From Hype to Impact: How AI Agents are powering the Autonomous Enterprise

Building the Autonomous Enterprise

The autonomous enterprise is a business that can operate, adapt, and optimize core workflows with minimal human intervention. The enabler is Agentic Process Automation (APA), AI agents that plan, decide, and execute across multi‑step, cross‑functional workflows, not just single tasks.

In this maiden issue of this newsletter, we unpack highlights from Forrester’s Opportunity Snapshot—Building the Autonomous Enterprise: How Organizations Are Preparing for and Adopting AI Agents, commissioned by Automation Anywhere, and dig into a video discussion with Mihir Shukla (CEO, Automation Anywhere), guest speaker Craig Le Clair (VP & Principal Analyst, Forrester), and our host Gabriel Carrejo (Head of Customer Advocacy). Together, they explore what’s hype, what’s real, and how enterprises are moving from pilots to production safely, quickly, and with measurable outcomes.

We’ve broken down the discussion into bite-sized segments below, but the full recording is available here on demand, and you can download your complimentary copy of the Forrester Opportunity Snapshot here.

What the research shows: adoption, trust, and outsized returns

Forrester’s global study (500 respondents; broad industry and geography; significant C‑suite involvement) finds enterprises are optimistic, but pragmatic, about AI agents. While many favor augmenting a step in a workflow today, there’s growing confidence in more autonomous operations, with leading companies already seeing huge impact on their way to becoming an autonomous enterprise.

Key signals from the conversation:

Trust and control still shape adoption: many leaders prefer model‑assisted steps inside a deterministic automation wrapper before handing over end‑to‑end control.

Outsized returns accrue to two profiles:

a.     companies that treat AI as core to the business model

b.     leaders who reimagine end‑to‑end departmental workflows (e.g., customer service, finance, HR, supply chain)

Example: Sumitomo Rubber expanded from shipping‑calc automation to real‑time inventory + shipping orchestration, improving revenue, cost, and inventory metrics simultaneously.

Where organizations are getting started (and what’s changing)

We’re witnessing a shift from RPA’s “clever efficiency” to APA’s process reinvention. Early traction appears in:

• Customer support: higher‑context agents drive measurable ROI via deflection and resolution
• Employee augmentation: agents support research, operational decisions, and sales workflows
• Creativity and adaptability: AI helps teams spot and test winning pivots faster

One bank reduced a six‑hour EV financing process to 15-45 minutes with an autonomous agent, yielding $150M in revenue — this is not just front office, back office or customer experience it’s all of it. As one customer put it, “RPA saves time; APA redefines it.”

Why some companies see disproportionate returns from AI

Not all AI adoption strategies are created equal. Mihir explains why the biggest winners aren’t dabbling in pilots, they’re either:

• Embedding AI as a core business pillar, or
• Reimagining entire workflows in key functions like finance, HR, and supply chain

These leaders aren’t just automating tasks, they’re redesigning how work gets done. The result? Outsized returns and competitive advantage.

From “Agentish” to Agentic

Most AI “agents” in the market today are what Forrester calls “agentish,” useful, but still limited in autonomy and action. This reflects the broader industry reality, not the capabilities of every solution.

To help leaders understand the maturity curve, Forrester’s framework maps the journey across two axis:

• Autonomy: from prescriptive rules to model‑driven control
• Action: from zero action to long‑running, multi‑tool execution  The path moves from solver agents (single step) → worker agents (multi‑step, tool‑triggering) → executive agents (system‑of‑agents that plan, create tools, and coordinate to a goal)

How enterprises are deploying agentic systems

The Forrester framework shows the maturity curve but how are organizations actually moving along it? Leaders are taking a measured, trust‑building approach:

• Today: most start by letting models handle specific workflow steps (solver), wrapped by deterministic automation to build trust.
• Next: worker agents trigger existing tools (RPA/API) but generally avoid inventing new patterns.
• Future: executive agents create new tools and coordinate sub‑agents to deliver end‑to‑end outcomes, requiring interoperability standards and stronger guardrails.

While most enterprises are still early in this progression, technology is advancing faster than expected.

On stage at Imagine, Mihir demonstrated our Process Reasoning Engine (PRE) that lets leaders communicate a complex process goal and have the system construct and execute the workflow on the fly — akin to a “Star Trek computer” for enterprise operations. Technology is arriving faster than many expected; the pace of adoption will be determined by imagination, trust, and change readiness.

What counts as an AI agent (and why it matters for cost)

Not all AI agents are created equal.

• Personal productivity agents (email summarizers, etc.) delight users but rarely satisfy a CFO.
• Application‑specific (“captive”) agents optimize within a silo (ERP, CRM) but don’t transform how work flows across the enterprise.
• Cross‑enterprise agents orchestrate end‑to‑end processes spanning multiple systems —where the biggest ROI lives. 

When you reimagine flows across silos, you often reduce seat licenses and shift work to autonomous execution. That can translate to significant IT budget reductions, because fewer people need to “touch” every step in a process.

Trust, ROI, and moving from POC to production

The top barriers are non‑technology as much as technical:

• Skills & domain expertise, and the balance of human oversight vs. AI autonomy, all roll up to trust.
• Explainability is the No. 1 trust driver: how was the decision made at runtime?
• Enterprises also need stronger observability and guardrails to scale beyond early adopters.

A pragmatic path forward:

1. AI‑first operating model analysis to show the from→to value (e.g., $200M–$1B+).
2. Pick high‑ROI hotspots first — create quick wins that fund broader transformation.
3. Enterprise‑grade security & governance from Day 1.
4. Pilot‑to‑Production programs where providers co‑invest and help teams operationalize at speed.

Summary

From Forrester’s vantage point:

• Deterministic automation remains the safe wrapper for inserting AI models into mission‑critical flows.
• 74% of leaders expect meaningful opportunities for fully autonomous agents in the next two years — momentum is real.
• For long‑running, end‑to‑end processes, intelligent automation platforms are the preferred route.
• Meanwhile, a majority of IT leaders doubt that personal productivity co‑pilots will reach economic viability on their own.

From the field, Mihir’s view is consistent: lead with process outcomes, then orchestrate RPA, APIs, and AI agents to deliver them. Real examples:

• Banking (loan origination): 6 hours → 15–45 minutes.
• Healthcare (prior authorization): days → <2 hours.
• Manufacturing (order fulfillment): faster clears, lower inventory.

AI transformation is not just an IT task. It requires IT + business aligned to three CEO outcomes: grow revenue, reduce cost, manage risk. Some customers are already targeting $1B+ impact.

Closing thoughts

Craig: Empower your people to accelerate adoption

Successful programs recruit early adopters across the business, put tools in their hands, and let them surface transformative use cases from the ground up. It’s practical change management and skills development rolled into one and it keeps the focus on outcomes, not tools.

Mihir: The AI‑first enterprise mindset

“For the first time in human history, you have an unlimited workforce. If you have an unlimited workforce, what is it you cannot do?” - Mihir Shukla

Leaders who shed “limited workforce” assumptions will reset cycle times, OKR expectations, and operating models. Challenge your teams: What could we turn from red to green in two days? When you reframe constraints, many more things become possible.

Watch the webinar in full + download the Forrester study

Get the complete discussion and the full research to go deeper into:

• How AI reshapes roles
• How to build a competitive edge
• The AI skills gap
• Metrics & ROI

From Point Apps to Agentic AI: Building the Autonomous Enterprise

Ready to move from hype to impact? Subscribe to this newsletter for exclusive insights on AI agents, enterprise automation, and real-world success stories from leaders creating impact with AI right now.