AI in Action: Early Lessons from Small Businesses on the Front Lines of Adoption

AI in Action:

Early Lessons from Small Businesses on the Front Lines of Adoption

Executive Summary

Artificial intelligence (AI) is moving from experimentation to early implementation across U.S. small and mid-sized businesses (SMBs), including construction, manufacturing, and home services. While awareness of AI is high, sustained and scalable adoption remains uneven. Evidence from recent research and industry surveys suggests that SMBs are beginning to realize measurable benefits, particularly in productivity, scheduling efficiency, safety, and back-office operations, yet face persistent barriers related to cost, skills, data readiness, and trust.

This white paper synthesizes current research on AI adoption in SMBs, with a focus on laborintensive, blue-collar sectors where adoption has lagged behind white-collar industries. It draws on findings from government agencies, industry associations, academic research, and small business surveys conducted since 2020. The goal is to provide decision-makers with a grounded view of where AI is delivering value today, what is preventing broader uptake, and what practical steps can accelerate responsible adoption.

Key Findings

AI adoption is accelerating amid economic pressure, with more SMBs viewing AI as a practical tool for resilience and competitiveness.

Early adopters report tangible gains in productivity, scheduling accuracy, safety monitoring, and administrative efficiency.

Workforce readiness and skills availability are consistently cited as the most critical determinants of successful adoption, alongside uncertainty about which tools to prioritize.

Data fragmentation and low digital maturity constrain the scalability of AI solutions, particularly in manufacturing and construction.

Trust, liability, and compliance concerns remain significant for owners without dedicated legal or IT teams.

Recommendations

Form a Small, Cross-Functional AI Working Group Focused on Real Business Problems | Small businesses that made the most progress with AI did not leave experimentation to one person or silo it within IT or marketing. Instead, they pulled together a small group of interested staff across roles like operations, marketing, finance, and leadership to test specific, practical use cases. SMBs described success when AI efforts were anchored in dayto-day pain points, such as drafting emails and proposals, reviewing contracts, automating hiring screens, or generating reports that previously took weeks.

““As a 65–70 person manufacturer, we realized that ‘we need AI’ was too vague to be a strategy; we had to start with specific business problems and bring a small, cross-functional group together to tackle them. By working closely with our software vendors, we’ve embedded AI directly into tools we already use, from flagging risks in complex customer drawings to analyzing machine data across shifts to uncover productivity gaps. For us, the real progress has come from applying AI to concrete, dayto-day challenges and gradually connecting our systems so the data becomes more actionable across the organization.”

— Thomas Deslongchamps, Pindel Global Precision

Start Small and Pilot Immediately, Even with Low-Risk, Everyday Tasks | The most consistent advice from SMBs was to stop waiting for a “perfect” AI strategy and instead begin with small, low-stakes pilots. Many businesses reported their first wins came from simple applications like marketing copy, social media calendars, email drafting, speech writing, or summarizing long documents, tasks that saved hours each week without requiring deep technical expertise or paid tools.

““At Zinc, we started small by piloting AI-driven estimation tools that pull accurate market data without requiring hours of manual, line-by-line input. While the tools aren’t fully integrated with our CRM, they’ve still significantly streamlined our preconstruction process and improved the accuracy of our numbers. The result has been meaningful time and cost savings, freeing up 8 to 12 hours a week and reducing the need to add additional full-time staff.”

— Courtney Zaugg, Zinc Contractors

”Leverage Ecosystem Partners to Reduce Risk and Accelerate Adoption | Small businesses consistently emphasized the value of trusted ecosystem partners in helping them navigate AI adoption with greater confidence. Training programs, shared tools, and industryspecific implementation guidance provided practical guardrails, clarified best practices, and reduced the risk of missteps. SMBs reported that structured support from partners, including tailored courses and real-world use case examples, helped them move from experimentation to more disciplined, sustainable implementation.

As AI becomes a critical driver of economic growth, small businesses are facing a daunting skills gap, with recent research showing only 14% of workers have been offered AI training. To ensure small businesses can benefit from this technology, Google is providing them with no-cost access to the Google AI Professional Certificate, delivering practical training on everyday use cases like analyzing data and producing professional marketing materials.

Introduction: Why AI Matters for Small Business Now

Artificial intelligence has become a central topic in discussions about productivity, competitiveness, and workforce transformation. Based on the latest research from the Connected Commerce Council, a majority of SMB leaders are investing in AI adoption as a critical strategy to keep their businesses competitive, successful, and growing.1

For small and mid-sized businesses in construction, manufacturing, and home services, however, the conversation is often marked by tension between promise and practicality. Headlines emphasize breakthrough technologies, while owners and operators grapple with labor shortages, tight margins, and operational complexity.

In these sectors, AI’s relevance is increasingly concrete. Scheduling tools can reduce downtime and missed appointments. Predictive maintenance can anticipate equipment failures before they halt production. Computer vision can improve jobsite safety and quality assurance. At the same time, many SMBs lack the capital, staff capacity, and digital infrastructure to adopt advanced tools at scale.

This paper is informed by a series of roundtable discussions sponsored by the U.S. Chamber of Commerce Foundation and Google, designed to surface real-world experiences of SMB leaders navigating AI adoption. It focuses on firms that have been operating for at least 18 months and employ fewer than 500 workers, with particular attention to construction, manufacturing, and home services such as HVAC. The intent is to clarify where AI is delivering value today and what conditions are required for responsible, scalable use.

The State of AI Adoption in SMBs: Levels and Outcomes

Across the broader small business landscape, AI awareness is high, and scaled implementation is growing. Approximately 79 percent of small business leaders report understanding how AI can be applied in their operations, and 53 percent say AI tools are already critical to their success or will be in the near future.2 Early adopters cite improvements in productivity and operational efficiency. The prevailing challenge is not skepticism about AI’s potential, but uncertainty about which tools to prioritize and how to deploy them responsibly at scale.

AI adoption among small and mid-sized businesses in manufacturing and construction is also advancing, but remains early and uneven. In manufacturing, only 8 percent of firms surveyed report fully using AI at scale in their operations today, underscoring how limited enterprise-scale deployment still is across the sector. Most manufacturers are still at the front end of the adoption curve; 58 percent report being in awareness or research stages, rather than implementation, and predictive analytics and advanced AI remain among the least mature technologies in use.3 These patterns reflect that manufacturers are interested in the technology but may need additional support in order to begin using it within their organizations.

Construction presents a related but distinct profile. Executive intent is high: 66 percent of construction leaders expect AI to become essential within the next two to three years, particularly for improving project delivery, safety performance, and cost control.4 Industry reports increasingly highlight AI use cases across estimating, scheduling, computer vision safety monitoring, and equipment automation.

SMB leaders understand how AI applies to their operations

Yet adoption on the ground lags stated ambition. Many contractors are still developing the data maturity, process discipline, and workforce readiness needed to translate intent into measurable outcomes.

Where AI has produced results in manufacturing, successes are often difficult to replicate. National research characterizes many current wins as “heroic efforts,” highly customized implementations requiring advanced expertise and intensive integration.5 While these efforts demonstrate proof of value, they do not readily scale across equipment, facilities, or firms, limiting their broader impact and slowing diffusion across the SMB landscape.

Evidence from the wider small business ecosystem points to a clearer near-term path. SMB owners most commonly deploy AI to streamline routine operational and administrative tasks, with consistent gains in efficiency and productivity. Importantly, these gains are amplified when workers receive targeted training to use AI tools effectively.6 Across manufacturing, construction, and home services, early outcomes such as time saved, fewer errors, and faster response cycles are already achievable, but they remain tightly linked to workforce readiness and clearly defined, operationally grounded use cases.

Drivers and Barriers to Adoption

DRIVERS

Across construction, manufacturing, and home services, interest in AI is being driven less by experimentation and more by operational pressure. Persistent labor shortages, rising costs, and tighter customer expectations are pushing SMBs to seek tools that improve efficiency, reduce downtime, and streamline routine work. Evidence from broader SMB surveys shows that when AI is deployed effectively, particularly for scheduling, maintenance, and administrative tasks, firms can achieve measurable productivity gains. These gains are often amplified when workers receive targeted training, suggesting that AI’s value is closely tied to workforce readiness rather than technology alone.

BARRIERS

Despite these drivers, barriers to adoption remain substantial and interconnected. Cost and return on investment dominate decision-making. National survey data indicates that 30 percent of SMB owners view AI tools as too expensive, while 34 percent report that available tools are not useful for their business.7 These perceptions are especially common when benefits are framed as incremental productivity improvements without clear financial translation.

In manufacturing, workforce constraints are as significant as financial ones. Manufacturers most frequently cite the absence of clear business cases and workforce skill gaps as the leading obstacles to adoption.8 Even relatively affordable tools require staff who can integrate AI with existing systems and interpret outputs in operational contexts.

Data readiness further limits scale. Research consistently identifies unstandardized and fragmented data such as inconsistent equipment logs, schedules, and material records as a major barrier to training and deploying AI across sites.9 These challenges are compounded by low digital maturity: many SMBs lack integrated data systems or digitized workflows, which are prerequisites for effective AI use.

Finally, awareness itself is uneven. Adoption narratives and outreach efforts are disproportionately concentrated in white-collar sectors, leaving construction, manufacturing, and home services with fewer visible, sector-relevant examples.10 This imbalance reduces confidence and slows adoption, even where practical use cases exist.

How SMBs Are Using AI Today: Priority Use Cases

CORE OPERATIONAL USE CASES: SCHEDULING, SAFETY, AND MAINTENANCE

Among construction and manufacturing SMBs, AI adoption is most visible in discrete operational functions where value can be measured quickly. In construction, AI-enabled project scheduling and crew optimization tools are emerging as practical applications. Contractors piloting these tools report reductions in costly overruns by flagging schedule risks earlier and reallocating labor more efficiently. Safety applications are also gaining traction. Computer vision systems, site cameras, and AI-enabled wearables are being used to improve hazard detection and incident reporting, contributing to lower incident rates and supporting compliance efforts.11

In manufacturing, predictive maintenance represents a high-potential use case, though adoption remains uneven. Where deployed, AI models help anticipate equipment failures and reduce unplanned downtime. However, these applications often require cleaner, more consistent machine data than many SMBs currently possess, limiting broader uptake.

BACK-OFFICE AND CUSTOMER-FACING APPLICATIONS

Across home services and smaller construction firms, backoffice and customer-facing automation is the most common starting point for AI adoption. Surveys document HVAC and service businesses using AI scheduling assistants to reduce missed appointments and increase weekly service calls.12 Additional applications include billing, invoicing, inventory tracking, and basic customer service functions. These tools typically require lower upfront investment, integrate more easily with existing systems, and deliver faster time-tovalue, making them attractive entry points for resourceconstrained SMBs.

“

“I didn’t wait for a formal strategy, I just started experimenting with tools like ChatGPT and Gemini on real, everyday tasks. I used AI to build an entire year-long social media and email plan, draft the content, and even flag where we could consolidate messages to avoid overwhelming customers. What began as a time-saving experiment quickly became a practical way to handle marketing, structure messaging, and reclaim hours each week.”

— Emily Dupree, Thunderbird Golf

PROCUREMENT, SUPPLY CHAIN, AND THE LIMITS OF SCALE

Procurement and supply chain forecasting represent another promising but underdeveloped area. AI-driven demand forecasting can reduce inventory costs and prevent shortages by anticipating fluctuations.13 However, national research finds that many SMBs lack the data depth and standardization required to support these models, making vendor partnerships or industry-level solutions critical. Across all use cases, most AI deployments remain isolated pilots rather than scaled solutions. Research consistently emphasizes the need for reusable use-case blueprints and human oversight to validate outputs, build trust, and enable repeatable adoption.

Workforce, Skills, and Change Management

WORKFORCE READINESS AS A LIMITING FACTOR

Across construction, manufacturing, and home services, AI adoption is constrained as much by workforce capacity as by technology or cost. In manufacturing, workforce shortages are acute: the sector faces an estimated shortfall of roughly 500,000 workers, and one in four employees is age 55 or older. 14 Many owners view AI as a potential tool to offset labor constraints and extend workforce productivity. Yet survey evidence shows that the scarcity of workers with the skills to implement, integrate, and operate AI systems remains a primary barrier.

Manufacturers consistently rank workforce skill gaps alongside ROI uncertainty as the leading obstacles to adoption.

These challenges are especially pronounced in environments with legacy equipment and fragmented data systems. Owners report difficulty finding staff who can connect AI tools to existing machines, interpret model outputs, or translate insights into operational decisions.15 As a result, even firms motivated to adopt AI often struggle to move beyond pilots.

TRAINING, UPSKILLING, AND HUMAN OVERSIGHT

National guidance underscores that workforce development is foundational to AI adoption. Research calls for coordinated training pathways spanning community colleges, technical schools, and on-the-job learning within SMBs, supported by public investment and financial incentives.16 Without structured, accessible training, small firms lack the capacity to close skill gaps at the pace required for adoption. Recent data shows that only 37 percent of employees have received organizational guidance around AI use, and a mere 14 percent have been offered AI training; small businesses are nearly three times less likely to offer AI training than large companies.17

Change management is equally critical. Evidence emphasizes that AI systems are most effective when designed to preserve human authority and oversight.18

Human-in-the-loop approaches, where workers validate outputs and retain decision-making responsibility, support trust, reduce resistance, and promote responsible use. This is particularly important for aging workers, whose experience and judgment remain valuable but may require targeted upskilling to adapt to new tools.

of employees have received organizational guidance around AI use 37% of employees say their organization offered AI-related training in the last 12 months 14%

small businesses are nearly three times less likely to offer AI trainings than large companies 3X

In construction, workforce readiness is widely viewed as the determinant of value capture. Industry reports stress that AI investments only translate into improved project outcomes when superintendents, project managers, estimators, and tradespeople are trained to apply tools in scheduling, safety monitoring, and cost control.19 Without this readiness, software investments risk being underutilized, limiting impact.

“Training has looked different depending on the tool. For rendering software, it was largely hands-on experimentation: sitting down together, testing prompts, reviewing quick tutorial videos, and learning through trial and error to see what produced the best results. For our preconstruction software, the approach was more structured, with formal walkthroughs and guided in person training to ensure the team understood how to input data and use reporting features effectively. As a result of implementing these tools, we’re now able to generate accurate, comprehensive reports on demand and walk into leadership meetings fully prepared without the manual work, delays, or risk of human error that used to slow us down.”

— Megan Williams, Compton Construction

Trust, Policy, and Compliance: Navigating Risk in a Small Firm

For small and mid-sized businesses, uncertainty around trust, compliance, and liability remains a material barrier to AI adoption. Many SMBs operate without in-house legal counsel or compliance officers, making plain-language guidance on data privacy, intellectual property, and workforce-related risks essential. Survey evidence shows that without clear regulatory interpretation, owners are hesitant to deploy AI in customer- or employee-facing systems, particularly in safety-sensitive environments.20

In manufacturing, the absence of consistent data and interoperability standards further amplifies risk. National research finds that AI pilots frequently remain siloed because systems cannot integrate across equipment, facilities, or vendors.21 This fragmentation limits scalability and increases liability exposure, as firms rely on bespoke solutions without shared standards to clarify accountability, data ownership, or system performance expectations.

Construction firms face heightened sensitivity to these issues. Liability considerations shape technology decisions, and AI tools must align with business- and sector-specific compliance frameworks.22 Without explicit guidance recognizing AI-enabled monitoring and decisionsupport tools as compliant and low-risk, firms are likely to delay adoption despite potential safety and productivity benefits.

Policy incentives emerge as a critical counterbalance to these risks. Evidence highlights that tax credits, adoption grants, and subsidized training programs can reduce upfront costs while encouraging responsible use aligned with emerging standards.23 For capital-constrained SMBs, these incentives lower financial exposure and make investments in both technology and workforce readiness more feasible.

Public–private partnerships play a central role in translating policy into practice. Coordinated efforts among government agencies, technical centers, industry associations, and vendors can provide shared guidance, standardized resources, and trusted compliance pathways.24 Together, clear standards, accessible guidance, targeted incentives, and coordinated partnerships form the regulatory backbone needed to support trustworthy AI adoption across construction, manufacturing, and home services.

Enablers and Success Patterns

SOCIAL AND APPLIED ON-RAMPS

Evidence consistently shows that SMBs adopt AI through practical, trust-based pathways rather than abstract exposure. Manufacturers most often report learning about AI and deciding whether to implement it through peers (22 percent), case studies (21 percent), and vendor support (20 percent).25 These findings underscore that community validation and applied examples are more influential than generalized awareness campaigns. Peer learning and vendor-assisted pilots help firms translate interest into action by reducing perceived risk and clarifying operational relevance.

Trusted local intermediaries further strengthen these on ramps. Small Business Development Centers, chambers of commerce, and technical assistance providers play a critical role when they combine candid risk guidance with industry-specific training.26 This approach increases confidence among owners and crews, particularly in sectors with limited prior exposure to AI.

HUMAN-IN-THE-LOOP AND SCALABLE INFRASTRUCTURE

Across construction and manufacturing, the most effective operating model is augmented intelligence rather than automation. Industry research emphasizes AI systems that support, rather than replace, superintendents, project managers, estimators, and safety personnel.27 Human-in-the-loop designs clarify accountability, reduce workforce resistance, and allow firms to target immediate, decision-supported use cases with measurable outcomes.

Scaling these successes requires technical and institutional scaffolding. National research converges on three enabling conditions: standardized data and interoperability, shared platforms for AI tools and models, and reusable use-case “blueprints” that codify ROI assumptions and implementation steps.28 Together, these elements convert isolated pilots into repeatable playbooks that can be adopted with lower cost and risk.

Because SMBs face tight capital and capability constraints, financial incentives and workforce training are critical complements. Evidence points to the value of funded training pathways, including community colleges, and incentives for digital upgrades that allow

legacy equipment to participate.29 International experience reinforces this approach: coordinated testbeds, applied mentors, and sustained investment accelerate adoption when adapted to regional and sector-specific contexts.

Gaps in the Evidence and Questions for the Field

Limited industry-specific insight. Most existing research examines small businesses in aggregate, with relatively little focus on construction, manufacturing, and home services. As a result, sector-specific operating conditions, data environments, and workforce realities are underrepresented.

Inconsistent evidence on ROI. Case studies frequently report productivity or efficiency gains, but rarely provide systematic data on costs, payback periods, or scalability. This limits SMBs’ ability to assess financial risk and compare investment options.

Unclear pathways from pilot to scale. Many reports highlight isolated successes but offer little guidance on how SMBs can replicate AI deployments across multiple sites, projects, or business units.

Underdeveloped training models. While workforce skill gaps are well documented, there is limited comparative evidence on which training approaches (vendor-led, community college–based, peer learning, or on-the-job) are most effective for frontline workers.

Ambiguity around compliance and liability. Small firms remain cautious about regulatory risk, particularly in safety-driven sectors. Existing guidance often lags behind available tools and lacks clarity on compliance expectations.