Managing Engineering Projects: A 2026 Leadership Guide

Why Engineering Firms Are Stuck at 27%

The AI adoption gap in engineering isn't a technology problem. It's a leadership and culture problem. The data makes it clear: firms that adopt AI see major returns, but most firms can't get past the starting line because they're underinvesting in the human side of adoption.

Start with training. 65% of AEC firms invest less than 10% of their technology budgets on training¹. They buy tools and assume people will figure them out. That assumption is expensive. In practical terms, firms are spending six figures on AI tools and single digits on helping people use them.

Then there's the leadership blind spot. C-suite leaders are more than twice as likely to blame employee readiness as they are to acknowledge their own role as a barrier to AI adoption². But here's the thing— employees are actually three times more willing to embrace AI than leaders realize². The bottleneck isn't the team. It's the corner office.

And even where individuals do adopt, it stays individual. 49% of AI workflows are built for solo use, which drives only 6% of downstream adoption by colleagues³. When AI lives on one person's desktop, it doesn't spread.

The irony is that firms that DO adopt see real results:

• 68% of AI-adopting AEC firms saved at least $50,000¹
• 46% reclaimed 500–1,000 hours using AI tools¹
• 94% of those firms plan to expand AI investment next year¹

The payoff isn't theoretical. But getting there requires something most firms aren't investing in. As the Zweig Group put it: "AI adoption doesn't deliver value on its own. It only works when paired with cultural transformation."⁴ (For a deeper look at these barriers, see our analysis of why engineering firms struggle with AI adoption.)

So why are engineering firms stuck at 27% when the rest of the business world has moved to 78%? The answer isn't what most leaders expect.

The Evidence for "Show, Don't Tell"

The most effective strategy for driving AI adoption is leadership modeling— leaders and peers who demonstrate how they use AI to solve real problems give others a concrete template to follow, according to an 18-month Stanford study published in Harvard Business Review⁵.

Over 18 months, the researchers studied nearly 2,000 professionals (including hundreds of interviews at Google) and landed on a four-skill framework: Define the Problem, Evaluate Tools, Experiment, and Integrate. Weekly demos walk participants through all four naturally.

As the researchers wrote: "When leaders and peers show— not tell— how someone identified a high-value problem, evaluated tools, ran a small experiment, and integrated the result into daily work, they give others a concrete template to follow."⁵

This isn't just academic theory. Multiple companies have independently validated the approach— and their results point to what building a culture of AI adoption actually requires:

Here's what's interesting about these results. At Shopify, the VP of Engineering described weekly demos as "the best way to determine if progress is happening." At Plaid, teams with AI-savvy managers adopted fastest— so they targeted manager enablement specifically. And at Varonis, enrichment sessions attended by hundreds became the cultural backbone of adoption.

Now, these are tech companies. But the underlying principles— leadership modeling, peer demonstration, team-level champions— translate directly to engineering firms. Why? Because engineers are evidence-based thinkers. Seeing a colleague use AI to review a spec or draft an RFP response is more persuasive than any mandate.

And the data backs up the training investment: regular AI usage is sharply higher for employees who receive at least five hours of training with access to in-person coaching⁹. Demos ARE that training.

Building a Weekly AI Demo That People Actually Attend

A weekly AI demo meeting that drives real adoption follows a consistent format: one team member presents a real workflow they improved with AI, the group discusses how to adapt it, and the firm tracks what gets implemented. All in 30 minutes.

Who presents: Rotate across the firm. Not just the tech-savvy people. The managing principal goes first— that models vulnerability and signals that this matters. Then project managers, engineers, administrative staff. The more diverse the presenters, the more people see themselves in the possibilities.

What to demonstrate: Real problems, not parlor tricks. The difference between a demo that drives adoption and one that becomes "AI show and tell" is whether someone walks away ready to change how they work tomorrow.

AEC-specific demonstrations that get people leaning forward:

• Document review: Feeding a 200-page specification into AI for summary and gap analysis. Bechtel developed a custom language model that reduced manual O&M documentation review from days to minutes¹⁰.
• RFP response: Using AI to draft proposal sections from project history and firm qualifications
• Scheduling and resource planning: AI-assisted conflict detection and resource optimization across active projects
• Safety compliance: Real-time guidance on regulatory requirements. Skanska launched Safety Sidekick, an AI assistant delivering real-time safety guidance using GPT-4o¹⁰.
• Meeting documentation: Automated summaries and action items from project kickoffs and owner meetings

Structure each session:

1. Problem (2 min): "I was spending 3 hours reviewing submittals each week..."
2. Tool + Approach (5 min): "Here's what I did in Claude/ChatGPT/Copilot..."
3. Result (3 min): "Now it takes 30 minutes, and here's what the quality looks like..."
4. Discussion (15 min): "How could you adapt this for your projects?"
5. Commitment (5 min): "Who's going to try something similar this week?"

That commitment step matters. It's the bridge between watching and doing. Each demo effectively walks the team through the HBR/Stanford framework— define the problem, evaluate the tool, see the experiment, then commit to integrating it⁵.

And the team-based format solves the individual adoption trap. When AI workflows are designed for team use rather than individual use, downstream adoption increases dramatically³.

What not to do: Don't let demos become impressive tricks with no workflow relevance. Don't skip the commitment step. And don't make attendance optional for leadership.

The Support System— Champions, Guilds, and Guardrails

Weekly demos start the fire, but AI champions, communities of practice, and clear governance keep it burning. The firms achieving full adoption pair visible leadership with distributed support at the team level.

AI Champions. Varonis appointed champions across groups as "field agents" who bridge leadership vision and team execution⁸. These aren't the managing principal— they're mid-level advocates, often project managers or senior engineers, who troubleshoot and encourage daily.

Manager engagement. Plaid's most important insight was that teams with AI-savvy managers adopted fastest⁷. Target manager enablement specifically. When a project manager uses AI in front of their team, the permission signal is immediate.

Community of practice. Varonis's AI Guild shaped standards and shared reusable patterns⁸. In an engineering firm, this might look like a shared prompt library for common tasks (submittal review, RFP boilerplate, meeting summaries), an internal wiki documenting what works, or a dedicated Slack channel where people share wins.

Governance as enabler. This is where engineering firms have a distinct concern— and it's legitimate. Engineers remain personally liable for AI-assisted work under ASCE and NSPE ethical codes. A PE stamp on an AI-generated calculation still means a professional engineer validated it.

Both are true. Professional liability IS real, and AI adoption IS necessary. The answer isn't avoidance— it's governance. And the data confirms it— workers at companies with AI usage policies are 55% more likely to report productivity gains³. Clear guardrails don't slow adoption. They make it safe.

Enterprises integrating change management into their AI programs are 47% more likely to meet their implementation objectives¹¹. For engineering firms, that means treating adoption as a managed initiative with scope, resources, and milestones— the same way you'd manage any other project.

What does that look like in practice? One professional services firm saw it firsthand. When Jeremy Zug's team at Practice Solutions started using AI, the shift happened when the whole team got comfortable with it as a "sparring partner"— a tool that magnified what they were already doing rather than replacing their expertise. As Zug described it, AI became "a tool that helps us do what we do best and magnifies what we're doing." The team's comfort grew from using AI together, not individually.

Measuring What Matters

Measure AI adoption by tracking what changes in daily work— usage rates by team, time saved on specific workflows, and business outcomes. Not by counting tool licenses purchased or demo attendance. The question isn't "are people logging in?" It's "are people working differently?"

What to track:

• Adoption rate by team/department: Varonis published team-level adoption scorecards for benchmarking and friendly competition⁸. Track which teams are using AI regularly, not just which individuals.
• Time saved on specific workflows: Document review, RFP drafts, scheduling, meeting summaries. Measure the before and after.
• Quality metrics: Error rates, revision cycles, review turnaround time. These tell you whether AI is improving work or creating new review burdens.
• Business outcomes: Project margin improvement, proposal win rate, client satisfaction scores. These are the numbers your partners and board will ask about.
• Training hours delivered: BCG found regular usage is sharply higher for employees who receive at least five hours of training with in-person coaching⁹. Track training hours completed, not just training "offered."

ROI benchmarks from the industry: Among AEC firms using AI, 68% report saving at least $50,000 and 46% have reclaimed 500–1,000 hours¹. But only the firms measuring these outcomes can prove it to their partners and boards.

Avoid vanity metrics. Demo attendance is an input metric, not an outcome. The signal that matters is what people do after the demo— did they try something new this week, and did it stick? But here's what separates firms that actually improve from firms that just feel busy— they track what people do after the demo, not whether they showed up. (For a deeper framework on what to track, see our guide to measuring AI success.)

The Managing Principal's Real Project

The firms that close the AI adoption gap treat adoption itself as a managed project— with a champion (the managing principal), a cadence (weekly demos), a team (AI champions), and measurable outcomes.

Managing engineering projects has always meant managing complexity, people, and risk. AI adoption is the same discipline applied to a different kind of project. 90% of the challenge is organizational, not technical. The managing principal's competitive advantage is leading that 90%.

Start small. One demo. Next week. You present first. Show your team how you used AI to summarize a project proposal, flag risks in a contract, or draft a response to an RFI. Watch what happens when the person at the top goes first.

For firms navigating this transition, an experienced AI implementation partner can help design the adoption program, train champions, and avoid the pitfalls that stall most engineering firms. Dan Cumberland Labs helps AEC firms build AI adoption programs that stick— starting with strategy, not software.

The engineering firms that will lead the next decade are the ones whose managing principals are leading AI adoption today. One demo at a time.

FAQ: Managing Engineering Projects in the AI Era

What percentage of engineering firms use AI?

Only 27% of AEC firms currently use AI for automation, problem-solving, or decision-making, according to Bluebeam's 2025 survey of over 1,000 AEC decision-makers¹. This compares to 78% of companies across all industries².

How can engineering firm leaders drive AI adoption?

Research from Stanford and Harvard Business Review shows the most effective approach is "show, don't tell"— leaders who regularly demonstrate how they use AI to solve real problems drive organic adoption far more effectively than top-down mandates⁵. Weekly demos, AI champions, and structured training form the foundation of successful programs.

What ROI do engineering firms see from AI adoption?

Among AEC firms using AI, 68% report saving at least $50,000 and 46% have reclaimed 500–1,000 hours annually¹. Firms that pair adoption with change management are 47% more likely to meet implementation objectives¹¹.

How long does it take to achieve AI adoption across an engineering team?

Plaid's engineering team hit 75%+ adoption within six months using structured demos, champions, and training⁷. AEC firms may see different timelines depending on team size and baseline technical fluency.

How do engineering firms handle professional liability with AI?

Engineers remain personally liable for AI-assisted work under ASCE and NSPE ethical codes. AI assists, but humans validate. Governance frameworks and clear usage policies are enablers of safe adoption, not barriers to it. Firms with AI usage policies see 55% higher productivity gains³.

References

1. Bluebeam, "New Bluebeam Report Shows Early AI Adopters in AEC Seeing Significant ROI Despite Uneven Adoption" (2025) — https://press.bluebeam.com/2025/10/new-bluebeam-report-shows-early-ai-adopters-in-aec-seeing-significant-roi-despite-uneven-adoption/
2. McKinsey & Company, "The State of AI in 2025" (2025) — https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai
3. Asana, "5 Innovative Ways to Encourage AI Adoption in Your Org" (2026) — https://asana.com/resources/encourage-ai-adoption
4. Zweig Group, "How AI and Culture Drive Agility in Architecture and Engineering Firms" (2025) — https://zweiggroup.com/blogs/the-zweig-letter/architecture-engineering-ai-adoption-culture
5. Harvard Business Review / Stanford University, "To Drive AI Adoption, Build Your Team's Product Management Skills" (2026) — https://hbr.org/2026/02/to-drive-ai-adoption-build-your-teams-product-management-skills
6. Bessemer Venture Partners, "Inside Shopify's AI-First Engineering Playbook" (2026) — https://www.bvp.com/atlas/inside-shopifys-ai-first-engineering-playbook
7. Plaid, "Transforming Engineers: How We Grew AI Coding Adoption at Plaid" (2025) — https://plaid.com/blog/ai-coding-adoption-plaid/
8. Varonis, "From Hype to Culture: How We Turned AI Adoption into Everyday Impact" (2026) — https://www.varonis.com/blog/impact-of-ai-adoption-engineering
9. Boston Consulting Group, "AI Adoption Puzzle: Why Usage Is Up But Impact Is Not" (2025) — https://www.bcg.com/publications/2025/ai-adoption-puzzle-why-usage-up-impact-not
10. American Society of Civil Engineers, "Architecture, Engineering, Construction Sector Slow to Adopt AI, Survey Shows" (2025) — https://www.asce.org/publications-and-news/civil-engineering-source/article/2025/12/18/architecture-engineering-construction-sector-slow-to-adapt-ai-survey-shows
11. Augment Code / Prosci, "6 Change Management Strategies to Scale AI Adoption in Engineering Teams" (2025) — https://www.augmentcode.com/guides/6-change-management-strategies-to-scale-ai-adoption-in-engineering-teams