Modest Confidence, Real Results: What Midmarket Leaders Are Actually Building With AI

JR
Jun 18
7 min read

Updated: Jul 4

What the Survey Reveals About AI Readiness

Executive Summary

Midmarket leaders rate their 2027 AI readiness at 6.3/10—cautious but committed to moving forward.
Revenue growth and customer experience emerge as equally important outcomes; cost reduction and talent/skills improvement lag far behind.
Talent and skills remain the top blocker (60%), yet 80% of respondents have shipped at least one AI pilot and 40% have shipped three or more.
Clear ownership matters: 40% have a named CEO or GM owner; 30% have no clear owner and 30% are split between functional leaders and working groups.
Data readiness is solid for half the respondent base; 70% have governance protections in place or partially enforced, suggesting leadership is taking risk seriously.
What resonated in the workshop: frameworks that expanded imagination, thought-provoking content, and validation of existing progress. What needs adjustment: pacing, duration, and pre-session preparation.
This analysis is based on a small workshop cohort (n equals 10) and offers directional rather than statistically generalizable findings.

What the Survey Reveals About AI Readiness

Outcomes Leaders Want

Two priorities dominate the midmarket AI conversation: revenue growth and customer experience. In this 10-person cohort, four respondents cited revenue growth as their top outcome, and four cited customer experience. No other outcome claimed more than one vote.

This split is revealing. It suggests that midmarket leadership sees AI not as a cost-cutting tool or a way to backfill talent gaps, but as a lever for growth and competitive differentiation. Revenue growth is the survival metric; customer experience is the stickiness metric. Together, they form a coherent vision: use AI to expand market opportunity while deepening customer relationships.

The modest priority given to cost reduction (one respondent) contradicts the narrative that midmarket companies turn to AI primarily to do more with less. Instead, the data suggests a maturation: leaders are asking "what new capability does this unlock?" rather than "what can this eliminate?"

What's Blocking Progress

If outcomes are clear, the path to achieve them is not. Talent and skills emerge as the top blocker by a significant margin: six of ten respondents named talent shortage as their primary constraint. Three cited leadership buy-in, and one cited regulation or compliance.

This ranking confirms a widely reported challenge. Yet the data contains a paradox worth exploring. Despite naming talent as their top blocker, 80% of the respondents have shipped at least one AI pilot, and 40% have shipped three or more. Something is enabling motion despite the stated constraint. That something, the data suggests, is pragmatism: leaders are building with what they have, accepting that perfect talent is not a prerequisite for progress.

Leadership buy-in (30% of blockers) is a distant second. This is encouraging. It implies that in this cohort, executive alignment around AI is largely present; the problem is not conviction but capability.

The Ownership Gap and Why It Matters

One of the clearest patterns in the data is fragmentation of ownership. When asked who owns their primary AI initiative, respondents split four ways:

40% have a named CEO or GM owner (accountable, visible, in control).
30% have no clear owner at all.
20% rely on functional leaders (Sales, Operations, IT).
10% have a working group without a single owner.

The ownership gap correlates with other variables in the data. Respondents with a named CEO or GM owner report higher governance strength, faster response speeds, and more consistent KPI tracking. Respondents with no clear owner are more likely to report scattered data, informal governance habits, and no KPIs tied to AI.

Ownership is not about having an AI expert in the room; it is about having someone accountable for outcomes, reviews, and course correction. Without that clarity, even well-intentioned pilots stall in the backlog or drift into maintenance mode.

Industry Intelligence: How 5 Sectors Are Responding to AI Right Now

Professional Services and Legal

What's changing: Legal and professional services firms are under margin pressure. Client expectations for faster turnaround times are rising, and talent costs are climbing. AI is being deployed as a margin-recovery tool and a competitive advantage in client advisory.

Where AI is being applied: Legal research and contract review (identifying risk patterns, flagging anomalies, accelerating document review), practice management (timekeeping, billing, matter accounting), and business development (proposal generation, client research).

Common pitfalls: Professional services firms often underestimate governance risk. Using generative AI on client work without guardrails or audit trails creates liability. Many also assume that AI will replace junior staff; in reality, junior work often evolves rather than disappears. The firms that succeed use AI to elevate junior staff to more advisory roles, not to downsize.

Industry context: 62 percent of legal professionals report using AI tools at work, though only 31 percent of law firms have a formal AI governance policy in place (training-data; verify before publishing). Investment in legal-tech AI has grown 35 percent year-over-year through 2025 (training-data; verify before publishing).

Financial Services and Insurance

What's changing: Financial services and insurance organizations are automating customer journeys and underwriting. Regulatory pressure is intense; compliance is not negotiable. AI is being deployed to improve speed, consistency, and risk assessment while maintaining audit trails.

Where AI is being applied: Underwriting (risk scoring, anomaly detection), claims processing (triage, damage assessment, fraud detection), customer service (robo-advisory, chatbots for policy questions), and anti-money laundering (AML) screening and transaction monitoring.

Common pitfalls: The regulated environment creates a false sense of safety. Having controls in place does not guarantee they are working. Many firms implement AI governance for new tools but leave legacy systems untouched, creating a patchwork. Another pitfall: assuming that historical data is clean. Bias in training data gets amplified at scale.

Industry context: 73 percent of financial services leaders plan to increase AI investment over the next two years (training-data; verify before publishing). Regulatory frameworks like the EU AI Act are creating compliance requirements that organizations are still learning to navigate. Firms that lead in AI are those that build governance early, not as an afterthought (training-data; verify before publishing).

Construction and Real Estate

What's changing: Construction is one of the least digitized industries, but the labor shortage is driving urgency. AI is being deployed for project management, resource forecasting, and safety monitoring. Real estate and construction firms are also exploring AI for site inspection and quality assurance.

Where AI is being applied: Project scheduling and resource allocation, safety monitoring (detecting unsafe behaviors on job sites), design optimization (cost and waste reduction), and predictive maintenance (identifying equipment failure before it happens).

Common pitfalls: Construction companies often lack the data infrastructure to feed AI tools. Site conditions change rapidly, and models trained on historical data quickly become stale. Another challenge: adoption. Field teams are often skeptical of technology; AI tools that add friction or feel disconnected from how work actually happens get abandoned.

Industry context: 47 percent of construction firms report using some form of AI or automation in project management; only 19 percent have a dedicated AI strategy (training-data; verify before publishing). The sector is ripe for disruption but is moving slowly due to fragmented technology adoption and workforce resistance (training-data; verify before publishing).

Healthcare and Life Services

What's changing: Healthcare providers and life services organizations are facing dual pressure: rising costs and aging populations. AI is being deployed to improve diagnostic accuracy, streamline administrative work, and optimize resource allocation.

Where AI is being applied: Clinical decision support (diagnostics, treatment recommendations), claims processing and coding, staff scheduling and resource optimization, and patient engagement (appointment reminders, follow-up communication).

Common pitfalls: Healthcare organizations are heavily regulated, and many err on the side of caution, delaying deployment. Data silos are common; legacy EHR systems do not talk to one another, making it hard to feed AI tools with complete data. Privacy concerns are real, and mistakes are costly. Some organizations become so focused on compliance that innovation stalls.

Industry context: 64 percent of healthcare organizations have implemented or are piloting AI tools (training-data; verify before publishing). However, only 41 percent have a formal AI governance policy that covers both clinical and operational use (training-data; verify before publishing). Leaders in this space have separated innovation sandboxes from production, allowing them to experiment safely.

Manufacturing, Distribution, and Business Services

What's changing: Manufacturing and distribution are under pressure from supply chain volatility and labor shortages. AI is being deployed to optimize inventory, predict demand, and automate repetitive administrative work.

Where AI is being applied: Demand forecasting and inventory optimization, preventive maintenance (monitoring equipment health), quality assurance and defect detection, and supply chain visibility and risk assessment.

Common pitfalls: Many manufacturing organizations have poor data quality to begin with. Inventory records, maintenance logs, and quality metrics are scattered across systems or stored in silos. AI can amplify bad data faster than humans can catch it. Another pitfall: automation bias. Once an AI system is deployed, humans stop questioning its outputs, even when conditions change.

Industry context: 58 percent of manufacturing firms report implementing AI in some form; 35 percent cite poor data quality as a significant barrier to deeper AI adoption (training-data; verify before publishing). Investment in AI for manufacturing is expected to grow at 22 percent annually through 2027 (training-data; verify before publishing).

What High-Performing Organizations Are Doing Differently

Across the five sectors and the midmarket cohort, organizations that are shipping AI and seeing results operate differently on five dimensions.

Ownership. High-performing organizations have named their AI owner. The title varies (CEO, CTO, VP of Ops), but the responsibility is clear: one person is accountable for outcomes, reviews, and priority decisions. This owner has air cover from leadership and authority to cross functional boundaries.

Capability. Rather than waiting for the perfect hire or external expert, high-performing organizations conduct an honest audit of internal skills and external needs, then build a roadmap: what can we do today with existing talent? What do we need to hire or contract for? Where do we need training? This pragmatic approach unlocks motion.

Governance. High-performing organizations have not perfected governance, but they have not ignored it either. They have established guardrails for data access, tool usage, and output review. These guardrails are documented and checked regularly, not treated as suggestions. Governance is an enabler of faster decision-making, not a brake.

Workflow Design. Rather than trying to automate entire processes at once, high-performing organizations identify one bottleneck or high-impact decision point, deploy AI there, measure the impact, and iterate. This focused approach builds credibility and generates fast wins.

Measurement. High-performing organizations tie AI outcomes to business metrics: revenue, customer satisfaction, cycle time, cost. Every pilot has a named owner, a target metric, and a monthly review rhythm. If a pilot is not moving the needle after 60 or 90 days, they course-correct or kill it rather than letting it drift.

Continue the Conversation at GPS Summit

The Chicago workshop confirmed what we hear from midmarket leaders everywhere: confidence in AI readiness is modest, but commitment to shipping is real. Leaders know what outcomes matter and what blocks progress. What they need next is connection to peers solving similar problems, access to frameworks they can implement immediately, and air cover to invest in capability building without waiting for perfection.

The GPS Summit is designed for exactly this conversation. Bring your ownership questions, your data readiness challenges, and your governance concerns. Connect with leaders who have shipped pilots and lived through the hard decisions.