on government work.

The most consequential change to how South Carolina pays for outside knowledge work has already started — quietly enough that most agency budgets haven't named it yet, and mostly in federal data that is a leading indicator for what reaches the state.

Nationally, governments at all levels pay roughly $75–120 billion a year for outsourced, document-based, hours-billed work: consulting reports, policy analysis, monitoring narratives, evaluations, compliance paperwork. The first hard sign of AI displacing it is already in: federal management-advisory contract awards fell 19% in FY2025 to a five-year low, and the GSA placed $65 billion in forward consulting value under review. South Carolina is downstream of that shift — but to see how it lands here, you have to keep three layers apart.

Three layers, three budgets

When people talk about "the government's consulting bill," they usually blur three separate things. Keeping them apart is the whole key to understanding what AI changes in South Carolina, because the money comes from different places and moves on different clocks.

Federal · backdrop
$75–120B
National addressable pool. The leading indicator — where the 19% drop showed up first.
SC State government
$80–150M
State agencies, drawing on the state budget through centralized state procurement.
SC Local government
$50–120M
Counties, cities, and housing authorities — each with its own budget and procurement.

The federal layer is the national backdrop and the early-warning signal. The two South Carolina layers are what matter locally — and they are not the same budget. State agencies answer to the state budget and the state's centralized procurement system. Counties, cities, and local housing authorities each set and spend their own budgets, through their own councils and commissions. A dollar saved at the SC Department of Administration is a state dollar; a dollar saved at the City of Charleston or the Charleston Housing Authority is a local one. They rarely move together, and conflating them is the most common way this story gets told wrong.

The state layer: South Carolina's own agencies

South Carolina's state-government professional-services spending — consulting, advisory work, technical assistance, software-implementation services — is plausibly $80–150 million a year. It is unusually visible because state procurement is centralized: the Materials Management Office sits inside the Department of Administration and reports through the State Fiscal Accountability Authority, so larger contracts surface in the statewide SCEIS system. Four named contracts show the shape of it.

↓ All four below are state-government contracts, paid from the state budget.

3 active
Deloitte · State agencies

Salesforce managed services, security/tech consulting, and the MUSC ERP implementation

Confirmed in SCEIS data. The managed-services and documentation layers are the most directly exposed to AI substitution; the MUSC ERP work substitutes more gradually, mainly in its testing and training-material components. (MUSC is a state institution.)

$7.45M
KPMG · 9-year term, 2023

South Carolina eProcurement Solution

A long-duration platform-and-services contract running to 2032. Its length concentrates the period-of-performance lag that slows realizable savings — the natural re-pricing point is years out.

$115–400/hr
Gartner · SC Dept. of Social Services, 2025

Third-party consulting on a tiered hourly-rate structure

Tiered hourly billing is the pricing model AI productivity compresses most directly: the same scope at fewer billable hours is how a state-agency budget sees savings before any "AI replaced the consultant" headline is true.

disclosed
Guidehouse · SC Office of Regulatory Staff, 2024

Administrative consulting at a state regulatory agency

Research- and document-heavy regulatory work sits squarely inside the kind of task AI accelerates well — summarization, comparative research, first-pass drafting — with the judgment calls left to staff.

The common thread: most state professional-services spend is multi-year, billed by the hour, and document-heavy. Those three traits determine how AI substitution lands in a budget — and they are why the savings, when they come, tend to show up at contract re-compete time rather than mid-contract. Applied across the state layer, the arithmetic produces a defensible range.

South Carolina annual savings · five-year realization

What the substitution math implies, statewide

Bear · 10–15%
$13–40M
Shallow adoption, heavy human review, judgment-heavy task mix.
Mid · 20–30%
$25–80M
Active deployment for first-pass drafting, extraction, summarization.
Bull · 35–50%
$45–130M
On suitable tasks only — writing- and extraction-heavy work, high adoption.

The local layer: counties, cities, and housing authorities

This is the layer most easily mistaken for the state — and it is genuinely separate. South Carolina's local governments — counties (Charleston's total budget runs about $809M, Richland about $1.24B, Greenville large as well), municipalities like the City of Charleston and the City of Columbia, and special-purpose districts — set their own budgets and run their own procurement. Together their professional-services spending plausibly adds another $50–120 million a year statewide, spread across hundreds of small contracts rather than a few named ones. SC counties don't report a single "consulting" line in their financial statements, so this layer is an extrapolation, not a clean count.

Housing is where the state/local line is easiest to trip over, because federal HUD money flows through both levels. The SC State Housing Finance & Development Authority is a state-level body that administers federal housing funds. The public-housing authorities — Charleston Housing Authority, Columbia Housing, Greenville Housing Authority — are local entities, each with its own board and budget. The consulting work AI substitutes most directly (monitoring narratives, compliance documentation, complaint intake) sits mostly at that local-PHA level, where it totals an estimated $3–8 million a year.

Because each authority's budget is small, the dollar savings there are modest — a realistic $1–3 million a year statewide. The more visible effect for residents is speed: faster grant administration and quicker federal reporting, not a smaller line item. That pattern holds across the whole local layer — the work gets faster before the budget gets smaller — and it's why the local savings curve trails the state layer, which in turn trails the federal one by two to four years.

The catch: AI's "jagged frontier"

The most important caveat is not about how much AI can save — it's about how the savings are captured. A 2023 Harvard/Boston Consulting Group study found AI-assisted consultants outperformed peers by 25% on speed and 40% on quality on tasks inside the frontier of what models do well — and performed 19 percentage points worse on tasks outside it. The authors called it the "jagged frontier," and it maps directly onto government work: first-drafting a grant-monitoring narrative is inside it; deciding the corrective action that narrative recommends is not.

Two anchors keep the estimates honest. A 2025 Stanford/Yale study found 17–33% hallucination rates on the legal-grade AI research tools, on factual questions. And a Microsoft RCT of ~7,000 workers found only ~12% document-level acceleration with no detectable firm-level output shift — and only 40% of users using the tool regularly. For regulated government artifacts — a fair-housing determination, an audit memo — the realized savings depend entirely on how disciplined the adoption is.

The disciplined pattern uses AI for the first-pass draft and a human for the judgment call. The undisciplined one uses AI for both — and produces worse work than either alone.

— SCAIO editorial observation

Why South Carolina is well positioned

SC has structural advantages that predate the AI wave. Its procurement is centralized at the state level in a way most states' is not — the Materials Management Office and the State Fiscal Accountability Authority already provide the coordination capacity other states would have to build. It already runs an AI Center of Excellence (first meeting February 2025; 29 agency use cases under review by mid-year) with risk, compliance, and procurement subgroups. And it has a state Director of Artificial Intelligence (Rich Heimann) charged with executing the 2024 SC AI Strategy.

The capacity to translate "the federal consulting bill is falling, and the same is coming here" into specific state procurement guidance therefore exists in South Carolina at a level few states match. The local layer has no equivalent central body — which is exactly why the state layer is likely to move first and more coherently, and why the counties, cities, and authorities will each arrive on their own schedule.

A watch list for the next 24 months

Signals the curve is actually arriving

  • State layer: the AI Center of Excellence publishes procurement guidance for AI-augmented consulting — task-suitability criteria, audit standards, vendor-disclosure expectations.
  • State layer: FY2027–FY2028 re-compete bids on existing state advisory contracts come in at compressed pricing or scope versus the FY2024 baseline.
  • Local layer: SC Housing's next monitoring-contract round shows reduced scope, and the local PHAs report faster administrative turnaround.
  • Both layers: an SC-based firm wins meaningful government work on AI-augmented delivery against a national prime — a sign the shift is creating in-state opportunity, not just relocating it.

Why it matters

Most of South Carolina's AI story moves on visible clocks — bills filed, strategies published, facilities permitted. This one doesn't. It moves in multi-year re-compete cycles, in contract amendments that never make a press release, and in vendor pricing that is rarely disclosed. It is the least dramatic AI story the state will face this decade — and, in dollar terms, plausibly one of the largest, with the most direct line to public budgets at both levels.

South Carolina has the central procurement structure, the Center of Excellence, and the AI Director to manage the shift deliberately at the state level — capturing the savings, redirecting them toward priorities the legislature names, and protecting the in-state professional-services base. Whether it does, and whether its local governments follow on their own timelines, is the policy choice that will define how this lands. SCAIO will return to it as the FY2027–FY2028 cycles develop.

Source & method: This piece adapts the South Carolina lens of a deep research report, The Quiet Reset: AI and the Man-Hour Economy of Government Contracting (May 2026; ~6,500 words, 60+ sources), and adds SC-specific institutional context. The federal figures are well-anchored (GAO, Bloomberg Government, Deltek GovWin, USAspending.gov). The SC state and local figures are extrapolations: the state layer from named contracts confirmed in SCEIS data, the local layer from county and municipal budget shares typical of the professional-services category. SC counties do not report a dedicated consulting line, so the local figure is the least precise. The single largest source of uncertainty in the headline range is the five-year realization rate, which is driven by procurement-cycle structure rather than AI capability. Productivity figures: Dell'Acqua et al. (2023, Harvard/BCG); Noy & Zhang (2023, Science); Brynjolfsson, Li & Raymond (2023, NBER 31161); Dillon et al. (2025, Microsoft 365 Copilot RCT); Magesh et al. (2025, JELS). SCAIO welcomes corrections from agency and local procurement officers.

Companion pieces: South Carolina's AI Workforce Exposure and What We Know — and Don't — About AI's Effect on SC Workers.