Midyear update: 2025 AI predictions

Our prediction: Your AI strategy will put you ahead — or make it hard to ever catch up

What happened: 2025 has indeed started to set apart winners and losers. We’re seeing companies with a bold vision and decisive AI strategy pulling ahead. This is most visible in the tech sector, but we’re also seeing this play out in financial services, healthcare, retail, transportation and other sectors. The urgency is growing as more companies see the AI imperative and it continued to be a key topic in quarterly earnings commentary and analysis. This is especially the case now that AI agents are proving they’re ready for the real work.

Despite macroeconomic uncertainties, companies aren’t pulling back on AI investment. In fact, in our May 2025 AI Agent Survey, 88% of executives told us their team or function plans to increase AI-related budgets over the next 12 months — specifically because of agentic AI’s potential. And 73% agree that how they use AI agents will given them a significant competitive advantage in the coming 12 months.

What to do: For AI strategy that will keep you ahead, think big. Consider how AI can and will drive new business and operating models. AI itself can help. AI-driven strategic planning continuously gathers and analyzes data — real-time market signals, third-party datasets, proprietary business information — providing up-to-date insights into market trends, customer behaviors and operational performance. Tools like PwC’s Agent Powered Performance engine can help drive your strategy end to end, scanning enterprise data, simulating scenarios and finding performance levers — letting you conceptualize business outcomes and work backwards to guide technology enablement.

Our prediction: Your workforce could double — thanks to AI agents

What happened: This prediction is also on track: AI agents are proving able to double workforce capacity — while increasing human workers’ value. PwC’s 2025 AI Jobs Barometer, which analyzed close to a billion job ads from six continents, found that industries exposed to AI have nearly three times higher growth in revenue per employee than less exposed industries.

At one hospitality company, we helped embed AI agents across the business to fuel growth, speed and customer satisfaction. The company realized gains in productivity up to 90% in some areas while enabling its franchise owners to innovate and enhance the guest experience. We’re also seeing the impact of AI agents firsthand. PwC US has deployed hundreds of agents — some as assistants to our people and others working autonomously with human oversight. They’re helping us achieve productivity gains of up to 50% in IT, finance, tax and more.

What to do: The key to quickly achieving value at scale with AI agents? Orchestration. A platform like PwC’s agent OS connects multiple agents — across apps and platforms — into seamless workflows. It brings agents and people together as integrated teams with enterprise system access and built-in oversight. Because while one agent can handle a narrow task, it takes a coordinated system of agents to securely transform complex processes like financial close, customer service, talent sourcing and software development.

Our prediction: ROI for AI depends on Responsible AI

What happened: Leading companies, especially in healthcare, financial services and technology, are seeing responsible AI help grow ROI. That was the case for one of our pharma clients, which prioritized a responsible approach. The company engaged its legal leadership early — inviting its general counsel and hundreds of attorneys to explore generative AI firsthand, not to flag risks but to spot opportunities. In doing so, it helped build trust and momentum across the organization. When AI is high-performing, secure and compliant, with humans at the helm, people trust it enough to use it. Its outputs are reliable enough to create value.

But in many companies, trust concerns are still holding AI back. In our AI agent survey, over a quarter (28%) of executives ranked lack of trust as a top challenge to realizing value. Some risks are real. When AI agents act autonomously within complex, high-value workflows, they could make unintentional mistakes — or a malicious actor could corrupt them.

What to do: Embed responsible AI from the start and make sure your approach and solutions are evolving with the technology. That means putting layered controls and independent assurance in place. It requires refreshed third-party risk management and a stronger cybersecurity posture. And it calls for orchestration tools with built-in security measures — real-time data access, API restrictions, prompt controls — and automated checkpoints that trigger human or machine intervention. Evolve oversight from one-time reviews to dynamic, ongoing monitoring. Whether through internal audit or independent assessment, consistent AI governance is critical to short- and long-term value creation.

Our prediction: AI will be a value play — and a boon for sustainability

What happened: Six months after our prediction, AI is no longer just a support tool for sustainability. It’s becoming central to how companies drive value amid shifting energy markets, grid capacity constraints, regulatory pressure and rising consumer expectations. While training large language models can be energy-intensive, AI is also proving essential in reducing overall energy demand. In fact, PwC research released after our original prediction suggests that, over time, AI could enable energy savings that rival its own consumption.

In the consumer-packaged goods sector, for example, AI is helping brands embed sustainability into every stage of the product life cycle — from predicting demand for product features to selecting low-impact materials and recyclable packaging. These decisions can reduce environmental impact while tapping into customer demand for sustainable options, a move that can command a 26% average price premium.

AI is also improving systems like HVAC and lighting, forecasting energy loads and managing renewable supply variability. On the regulatory front, AI tools are flagging disclosure gaps, tracking global policy shifts and drafting early versions of climate-related reports — helping companies keep pace with evolving standards while tying sustainability to financial outcomes.

What to do: This prediction is no longer just about AI enabling sustainability. It’s also about sustainability amplifying the value of AI. In the years ahead, businesses that embed sustainability into their AI strategies — and AI into their sustainability strategies — will gain more efficiency while also driving drive more value. They’ll capture market share, build trust and unlock entirely new levers of growth.

Our prediction: AI will cut product development lifecycles in half

What happened: AI has started to significantly cut product development life cycles — but so far only in a few areas. In software development, AI is writing code, finding bugs and supporting testing, often in near real time. It’s enabling smaller, AI-native companies to quickly scale and take on industry leaders. In consumer goods and fashion, AI streamlines early-stage decisions, reducing iteration time and helping teams get products to market with greater speed and confidence. But in other areas, with longer development timelines or regulatory requirements, the change may be more incremental. Innovative pharmaceutical companies, for example, may use AI to help identify molecules that humans could not have found on their own.

While we haven’t seen rapid progress across the board yet, we still expect this prediction to come true in the coming years. That’s in part because R&D is also leveraging agentic AI. In our AI agent survey, about one-third of executives (35%) told us they’re using AI agents to innovate products and services. As multi-agent systems mature, product development will shift from periodic sprints to continuous cycles of learning, testing and refinement. This acceleration isn’t just operational — it’s strategic.

What to do: AI is ready to reinvent product development — but most companies aren’t. If your teams aren’t yet using AI to design, simulate and test physical products, now’s the time to start. Multimodal AI can process everything from CAD files to sensor data, helping engineers explore more possibilities, iterate in hours and troubleshoot before a prototype is built. To get ahead, you should modernize your data infrastructure, embed AI tools into engineering workflows and close the talent gap with upskilling and new hires.

Our prediction: AI will transform industry-level competitive landscapes

What happened: Signals of disruption are already here. AI isn’t just transforming industries — it’s reshaping where and how value is created. PwC’s research highlights areas for business growth and innovation that align to core human needs. In these domains, AI acts as a force multiplier — accelerating design, unlocking precision and enabling partnerships that weren’t viable before.

The upside is significant. If responsibly deployed, AI could expand global GDP by nearly 15% by 2035 — adding more than 1 percentage point of growth per year, comparable to the economic lift of 19th-century industrialization. In healthcare, for example, those companies that leverage their unique data sets could innovate new, high-value offerings. In manufacturing and infrastructure, early adopters are cutting development cycles, improving quality and reshaping supply chains in real time. Industries that combine data intensity with operational complexity — like healthcare, manufacturing and financial services — are often well positioned to lead this shift.

Consumer markets companies are starting to make bold AI moves. Leading brands are shifting away from traditional media and search, treating GenAI as the new front door to product discovery. They're rethinking content for AI optimization, structuring product data, anticipating queries and building for seamless, AI-native experiences. Many are also adopting dynamic pricing models, using AI to respond in real time to demand signals, competitor actions and inventory levels.

What to do: Start by identifying where your business intersects with these emerging growth areas. Look for areas where AI can streamline operations, enhance decision-making or unlock new value through collaboration — especially across industries. Make trust a foundational principle by embedding responsible AI practices into every deployment. Modernize your data infrastructure so AI tools can access, learn from and act on the right signals at speed. And close your capability gaps now. Upskill your teams in AI fluency, bring in new talent and empower leaders to experiment and scale what works. Reinvention starts with readiness — and it’s time to get ready.