# The 18-Month Ramp Is Negotiable **By Dan Cumberland** · Published August 23, 2026 · Categories: AI Strategy ![Featured image for The 18-Month Ramp Is Negotiable](https://cms.dancumberlandlabs.com/uploads/featured_image_fa54842175.webp) Illustration: Dan Cumberland Labs with Gemini. > Your newest project manager has been with the firm two years. She's getting better— you can see it. But she's not someone you'd put on a client call alone,... Your newest project manager has been with the firm two years\. She's getting better— you can see it\. But she's not someone you'd put on a client call alone, and you know it'll be at least another year before that changes\. That timeline isn't a coincidence\. The on\-the\-job training required to develop an architectural project manager runs 2–4 years after hire, with full PM readiness typically requiring 5–7 years of total professional experience[1](/blog/blog-the-18-month-ramp-is-negotiable#ref-1)\. That's the industry reality for mid\-level AEC professionals— and 72% of AEC firms identify the need to better train staff as a current priority[2](/blog/blog-the-18-month-ramp-is-negotiable#ref-2), which means most principals are living with this problem while also recognizing they haven't solved it\. The ramp takes as long as it does for three specific reasons: institutional knowledge lives in senior people's heads, not in any system new hires can access; mentors are stretched too thin to be reliably available; and judgment requires real project time to develop\. AI addresses the first two directly\. It does not address the third— and that distinction is where most of the current AI\-in\-training conversation goes wrong\. ## The Missing Middle— Why This Is Getting Worse Seventy\-six percent of AEC firms cite "The Missing Middle" as a current challenge— a shortage of experienced mid\-career professionals with 5–12 years of experience who can lead projects and develop junior staff[2](/blog/blog-the-18-month-ramp-is-negotiable#ref-2)\. The problem isn't new\. But the timeline is accelerating\. Industry projections suggest that at least 41% of AEC workers, including many in management roles, expect to retire by 2031, reducing not just headcount but the pool of experienced mentors available to develop the next generation[3](/blog/blog-the-18-month-ramp-is-negotiable#ref-3)\. What leaves with those senior professionals isn't just labor hours\. As Certis AEC put it in their 2026 industry challenges report, it is "decades of accumulated field judgment, site problem\-solving, and project delivery expertise that cannot be replaced through classroom training alone\."[4](/blog/blog-the-18-month-ramp-is-negotiable#ref-4) Two pressures are hitting simultaneously: - **Senior departures**— 40%\+ of the construction workforce retiring by 2031, taking with them the mentoring capacity that has always carried knowledge forward - **Junior contraction**— According to AIA Work\-on\-the\-Boards survey data, 17–22% of design firms are already eliminating positions for staff with less than six years of experience[5](/blog/blog-the-18-month-ramp-is-negotiable#ref-5) Fewer juniors entering today means a deeper Missing Middle in five years\. You can't hire your way out of this fast enough\. Training must get more efficient— and a specific part of the ramp is now genuinely compressible\. This is the context in which AI training tools arrive\. The question isn't whether you should pay attention\. It's what they can actually move\. ## What AI Can Compress— and What It Still Can't AI can meaningfully compress the knowledge\-transfer layer of professional ramp time\. It cannot compress the judgment layer\. That distinction matters— and most AI\-in\-training content ignores it\. ```html-table
AI Can AccelerateAI Cannot Replace
On-demand retrieval of institutional knowledge: policies, contract language, firm-specific decision frameworksJudgment— the pattern recognition built through supervised experience on real projects with real stakes
Simulated practice in scenarios that normally take years to encounterLearning to recognize when AI output is wrong (requires foundational judgment first)
Personalized feedback on document drafts, scope reviews, budget narrativesSoft skills from real client exposure, conflict, and accountability
Recall of the reasoning behind firm procedures— "why we handle RFIs this way"The feedback loop of real decision-making with a mentor present
``` The adjacent\-industry signals are worth taking seriously\. In software development, ramp\-up time has fallen more than 50% since Q1 2024, driven primarily by AI coding assistants[7](/blog/blog-the-18-month-ramp-is-negotiable#ref-7)\. LinkedIn's 2025 Workplace Learning Report— based on surveys of nearly 1,000 L&D professionals— found that AI tutoring tools increase knowledge retention by 25–40%, and skills\-based development programs reduce time\-to\-proficiency by 30–50%[6](/blog/blog-the-18-month-ramp-is-negotiable#ref-6)\. These aren't AEC\-specific numbers— but they're the best directional signal available, and they're consistently pointing the same direction\. There's a parallel risk worth naming directly\. The Thomson Reuters Institute's research on legal training put it plainly: "Because AI produces confident, polished output regardless of whether it's correct, professionals who never built judgment through direct experience have no reliable way to distinguish the tasks where AI excels from those where it fails\."[8](/blog/blog-the-18-month-ramp-is-negotiable#ref-8) And McKinsey research found that AI\-exposed junior roles are already seven times more likely to demand traditionally senior skills like leadership and strategic thinking[9](/blog/blog-the-18-month-ramp-is-negotiable#ref-9)— which means the pressure on judgment development is increasing, not decreasing, as AI handles more routine work\. The goal is AI for knowledge, mentors for judgment— not AI instead of mentors\. ## What This Looks Like in Practice— AEC Firms Already Doing It Several AEC firms have already used AI to capture and distribute institutional knowledge— and the same infrastructure that preserves expertise from retiring staff can be redeployed to accelerate training for incoming professionals\. **MBH Architects** captured their CFO's decades of contract judgment through structured conversations before her retirement\. Knowledge Architecture[10](/blog/blog-the-18-month-ramp-is-negotiable#ref-10) helped distill those conversations into an NDA review agent\. Managers can now upload an NDA and receive clause\-by\-clause analysis powered by accumulated firm judgment— knowledge that previously walked out the door with a retiring executive, now accessible firm\-wide\. **Shepley Bulfinch** invested roughly five hours of expert time— two hours of preparation, three hours of recording— to document specialized knowledge that AI then transformed into written summaries and guides estimated to represent 200 hours of traditional documentation work[10](/blog/blog-the-18-month-ramp-is-negotiable#ref-10)\. **Boulder Associates** ran 30\-minute expert conversations with individual practitioners and converted them into shared knowledge resources for healthcare construction expertise that had previously lived with one person\.[10](/blog/blog-the-18-month-ramp-is-negotiable#ref-10) What all three have in common: - They captured reasoning, not just procedures - The knowledge became accessible firm\-wide, not person\-dependent - The assets that preserve expertise from departing staff can also train the people arriving This is what AI\-assisted training looks like in practice\. And it connects to [how we approach institutional knowledge in AEC](/blog/institutional-knowledge-aec): the infrastructure for knowledge preservation and the infrastructure for ramp acceleration are the same infrastructure— just aimed in two different directions\. ## What You Need in Place Before AI Can Help All three of those firms shared one prerequisite that's easy to overlook: the expertise was already documented— or at least documentable— before AI got involved\. AI training tools work only if there is something for AI to train on\. Deploying AI before you've captured your institutional knowledge doesn't compress the ramp— it gives new hires AI access to nothing\. The firms getting results from AI\-assisted training didn't start with the AI\. They started with the knowledge capture\. And the sequence matters\. Three prerequisites make the difference: 1. **Documented SOPs**— not procedural checklists, but the reasoning behind them\. "Why we handle scope changes this way" is more valuable for training than "what form to submit\." 2. **Captured expert knowledge**— structured conversations with senior staff organized by scenario type \(contract disputes, client conflicts, scope creep decisions\), not just recorded and left as unstructured transcripts\. 3. **A delivery method**— a way for incoming staff to access this knowledge at the moment they need it: a custom AI tool, an embedded query interface, a structured Q&A system within their existing workflow\. This infrastructure\-first argument isn't theoretical\. A professional services consultant with a decade of domain expertise found exactly this in her own practice: building comprehensive SOPs before deploying AI dramatically accelerated her ability to work with the technology\. "If I hadn't done all this work to establish SOPs, AI would have been a lot less useful," she said\. "Having that infrastructure already in place allowed me to move faster\." The principle scales directly\. Firms that have done the documentation work are positioned to convert it into AI\-deployable training assets immediately\. If the institutional knowledge lives only in the heads of your three most senior project managers, AI can't transfer it\. You have to externalize it first\. For principals thinking through whether they're ready for this kind of work, an [AI decision framework for founders](/blog/ai-decision-framework-founders) can help clarify where your firm sits before you commit to the infrastructure\. And [building an AI culture in your firm](/blog/building-ai-culture) covers the change management layer that determines whether these tools get used once they're built\. ## Getting Started— The Minimum Viable Approach The minimum viable entry point is one knowledge area, two to three senior staff members, and three to five hours of structured conversation\. That's enough to build an AI\-accessible training asset that mid\-level staff can query when the senior person isn't available\. 1. **Choose one high\-stakes knowledge area** where junior PMs most often need senior guidance and where the guidance is largely consistent— RFI process, contract language review, scope change documentation\. 2. **Record two to three hours of expert reasoning** with one or two senior staff, structured around realistic scenarios, not abstract principles\. 3. **Build a searchable tool**— at the basic level, a well\-prompted AI assistant with your documentation loaded that staff can ask questions; at a more capable level, a dedicated platform similar to what Knowledge Architecture deployed for MBH Architects and Shepley Bulfinch\. 4. **Deploy alongside real project mentoring**, not instead of it\. The goal is increasing the mentor's bandwidth, not eliminating the mentor\. Start with the question your junior PMs ask most often\. Build one tool that answers it well\. See what happens\. That's a better starting point than a firm\-wide training platform deployment\. Measure as you go: Are junior staff asking fewer entry\-level questions? Are their first drafts better? Is mentoring time shifting toward judgment\-development conversations rather than knowledge\-retrieval? Those are the signals that the knowledge layer is compressing\. For firms evaluating where to start or how to structure the knowledge capture process, [our AI implementation work](/services/ai-implementation) covers exactly this kind of ground— from the infrastructure questions to the tooling decisions\. ## FAQ ### What is "The Missing Middle" in AEC? The Missing Middle refers to the shortage of experienced mid\-career professionals in architecture, engineering, and construction firms— specifically those with 5–12 years of experience who can lead projects and develop junior staff\. Seventy\-six percent of AEC firms currently cite it as a challenge[2](/blog/blog-the-18-month-ramp-is-negotiable#ref-2)\. The gap is caused by accelerated retirements of senior staff and the simultaneous failure to develop a robust mid\-level pipeline, and it compounds as firms continue to reduce entry\-level hiring positions\. ### How long does it take a new hire to become a productive project manager at an AEC firm? Architectural project managers typically require 2–4 years of post\-hire on\-the\-job training to operate independently, with senior PM roles often requiring 5–7 years of total professional experience[1](/blog/blog-the-18-month-ramp-is-negotiable#ref-1)\. This timeline reflects the time needed to build independent judgment on real projects— not just skill acquisition— which is why it resists standard training interventions\. The knowledge layer can be compressed; the judgment layer still requires supervised time on real work\. ### Can AI actually reduce employee training time? For the knowledge and skills layer, yes\. AI tutoring tools increase knowledge retention by 25–40%, and skills\-based development programs reduce time\-to\-proficiency by 30–50%, according to LinkedIn's 2025 Workplace Learning Report[6](/blog/blog-the-18-month-ramp-is-negotiable#ref-6)\. But AI does not compress judgment development— the pattern recognition built through supervised experience on real projects\. AI shortens part of the ramp, not all of it\. ### What do AEC firms need before AI training tools will work? Captured institutional knowledge: documented SOPs that explain the reasoning behind firm decisions, structured expert conversations organized by scenario type, and a delivery method that makes that knowledge accessible at the moment a junior professional needs it\. AI training tools work only when there is something for AI to train on\. Without documented expertise, deploying AI accelerates nothing\. ## The Window Is Open Now The 18\-month ramp isn't a law of nature\. It's a product of how slowly firms have transferred knowledge from senior professionals to the people behind them— and that's exactly what AI is designed to change\. The firms moving fastest on this aren't deploying AI training platforms and hoping for the best\. They're capturing reasoning first\. They're making explicit what used to be implicit, and then making it accessible to anyone who needs it, at the moment they need it\. The AEC firms already doing this— MBH Architects, Shepley Bulfinch, Boulder Associates— didn't start with AI\. They started with the insight that the knowledge had to get out of people's heads before it could go anywhere useful\. AI extends what each mentor can reach\. It doesn't replace what a mentor is\. If your firm is sitting on decades of institutional knowledge from senior project managers approaching retirement, you're also sitting on a training infrastructure you haven't yet built\. The window to capture that knowledge is open now— and once it's captured, it's deployable immediately\. ## References 1. Enginuity Talent Group, "How to Become an Architectural Project Manager" \(2025\)— [https://www\.enginuityadvantage\.com/how\-to\-become\-an\-architectural\-project\-manager/](https://www.enginuityadvantage.com/how-to-become-an-architectural-project-manager/) 2. Stambaugh Ness, "What's Ahead for AEC Firms in 2026: Top Emerging Challenges" \(2025\)— [https://www\.stambaughness\.com/blog/top\-aec\-challenges\-2026/](https://www.stambaughness.com/blog/top-aec-challenges-2026/) 3. Openquire, "Recruiting and Retention in the AEC Industry: Navigating the Pain Points" \(2025\)— [https://openquire\.com/recruiting\-and\-retention\-in\-the\-aec\-industry\-navigating\-the\-pain\-points/](https://openquire.com/recruiting-and-retention-in-the-aec-industry-navigating-the-pain-points/) 4. Certis AEC Solutions, "AEC Industry Challenges in 2026: Workforce, Skills, and the Path Forward" \(2026\)— [https://www\.certisaec\.com/aec\-industry\-challenges\-in\-2026\-workforce\-skills\-and\-the\-path\-forward/](https://www.certisaec.com/aec-industry-challenges-in-2026-workforce-skills-and-the-path-forward/) 5. BDC Network / AIA, "AEC Industry at Risk of Losing Its Next Generation Leaders Without Better Mentoring" \(2024\)— [https://bdcnetwork\.com/aec\-industry\-risk\-losing\-its\-next\-generation\-leaders\-without\-better\-mentoring](https://bdcnetwork.com/aec-industry-risk-losing-its-next-generation-leaders-without-better-mentoring) 6. LinkedIn, "2025 Workplace Learning Report" \(2025\)— [https://business\.linkedin\.com/learn/resources/workplace\-learning\-report](https://business.linkedin.com/learn/resources/workplace-learning-report) 7. GetDX, "Developer Ramp\-Up Time Continues to Accelerate with AI" \(2026\)— [https://newsletter\.getdx\.com/p/developer\-ramp\-up\-time\-continues](https://newsletter.getdx.com/p/developer-ramp-up-time-continues) 8. Thomson Reuters Institute, "How AI Simulation Could Reshape Legal Training and Education" \(2025\)— [https://www\.thomsonreuters\.com/en\-us/posts/legal/ai\-simulation\-legal\-training/](https://www.thomsonreuters.com/en-us/posts/legal/ai-simulation-legal-training/) 9. McKinsey & Company, "Building Expertise in the Age of AI: Who Trains the Next Generation?" \(2025\)— [https://www\.mckinsey\.com/capabilities/people\-and\-organizational\-performance/our\-insights/building\-expertise\-in\-the\-age\-of\-ai\-who\-trains\-the\-next\-generation](https://www.mckinsey.com/capabilities/people-and-organizational-performance/our-insights/building-expertise-in-the-age-of-ai-who-trains-the-next-generation) 10. Knowledge Architecture, "Modern Knowledge Capture: How AI\-Enabled Teams Are Changing the Way AEC Firms Capture Expertise" \(2025\)— [https://www\.knowledge\-architecture\.com/blog/modern\-knowledge\-capture\-how\-ai\-enabled\-teams\-are\-changing\-the\-way\-aec\-firms\-capture\-expertise](https://www.knowledge-architecture.com/blog/modern-knowledge-capture-how-ai-enabled-teams-are-changing-the-way-aec-firms-capture-expertise) --- **How we made this article:** We use AI in our research and writing so our small team can share more of what we learn. We verify the sources and take responsibility for every article we publish. [Read how we use AI.](https://dancumberlandlabs.com/how-we-use-ai/) --- ## About the author **Dan Cumberland** — Founder, Dan Cumberland Labs Dan Cumberland helps engineering and construction firms see where they stand with AI and decide what to build first. He created Pacemark, the AI maturity model behind that work, from research on more than 300 companies. - Take the assessment: https://pacemark.ai/signal/assessment/?track=aec&utm_source=dcl-site&utm_medium=link&utm_campaign=pacemark-assessment - Book a call: https://book.dancumberland.com/ai-strategy --- Source: https://dancumberlandlabs.com/blog/the-18-month-ramp-is-negotiable/