Top Software Investment Firm Gives Their Take On Agentic AI + Impact on Construction
Vista Equity Partners just dropped their 2026 outlook on Agentic AI and enterprise software. The takeaway for construction: the software you already use is about to shift from a tool you operate to a system that does the work for you. With $37 billion in enterprise AI spend and an estimated $3 trillion in labor value up for grabs, this isn’t a tech-sector sideshow. Here’s what it means for your projects, your stack, and your margins.
There’s a report doing the rounds from Vista Equity Partners – one of the largest enterprise software investors in the world. It’s called “Agentic AI and the Future of Enterprise Software,” and the thesis is aggressive: software is about to stop being something your team uses and start being something that works alongside them. Vista isn’t a startup pitching a seed round. The data they cite comes from the Federal Reserve, Bain, and Morgan Stanley. Hard to wave away.
The question for anyone managing projects, budgets, or teams in construction isn’t whether this shift is happening. It’s whether the software you’re already paying for will capture it – or whether you’ll be left holding a legacy system while competitors pull ahead.
What “Agentic AI” Actually Means (Without the Buzzwords)
From tools you click to systems that act
Vista draws a useful distinction. Most software today is a tool – it helps you do your job. You log in, pull the report, update the record, approve the change order. Agentic AI flips that. The software takes the action itself, inside rules and approval frameworks you’ve set up.
Think about what that looks like on a real project. Your estimating software doesn’t just store bid data – it flags cost anomalies against historical benchmarks and drafts a revised line item for your review. Your scheduling tool doesn’t just show you the critical path – it identifies resource conflicts across three projects and proposes a reallocation before you’ve opened your laptop.
Vista’s paper uses an insurance company called Duck Creek Technologies as its case study. Duck Creek built a system where multiple AI agents handle discrete underwriting tasks – guidelines, risk scoring, policy creation – each one auditable, traceable, and operating inside the insurer’s actual rulebook. The parallel for construction is obvious: imagine the same architecture applied to submittal review, RFI routing, or pay application processing.
Why This Matters More for Construction Than Most Industries
Thin margins, thick workflows, and a workforce cliff
Construction sits in a peculiar sweet spot for this technology. The industry spends just 1-2% of revenue on IT (compared to 3-5% in other sectors), and 70% of contractors have no formal tech roadmap. But the problems agentic AI is designed to solve – repetitive workflows, data-heavy decisions, labor scarcity – are construction’s defining challenges.
Vista notes that while 20-40% of workers across industries are using AI in some capacity, only 1% of enterprise data is actually feeding into those tools. In construction, that gap is almost certainly wider. Most firms are sitting on years of project data, cost history, and procurement patterns locked inside disconnected systems. That data is the raw material agentic systems need. Whoever unlocks it first gets a compounding advantage.
The relevant question is not whether enterprise software persists in an AI world. It is whether software can capture the next wave of AI-driven value creation.Vista Equity Partners, 2026 Outlook
And the labor angle can’t be overstated. The US needs 439,000+ additional construction workers. The UK workforce has hit a 25-year low. If software can take on discrete, repeatable tasks – processing change orders, reconciling invoices against contract terms, generating progress reports from field data – that’s not a nice-to-have. It’s how firms stay operational with the crews they can actually hire.
The Three Things That Determine Who Wins
Context, trust, and scale – and what they mean for your tech stack
Vista’s framework for which software companies will capture this opportunity comes down to three structural attributes. They’re worth thinking about the next time a vendor pitches you an “AI-powered” anything.
- Context. Does the software understand how your business actually operates? Generic AI can generate a nice email. Acting reliably inside your contract terms, approval chains, and cost codes requires years of domain-specific workflow data.
- Trust. Will the system protect your data and produce auditable decisions? Construction companies operate under regulatory scrutiny and litigation risk. An AI agent processing a pay application needs to show its working – and your data can’t leave your ecosystem.
- Scale. Can the software handle enterprise-grade volumes across multiple projects and geographies? Being embedded at the center of operations matters more than being bolted on as a point solution.
The punchline: if your current software vendor has deep workflow context, strong data governance, and is already embedded across your operations, they’re well positioned for this transition. If they don’t have those things, they’re at risk. That evaluation should be happening now, not in two years.
Stay ahead of the curve on construction tech
Get the sharpest takes on AI, labor, and the forces reshaping construction – straight to your inbox every week.
Join 2500+ ReadersThe Business Model Shift You Need to Watch
From paying per seat to paying per outcome
This is where Vista’s report gets genuinely interesting for anyone managing a P&L. Today, you pay for software per seat, per user, per license. Agentic AI changes that. When software executes work (reconciling an invoice, resolving a help desk ticket), vendors start pricing based on work completed rather than headcount.
Vista frames this as the evolution from “software as a product” (licenses) to “software as a service” (seats) to “software as a worker” (outcomes). For construction firms on 2-3% margins, the implication is real: your software costs may shift from a predictable line item to a variable cost that scales with project volume. That could be a good thing – you pay for what gets done, not for seats gathering dust. But it means procurement teams need to rethink how they evaluate and negotiate software contracts.
Vista believes this shift will be worth trillions. Bain & Company projects $5-7 trillion in cumulative value flowing to the software layer by 2030. Whether those numbers land precisely isn’t the point. The direction is clear, and construction firms that understand the shift early will negotiate better deals, pick the right platforms, and avoid getting locked into legacy pricing while their software does increasingly valuable work.
How the Software Model Is Changing
| On-Prem (Legacy) | Cloud/SaaS (Current) | Agentic (Emerging) | |
|---|---|---|---|
| What you’re buying | A product | A service | A worker |
| Pricing unit | Licenses | Seats/users | Outcomes, tasks, usage |
| Growth curve | Episodic upgrades | Linear ramp | Scales with work volume |
| Revenue cap | Number of licenses sold | Customer headcount | Volume of work executed |
| Construction example | Installed estimating software | Cloud project management platform | AI agents processing RFIs, change orders, pay apps |
Frequently Asked Questions
No, and Vista is explicit about this. Agentic AI targets specific, repeatable tasks where execution can be standardized – think document processing, invoice reconciliation, scheduling optimization. It’s not replacing your superintendent or PM. It’s handling the admin that eats up 30-40% of their week. Given the industry’s severe labor shortage, that’s not displacement – it’s survival. (Source)
Enterprise AI adoption is tracking faster than the cloud transition did, but it’s still measured in years, not months. Vista compares ServiceNow’s AI tier adoption to the Office 365 upgrade and public cloud migration – both took close to a decade for broad penetration. Construction will likely lag the enterprise average given the industry’s cautious adoption culture. But the early movers are already here. (Source)
Three things, based on Vista’s framework. First: what’s your AI roadmap beyond a chatbot bolted onto the existing product? Second: how are you handling data governance – does my data stay in my ecosystem? Third: are you building toward outcome-based pricing, and what does that look like for my contract? If your vendor can’t answer these clearly, that’s a signal. (Source)
Vista isn’t an AI startup selling a product. They’re one of the largest enterprise software investors globally, owning dozens of software companies. When they publish an outlook like this, they’re telling you where they’re putting capital and how they expect their portfolio to evolve. That’s a useful signal for anyone buying enterprise software. The sources cited include the Federal Reserve, Bain, Morgan Stanley, and Menlo Ventures. (Source)
Vista Equity Partners – 2026 Outlook: Agentic AI and the Future of Enterprise Software
Federal Reserve – Measuring AI Uptake in the Workplace (Feb 2026)
Bain & Company – AI Market Study (Q1 2024)
Menlo Ventures – The State of Generative AI in the Enterprise (Dec 2025)
Morgan Stanley Research – Software Gut Check: AI IS Software (Feb 2026)
