New G2 Research: How AI Agents Are Overcoming Market Hype to Deliver Real Business Impact

We already know that artificial intelligence (AI) is everywhere.

It feels like there's a new AI software feature or bot announced every other week. And with that comes an increase in reports, data, and surveys about all the hottest AI tools on the market.

For instance, a new studyfrom MIT claimed that 95% of AI projects never make their way into production. However, that report only examined custom generative AI programs, which are inherently limited in their scope and potential deployment.

My latest study at G2 dives deep into AI agents, which are the clear next step in AI evolution. Perhaps the role of large language models (LLM) is not limited to serving as chat-driven suggestion machines, but rather as underlying intelligence for agentic planning and decision-making.

A Leap of Trust: AI Agents Are Winning Hearts and Walletsdetails how agents are moving from pilot to production faster than anyone predicted - and they're proving value with cost savings, speed, and real business impact.

After polling over 1,000 B2B software buyers and analyzing thousands of G2 reviews, we found that nearly 60% of organizations already have AI agents in production, and over half of them are highly likely to expand the scope or budgets for them over the next 12 months.

The promise of AI is real: 4 trends from our new agent report

Below is a preview of some of the trends and findings from our 2025 AI Agents Insights Report, which provide a glimpse into the buyer's journey - from finding agent solutions to scaling them up within their organizations.

1. Agent adoption is happening at scale

We're officially past the "fear of missing out" (FOMO) era for AI. Companies are making deliberate, ROI-focused investments that start with solving a specific business pain point, not chasing hype. They work back from the problem to the agents, which show near-immediate results. This shift in strategy has led to lightning-fast, scaled adoption.

The most popular uses aren't experimental; they're straight to the bottom line: software development and customer service, where agents are easing major bottlenecks. AI is no longer a side project. It's an embedded operating technology owned and governed by IT.

• 57% of companies have agents in production today, demonstrating a surprisingly short cycle between testing and scaling.
• 40% of companies have an AI agent budget of over $1 million this year, and 1 in 4 large enterprises are planning to spend $5 million or more over the next 12 months
• Leading use cases for AI agents at enterprise companies include customer service, business intelligence, and software development.

2. Agents are driving speed and growth

Unlike generative AI (genAI) chatbots, which primarily drive individual productivity, agents are facilitating bottom-line growth at scale. But why do humans leak productivity, and agents don't? Unlike human processes, which often fall victim to Parkinson's Law (where work expands to fill the time allotted), agents can work without deadlines or distractions. As they become faster, deliverables roll out more quickly, giving organizations a new source of operating leverage and boosting confidence in their outputs.

This report found that this speed is a key outcome for agent adoption, directly translating into top-line growth.

• 34% of study participants have enough confidence in AI agents to follow a "let it rip" approach and make any reactive corrections or QA sprints as needed.
• 83% of buyers are satisfied with agent performance, indicating low churn and high confidence in the technology.
• Respondents reported a median 23% gain in speed-to-market, and G2 Review Data shows velocity gains of up to 50% in marketing and sales use cases.

3. Agents are winning trust in the boardroom -- and on the front lines

Agents are not just driving business outcomes - they are also improving employee experience. Early investments in AI agents are quickly paying off, with the median time to achieve a meaningful outcome being six months or less, and over a quarter of companies seeing results in three months or less. But agents are not just driving business outcomes. Employees are happier in their roles because they now have to do less repetitive work.

As organizations experience the benefits of agents and trust is earned at all levels of the organization, I predict leaders will increasingly give them more autonomy.

• Nearly 90% of study participants reported higher employee satisfaction in departments where agents were deployed.
• Nearly 50% of buyers would grant full autonomy to agents in low-risk workflows.
• 45% of leaders predict a net increase in jobs by 2028 due to talent redeployment to higher-value work and resulting growth in their businesses.
• Agent programs with a human in the loop were twice as likely to deliver cost savings of 75% or more than fully autonomous agent strategies.

4. SaaS faces a mandate to evolve

The rise of agents is forcing SaaS vendors to adapt their models. As companies integrate agents into their tech stacks, a new era of orchestration is emerging where the traditional SaaS subscription model is being challenged.

• More than one in three companies would switch from a current software-as-a-service (SaaS) vendor to acquire agent functionality.
• The majority of leaders (68%) believe the future will see agents augmenting SaaS in a complementary relationship, not replacing it entirely. The challenge, though, is figuring out one's place in this new landscape.
• The agent-to-agent (A2A) communications era is already beginning, with 50% of companies reporting that their agents are handing off work across different vendors and platforms.

Our strategic recommendations for sellers

That's a lot of information - my take?

AI agents are the most significant leap we've seen in enterprise technology in a decade, and they are powering an exciting shift from individual productivity to organizational velocity.

If you're interested in the future of AI agents (and you should be), here's what you need to know to leverage them effectively:

• Check hyperbole at the door: A majority of respondents (over 70%) felt the public narrative about agents is overhyped compared with the results they've seen. Sellers should be measured in their claims and transparent about current limitations (no, agents can't do everything - yet!).
• Offer variable pricing: Almost half of buyers prefer variable pricing for agents over subscriptions. This includes pricing based on consumption, conversations, actions, and outcomes. High costs, such as expensive subscriptions and seat licenses, were listed as the number one reason for non-adoption.
• Design for the agent-to-agent (A2A) future: The agent-to-agent era isn't next - it's now. Sellers must prepare for agents to act as co-workers to other agents, not just humans.
• Use third-party proof: Large enterprise buyers rated vendor-provided internal reliability metrics as the least trustworthy signals. To win them over, sellers need third-party proof from independent review websites (like G2), analyst reports, and evaluation-benchmark providers.

Read the full report today!

Download the complete new report, A Leap of Trust: AI Agents Are Winning Hearts and Wallets, to not only get these strategic insights, but also to uncover the top departmental "beachhead" for agent adoption, the exact trust signals needed to convert skeptical enterprise buyers, how to structure your pricing model to overcome non-adoption across the market, and so much more.

To check these out and discover additional insights from this new survey, download the complete 2025 AI Agents G2 Insights Report.