Hyperautomation by 2028: What Gartner's Forecast Changes for SMBs

The Number You Can't Ignore

Seventy percent. That is the share of global organizations that, according to Gartner, will have implemented some form of hyperautomation by 2028. This is not a vague projection about "the digital future" — it is a three-year deadline with a quantifiable scope. And for any Brazilian SMB that still treats automation as an optional investment, this number needs to urgently change the internal conversation.

Hyperautomation is not a single, isolated technology. It is an approach that deliberately combines at least four layers: artificial intelligence (language models, machine learning, computer vision), RPA (Robotic Process Automation, such as UiPath, Automation Anywhere, or the Brazilian-made Neobrain), system integration via APIs and iPaaS (such as MuleSoft or Make), and process management with BPM tools and process mining solutions like Celonis or Bizagi. Taken separately, each of these technologies already delivers point-in-time gains. United within a coordinated architecture, they create what Gartner calls an "automation fabric" — where distinct systems communicate, decisions are made automatically, and exceptions are escalated to humans only when necessary.

What sets hyperautomation apart from simple automation is precisely this orchestration. An RPA bot that fills in spreadsheets is automation. A workflow in which that same bot extracts data from electronic invoices, reconciles it with the ERP via API, triggers an AI model to approve or reject the transaction, and notifies the finance team only in out-of-pattern cases — that is hyperautomation.

Why 2028 Matters Now, Not in 2027

The classic trap with market projections is using them to justify inertia: "we still have time." The problem is that hyperautomation is not installed in a matter of weeks. It requires a foundation — and that foundation takes 12 to 24 months to build solidly.

This foundation rests on three non-negotiable pillars:

1. Clean and Centralized Data

Autonomous AI agents — the top of the hyperautomation pyramid — only work with reliable data. An SMB with customer records fragmented across three different systems, or with financial processes that depend on manual adjustments in spreadsheets, is not ready to automate decisions. It is ready to automate errors at scale.

2. Documented and Stabilized Processes

Before automating, it is essential to understand what actually happens in the process — not what the manual says should happen. Process mining tools like Celonis do exactly this: they map the real flow from system logs. The results are often surprising: between 30% and 50% of real process executions deviate from the designed flow.

3. Basic Automations Already Running

Gartner's recommendation is straightforward: start with the fundamentals before investing in autonomous agents. Automated invoicing, bank reconciliation, vendor onboarding, periodic report generation — these are high-volume, low-variability processes with a fast ROI. They are also the laboratory where teams learn to operate, monitor, and correct automations. Those who skip this step and go straight to generative AI on a critical process tend to backtrack at a high cost and with significant internal distrust.

What a Brazilian SMB Can Do in Practice

I work with SMBs in Brazil, Italy, and the United States, and the pattern is consistent: the companies that get ahead are not the ones that invest the most — they are the ones that start earlier and in the right place.

A concrete example: a client in the distribution sector, with annual revenue of R$ 40 million, had an order reconciliation process that kept two full-time employees occupied. We implemented a workflow with RPA (UiPath) integrated with the ERP via REST API, with a simple machine learning classification layer to identify discrepancies. Within 90 days, the process ran in under 2 hours per day, with human intervention required in only 8% of cases. Project cost: R$ 85,000. Return in the first year: equivalent to 1.4 full-time positions, plus a 94% reduction in reconciliation errors.

This is not a case of full hyperautomation — it is the first step. But it is precisely the step that needs to be taken now so that, in 2026 or 2027, the company is positioned to scale with AI agents without building on sand.

The Risk of Not Being Among the 70%

Gartner's forecast does not say that 30% of companies will fail. But it implies, unmistakably, that those who do not adopt hyperautomation will be competing against cost structures and operational speeds that are fundamentally different from those of companies that did. In markets with tight margins — retail, logistics, financial services, agribusiness — this efficiency gap tends to be decisive.

For Brazilian SMBs, the argument is twofold: scaling operations without proportionally increasing headcount is a necessity in markets where personnel costs are high and the availability of skilled labor is limited. Hyperautomation is not about replacing people — it is about not needing to hire ten people to do the work that a well-automated process handles with two.

The clock is ticking. And the three years until 2028 are shorter than they seem when you account for the time it takes to mature a well-built automation architecture.