Architecture Firm Project Portfolio: Rebuilding the Master List After a Leadership Transition

When the Master List Walks Out the Door

In most architecture firms, the master project list isn't a file. It's a person. And when that person retires, takes a buyout, or leaves at an ownership transition, the list dies with them— usually about three years before anyone notices.

You know the partner. The one who knew every photographer credit on every published project, every consultant relationship that survived the bad GC, every fee on every pursuit going back fifteen years, every quiet "we don't work with that developer again." When they walked out the door, they didn't take a binder. They took the firm's memory.

What gets lost first isn't the hard layer. Project metadata— names, dates, square footage— usually lives somewhere recoverable, even if scattered. What goes first is the soft layer: the lessons learned, the reference relationships, the political read on which past clients you can call. Unanet's succession-planning research¹ frames it directly— a senior engineer's retirement creates project-context loss and severs developer and MEP-consultant relationships that took decades to build.

"Most firms discover their master list is gone the week they need to put together a qualifications package for a pursuit they should have already won."

This isn't a CRM-selection article. It's a question of whether your firm still knows what it has done — and a practical sequence (schema → owner → tooling → AI-assisted rebuild) for principals who've realized the answer is no. The AIA frames knowledge management as an ongoing concern across the firm life cycle²— not a project, not a software purchase, an ongoing concern.

Once you've seen this happen once, the pattern is unmistakable. Here's why it happens by default.

Why the Best Master Architecture Records Fail at Leadership Transitions

Master project records most often fail at leadership transitions because the record is tacit, not documented— concentrated in one or two senior people. When those people leave, capture stops, and the institutional memory begins to decay from day one.

According to Gartner research³, roughly 42% of institutional knowledge resides solely with individual employees, meaning the firm cannot do nearly half of what it used to when those people leave. In a $20M–$100M practice, that 42% sits inside three or four heads. Sometimes one.

Three mechanisms make this the default outcome:

1. Tacit-knowledge concentration. The principal who built the practice owns the project history politically, not just operationally. Asking a junior PM to "update the master list" is asking them to formalize someone else's authority. It rarely sticks.
2. Ownership transitions surface what was never documented. Building Design+Construction and Architectural Record both treat ownership transition as a foreseeable event, but most firms only audit their record once a transition is already underway— too late.
3. The work itself suppresses capture. MDPI research on construction project-based organizations⁴ identifies project features, knowledge withholding, fear, and team-interaction patterns as active drivers of knowledge loss. This is a polite way of saying: the people who know things often have reasons not to share them.

"Software is not a system. Most firms with Deltek Vantagepoint or Unanet CRM still have a decaying master list, because no one inside the firm owns the discipline of keeping it current."

The Gartner number is the headline. The MDPI finding is the texture. Together they describe a firm where the record exists right up until the moment you need it. Once you see the pattern, you can't unsee it— and that's the opening.

If we're going to preserve the record, we have to first agree on what it is.

What a Durable Master Project Record Actually Contains

A durable master project record is a single, queryable inventory of every project the firm has delivered— capturing not just the names and dates, but the relationships, fees, photographer credits, lessons learned, and reference contacts that actually win pursuits. It is the firm's source of truth for every SOQ response, every RFP shortlist, every go/no-go conversation.

If your master project list can't produce a defensible qualifications package without three weeks of forensic email archaeology, you don't have a master list. You have a stack of files.

A minimum-viable master project record captures, at minimum, these twelve fields per project:

• Project name
• Client (legal entity + contact)
• Firm role (architect of record, design architect, EOR, associate)
• Dates (start, substantial completion, occupancy)
• Fee (contract value + final billed)
• Square footage and program type
• Building type / market sector
• General contractor
• Consultant team (MEP, structural, civil, landscape)
• Photographer credit and image rights status
• Lessons learned (one paragraph, written within 30 days of closeout)
• Reference contact name + permission status

This is different from a portfolio (which is selective and externally curated) and different from a CRM contact list (which tracks relationships, not project deliverables). Stambaugh Ness⁵ describes how Vantagepoint integrates CRM, project management, resource planning, time and expense, and accounting against a single project record— and that integration is the point. But it only works if the schema is decided first.

Pick the schema before the software. Monograph's succession framework⁶ makes the same case in different words: start early, document systematically. The discipline is the asset. The tooling is the substrate.

With the schema clear, the natural next question is which system carries it.

Why AEC ERP/CRM Software Alone Doesn't Save the List

The dominant AEC-purpose-built systems— Deltek Vantagepoint and Unanet CRM (formerly Cosential)— are designed around exactly this problem. But firms routinely defer history loading, and even firms that load it watch the record decay because no one owns the discipline of keeping it current.

Deltek's own help documentation⁷ acknowledges the weak point: project history can be entered at install, after install, or "not at all." Many firms quietly choose "not at all."

"Software is not a system."

What Vantagepoint and Unanet do well is consolidate the record into one place. What they cannot do is decide who, inside your practice, is accountable for keeping that record true. That decision is a cultural one, not a procurement one.

If software isn't the answer, what is? It starts with naming an owner.

The Cultural Question No Vendor Can Answer for You

Before you choose software, choose an owner. The master project list survives leadership transitions only when one named role inside the firm is accountable for keeping it current— and when contributing to it is treated as part of doing the work, not extra to it.

"Every firm with a healthy master list has a person whose job description includes 'the list is current.' Every firm with a dead list does not."

"Everyone is responsible" reliably means no one is. In $20M–$100M practices, three role configurations actually work:

• Practice manager / COO owns the list; PMs feed it at project closeout.
• Marketing director with PM authority owns the list because the qualifications package is their deliverable anyway.
• Dedicated knowledge manager (more common above $50M) owns the list as their primary remit.

Which configuration fits depends on the firm. What doesn't work is leaving it to whichever principal happens to remember. That's how the list dies the first time. And building an AI culture inside the firm starts here— at the cultural ownership question, not at the tool selection.

The replacement-cost honesty matters too. Gallup research³ puts the cost of replacing one employee at one half to two times their annual salary— and that math gets ugly when "the employee" is the de facto keeper of the firm's project memory. You're not just hiring a replacement. You're rebuilding institutional memory.

Make capture part of project closeout, not an afterthought. A 30-minute lessons-learned note, dictated and transcribed, is cheap insurance. Once you have an owner and a schema, the question becomes how to rebuild what you've already lost.

Where AI Actually Helps (and Where It Can't)

Generative AI can compress the rebuild of a decayed master project list by hours-to-minutes, by summarizing existing emails, proposals, project files, and meeting transcripts into structured project records. What it cannot do is recover knowledge that was never written down.

"AI is an accelerant on documentation that already exists. If a project's lessons-learned conversation lives only in a retired partner's memory, no model can retrieve it."

Native AEC tooling is already heading here. Deltek's Client Smart Summary⁸ uses generative AI to summarize client and project information for executives, BD managers, and PMs. The same approach works outside the stack — on email archives, SharePoint folders, and Teams transcripts a firm already owns.

What AI can recover from existing artifacts:

• Project metadata buried in old proposals
• Client and consultant relationships visible in email patterns
• Approximate fee and schedule data from contract files
• Reference contact lists from past correspondence

What AI cannot recover:

• The "we don't work with that GC again" conversation that was never written down
• The political read on a developer client
• The unwritten reason a project went sideways
• Anything that lived only in a retired principal's head

The implication is direct: sit with retiring partners now and capture the tacit layer to artifacts that AI can later index. Panopto research⁹ suggests a firm of 1,000 employees loses approximately $2.4 million in productivity annually to knowledge-loss inefficiencies. Scale that down to a 60-person practice and the number is still serious money— and the rebuild is cheaper than the loss. An honest read of AI consultant vs. in-house capacity helps decide who runs the capture.

Knowing where AI fits, here's the principal's first move.

The Principal's First Move — A 90-Minute Inventory

The first move is not buying software. It's a 90-minute exercise: gather the people who actually know what the firm has done— typically two or three principals and the longest-tenured project manager— and walk a single representative pursuit backward through every record it touches.

"If your firm can't reconstruct the data trail of a single won pursuit in 90 minutes, the master list is already past the point where memory will save it."

Run the exercise like this:

1. Pick one recent won pursuit. Ideally something within the last 24 months that's still fresh in memory.
2. Walk every artifact backward. Proposal, kickoff deck, signed contract, photos, references, lessons learned. Note where each lives.
3. Mark every gap. Where is there no artifact at all? Where is the artifact incomplete? Where does the only copy live in someone's inbox?
4. Build a one-page gap map. What we have, where it is, what's missing.
5. Sequence the rebuild. Schema decision (Section 3) → ownership decision (Section 5) → tooling decision (Section 4) → AI-assisted rebuild (Section 6).

The output is the gap map. The gap map drives every subsequent decision— and it's built by the people in the room, not by a vendor. Monograph's framework⁶ holds here: start early, document systematically. The 90-minute exercise is the start.

If the gap map reveals more than your team can rebuild internally, an implementation partner can help sequence the right schema, owner, and tooling for your specific practice. Dan Cumberland Labs works with founder-led AEC firms on exactly these decisions — bringing an AI decision framework for founders and, where it fits, a fractional AI officer engagement that lets you move without hiring. When the rebuild is bigger, our AI Strategy Services turn the schema-owner-tooling-AI work into a defined plan you own at the end.

The list dies quietly. Rebuilding it does not have to.

Frequently Asked Questions

What is a master project list in an architecture firm?

A master project list is a canonical record of every project the firm has delivered or is currently delivering, capturing client, role, fee, team, photographer credit, lessons learned, and reference contact⁵. It is the source asset behind every qualifications package, SOQ response, and pursuit decision— the firm's source of truth for what it has done and who it has done it with.

Why do architecture firms lose their master project list?

Because the list is held tacitly by one or two senior people rather than systematically captured. When those people leave through retirement, ownership transition, or departure, capture stops and the record decays from day one. Gartner³ estimates 42% of institutional knowledge resides solely with individual employees, and MDPI research⁴ confirms project-based organizations actively suppress knowledge sharing through team dynamics and fear.

What is the best system for tracking AEC project history?

Deltek Vantagepoint and Unanet CRM are the dominant AEC-purpose-built options that combine project history with CRM and financials⁵⁷. The right choice depends on firm size, existing financial systems, and— most importantly— whether someone inside the firm owns the discipline of keeping the record current. Software without an internal owner produces a decaying record, regardless of which platform you buy.

How much knowledge do firms lose when key staff leave?

Gartner³ estimates roughly 42% of institutional knowledge resides solely with individual employees. Panopto research⁹ suggests a firm of 1,000 employees loses approximately $2.4 million in productivity annually to knowledge-loss inefficiencies. In a smaller AEC practice, the absolute dollar figure is smaller but the proportional impact is often larger, because each principal carries more of the firm's institutional memory.

Can AI rebuild a decayed master project list?

AI can accelerate the rebuild by summarizing existing emails, proposals, and project files— Deltek's Client Smart Summary⁸ is one example built directly into the AEC stack. It cannot recover knowledge that was never written down, which is why capturing tacit knowledge from retiring principals before they leave remains essential. AI is an accelerant on documentation; it is not a substitute for documentation.

References

1. Unanet, "Why Succession Planning Is a Must for Any AEC Firm — and How to Do It Right" (2024) — https://unanet.com/blog/why-succession-planning-is-a-must-for-any-aec-firm-and-how-to-do-it-right
2. American Institute of Architects, "Pathways to Firm Ownership & Leadership" (2024) — https://www.aia.org/resource-center/pathways-firm-ownership-leadership
3. Gartner, "Institutional Knowledge: How to Safeguard It When Employees Leave" (2022) — https://www.gartner.com/en/articles/how-to-safeguard-institutional-knowledge-in-the-face-of-the-great-resignation
4. MDPI / Sustainability, "Knowledge Loss in Construction Project-Based Organizations: The Role of Project Features, Knowledge Withholding, Fear, and Teams Interaction" (2025) — https://www.mdpi.com/2071-1050/17/21/9880
5. Stambaugh Ness, "7 Reasons AEC Firms Choose Deltek Vantagepoint CRM for Growth" (2024) — https://www.stambaughness.com/blog/7-reasons-aec-firms-choose-deltek-vantagepoint-crm-growth/
6. Monograph, "Engineering Firm Succession Planning: A 4-Step Framework" (2024) — https://monograph.com/blog/engineering-firm-succession-planning
7. Deltek, "Vantagepoint Project History (Help Documentation)" (2024) — https://help.deltek.com/product/Vantagepoint/4.5/hc_project_history.html
8. Deltek, "Transform Vantagepoint Client Information Into Intelligence" (2024) — https://www.deltek.com/en/blog/vantagepoint-client-smart-summary
9. Panopto Workplace Knowledge and Productivity Report, cited via Iterators, "Cost of Organizational Knowledge Loss and Countermeasures" (2024) — https://www.iteratorshq.com/blog/cost-of-organizational-knowledge-loss-and-countermeasures/