The most underestimated skill in estimating: reading between the drawings

Reading between the drawings means spotting functional requirements that are missing from the explicit design so you can anticipate true project cost. It separates basic quantity extraction from real risk mitigation: applying historical construction knowledge to incomplete plans, and accounting for trade interfaces, structural supports, and constructability before a binding commitment.

The job of a preconstruction professional is not merely to calculate a budget. It is to turn uncertainty into binding corporate commitments. When an estimating department reviews a bid package, the critical skill is distinguishing literal extraction of what exists on paper from applied judgment about what will need to exist in the field. Because design documents often reach bidding with significant information missing, relying solely on explicit line items guarantees systemic scope gaps. Advanced estimating requires pattern recognition, historical validation, and the ability to project design intent across 30%, 50%, and 75% complete documents.

For the broader case that estimating is risk management, not cost calculation, start there. This piece focuses on the skill itself: trade-interface gaps, incomplete design, bid leveling, and procurement models that put estimators inside the design process.

What "reading between the drawings" means

Reading between the drawings means identifying scope implied by design intent but physically absent from the current blueprints. The estimator looks beyond explicit line items and applies pattern recognition to determine which missing details, trade overlaps, and structural supports are realistically necessary to construct the facility.

A shallow view of estimating reduces the discipline to a linear equation: analyze drawings, extract quantities, apply unit prices, generate a total. Experienced preconstruction professionals know that picture is incomplete. The dollar value at the bottom of a bid sheet is not the work of estimating. It is the financial output of thousands of smaller qualitative judgments about project risk.

Operating at a high level requires working on two layers at once:

Layer 1 is understanding exactly what is explicitly present in the bid package: specifications, subcontractor proposals, addenda changes, and the dimensions the architect provided. It answers: what do the documents say?

Layer 2 is where expertise and judgment reside. It asks what the project actually requires to be built successfully, regardless of what is currently drawn: implied design intent, historical precedent, and constructability.

A junior estimator may look at a 50% Construction Document (CD) set, correctly note that a structural detail is not shown, and assume it need not be priced. An experienced estimator looks at the same set and recognizes that while the detail is not drawn, there is no realistic way to construct the specified assembly without it, so the cost must be carried. That pattern recognition takes years: noticing a suspiciously low subcontractor bid because an implied requirement was missed, or realizing two overlapping trades are each assuming the other owns a critical interface. Related playbooks: spotting scope gaps before you carry and reviewing exclusions and qualifications.

Incomplete documents at 30%, 50%, and 75% design

Incomplete documents force estimators to price quantified risk rather than deterministic material counts. At 30%, 50%, and 75% design completion, frameworks such as the AACE Cost Estimate Classification System help apply appropriate contingencies, bridging schematic representations and the assemblies that will eventually be required in the field.

AACE International Recommended Practices 18R-97 and 56R-08 map projects from concept to firm contractual price, classifying estimates from Class 5 (least defined) to Class 1 (most defined). The core idea: estimate quality and accuracy are driven by the maturity of engineering deliverables and project definition, not by the sophistication of the software used.

Ranges summarized from AACE International Cost Estimate Classification matrices for building and general construction.

The hard part for precon teams is that they are often asked for binding commitments, such as a Guaranteed Maximum Price (GMP) under Construction Manager at Risk (CMAR), when documents are only at Class 3 or Class 2. At 30% to 75% design maturity, major architectural concepts and massing are established, but important micro-details remain absent.

At Class 3, mechanical systems may appear as single-line diagrams, and structural drawings may lack final connection designs or secondary framing. The estimator must read design intent and carry cost for inevitable evolution. When fire-stopping, waterproofing terminations, or secondary steel supports are missing from a 50% CD set, those costs do not vanish. They surface in execution as margin-eroding change orders.

Teams often use a Project Definition Rating Index (PDRI), a weighted checklist that scores scope-document completeness against industry standards. A favorable (low) PDRI score supports a tighter, more deterministic budget; a weak score signals that scope extrapolation and stochastic modeling are required. Mentally completing the missing 25% to 50% of the drawings from historical baselines and industry norms is a defining characteristic of a profitable estimating department. Delivery-method context: hard bid vs CMAR vs design-build.

Risk mitigation, not budget calculation

Estimating is about risk mitigation because a construction budget is the financial expression of quantified uncertainty. Rather than merely tabulating quantities, estimators evaluate historical data, trade overlaps, and design maturity to determine whether the firm can safely commit corporate resources to a binding contract.

Directors of estimating who operate at a high level treat estimating as a forward-looking risk discipline, not an accounting function. That distinction shows up clearly when analyzing win rates. A common structural flaw is the lack of a closed loop connecting precon assumptions to field performance. In many firms, actual productivity, true material escalation, and post-mortem findings never feed back into the estimating database. See the estimate-to-actual gap and why estimates start from zero.

Without verified historical actuals, estimators rely on anecdotal memory. If a project manager criticized an estimator because a specialized concrete pour took twice as long as budgeted, that estimator develops scar tissue: inflated unit rates and hidden cushions on the next bid. That makes one estimate safer to execute while systematically destroying competitiveness. Compounded conservatism across dozens of trade packages produces bids disconnected from market reality and stagnant win rates.

True risk mitigation is the opposite of blind conservatism. It requires analytical specificity. Rather than adding a blanket 15% contingency because drawings are vague, a high-functioning team identifies which elements of design intent carry which risk profiles. They differentiate:

• Design contingency: funds reserved for incomplete architectural and engineering details that will be added as drawings mature
• Construction contingency: funds reserved for unforeseen field conditions, weather, and execution risk

By isolating where the risk lives, the team can price known, deterministic quantities aggressively while keeping defensible buffers strictly where documents remain incomplete. Carry methodology: carry number methodology for construction bids.

Common scope gaps at trade interfaces

The most common precon scope gaps appear at trade interfaces where responsibility is poorly defined. Frequent omissions include parapet roof coping and waterproofing transitions, structural steel fabrication embeds, and in-wall fire-retardant plywood backing for Division 10 restroom accessories.

Gaps rarely occur in primary building elements. Estimators seldom forget main structural columns or exterior drywall square footage. Gaps hide in the seams, where specialty trades meet, or where design intent requires supporting infrastructure that never appears on the finish schedule.

Parapet roof design and waterproofing transitions

Parapets are a notorious source of catastrophic scope gaps. Architecturally, a parapet hides rooftop equipment and frames the silhouette. From a construction standpoint, it is a high-risk liability because waterproofing requires coordination across multiple independent trades.

Failure rarely starts at the visible metal coping. It usually starts at the hidden wall-to-membrane junction. Proper waterproofing typically requires roof membrane base flashing to extend up the parapet face (often eight to twelve inches), overlapped by counterflashing mechanically attached to the wall. Between those layers sits a vulnerable transition that depends on sealant integrity.

Estimators frequently miss the nuanced costs of that transition. The roofing sub may assume the mason installs through-wall flashing; the mason may assume the roofer owns all moisture barriers and sealants. Coping thermal movement matters too: metal expands and contracts under UV exposure, and attachment methods that do not allow cycling fail joints and admit wind-driven rain. Without explicitly assigning wood blocking, sheet metal counterflashing, sealants, and coping to specific subcontractors during bid leveling, the GC absorbs the bridge components.

Structural steel embeds and design intent

Another Layer 2 problem is translating structural steel from preliminary engineering to shop fabrication. Structural engineers often provide LOD 300 drawings: design intent, primary member sizing, load paths. Steel cannot be fabricated or erected from LOD 300 alone. The project needs LOD 400 fabrication detail: gusset plates, weld clearances, bolt-hole patterns.

Massive gaps appear when estimators fail to account for secondary steel and cast-in-place embeds required by broader design intent. A heavy unitized curtain wall may be clear on architectural drawings while structural drawings lack embeds at the slab edge to support facade loads. Extracting only primary I-beam tonnage from the engineering schedule underbids the package. The estimator must read facade design intent, understand cantilevered load requirements, and clarify (via RFI or scope clarification) who supplies and installs the embeds. The gap between raw uncoordinated tonnage and fully fabricated, erected steel with connections and coordination is frequently a multi-million-dollar variable on large commercial work.

Division 10 restroom accessories and in-wall backing

Interior specialties, especially Division 10 restroom accessories, are a classic source of schedule delay and margin fade. A restroom accessory schedule may list ADA grab bars, recessed towel dispensers, mirrors, and folding shower seats. The inexperienced estimator counts the items, solicits a lump-sum specialty quote, and marks the package complete.

The experienced estimator knows heavy accessories cannot mount to standard gypsum alone. They need in-wall backing (typically fire-retardant treated wood blocking or welded steel plates) installed before drywall is hung. Recessed accessories need framing cavity depth that does not conflict with plumbing; powered accessories need dedicated electrical rough-in.

Without a scope matrix that assigns backing to rough framing, electrical rough-in to the electrician, and final install to specialties, the schedule fails late: discovering omitted backing after tape, mud, and paint means destructive rework, out-of-sequence work, delayed inspections, and contentious change orders. The design intent was clear (a functional, load-bearing grab bar). The precursor construction steps were missing from the drawings.

How bid leveling exposes design intent

Bid leveling exposes hidden gaps by normalizing disparate subcontractor proposals to reveal conflicting assumptions about design intent. When specialty contractors submit widely varying prices for the same package, that spread often highlights undocumented interfaces, missed exclusions, or ambiguous specifications that need clarification.

During buyout, competing subs exclude different peripherals, qualify pricing on different labor assumptions, and interpret incomplete drawings in ways that favor their operations. Bid leveling is the process of normalizing those proposals for an apples-to-apples comparison. It is not merely administrative. It is one of the strongest risk tools available to precon.

Example: three electrical bids at $1.2M, $1.25M, and $1.15M suggest the market understands the package. Bids at $1.2M, $800k, and $1.4M signal severe ambiguity. The $800k bidder has almost certainly excluded a major component (integrated fire alarm, low-voltage data, heavy switchgear) because they assumed it belonged elsewhere.

Effective leveling maintains two distinct records:

1. An internal leveling sheet capturing every raw exclusion, assumption, and cost delta
2. A clean scope of work that will attach to the binding subcontract

A seasoned estimator also cross-pollinates knowledge across the bidding pool. If one observant sub excludes "temporary shoring required for structural sequencing," that shoring may be true design intent even if the engineer never drew it. The estimator then requires every other bidder to acknowledge and price it, closing the gap before contract. See also bid spread reporting.

How CMAR and ECI change early contractor involvement

Procurement models such as Construction Manager at Risk (CMAR) and Early Contractor Involvement (ECI) shift the GC from passive bidder to active design consultant. Estimators provide continuous constructability feedback, open-book cost modeling, and trade coordination while drawings are still developing.

Traditional Design-Bid-Build (DBB) separates design from construction and withholds constructability feedback until drawings are complete. That often produces adversarial RFIs and mid-project redesign when budgets overrun. Owners increasingly use alternative models that put the estimating team inside the design process.

Under CMAR, the construction manager engages early for iterative cost modeling, schedule development, and constructability reviews. As the architect moves from schematic to Design Development, the estimating team continuously reads between evolving drawings to project final cost implications of site constraints, materials, and structural systems.

ECI similarly lets the builder advise on methodology, compliance pathways, and value engineering before construction. Published comparisons often cite meaningful total-project savings versus traditional lump-sum procurement when that involvement is real. In both environments, identifying missing implied scope matters most when advisory modeling becomes a binding GMP: the estimator locks price on incomplete documents. Fail to carry implicit design intent, and overruns come out of CM contingency or corporate profit. Workflow detail: hard bid vs CMAR vs design-build estimating workflows.

Why traditional AI tools struggle here

Traditional AI tools often fail in precon when they operate as passive calculation engines rather than contextual workflow systems. Early LLM deployments struggle with precise geometric calculations and lack the ability to cross-reference fragmented project changes across hundreds of dynamic quotes and addenda.

The industry's first wave of AI adoption was shaped by a misunderstanding of where the technology helps. Early adopters tried generative models and generic computer vision as automated takeoff engines, for example volumetric rebar calculations across complex structural members that produced divergent, unreliable answers. When identical prompts yield different mathematical outputs, organizational trust collapses.

That failure mode is structural: large language models are built for probabilistic semantic extraction and pattern recognition, not absolute spatial geometry or deterministic mathematics. The useful value of AI in precon is not replacing linear counting. It is resolving the coordination problem created by fragmented, constantly evolving project data.

Preconstruction is dynamic. Teams receive a 50% design set, then Addendum 1, an RFI answer on substitutions, a revised quote altering exclusions, and a VE directive reducing scope. The human limiter is rarely the inability to read one specification. It is the inability to instantly cross-reference how a minor architectural change in Addendum 2 impacts a quote received three weeks earlier, which alters a schedule assumption buried in an internal leveling sheet. Chatbots and point solutions lack the persistent memory and systemic awareness required to map those relationships. Addenda discipline: construction addenda management for GC estimators.

What a proactive precon system changes

A proactive preconstruction system automatically parses document changes, tracks cascading impacts of addenda across trade packages, and drives bidding workflow forward without waiting for manual routing. It turns passive data storage into actionable risk mitigation.

The useful shift is from a passive system of record to an active system of action. The core tension in modern estimating is extreme analytical specificity under immovable bid deadlines. Under pressure, even experienced teams skim addenda, rely on unverified memory for inclusions, or apply blanket contingencies instead of investigating granular risks.

When a new addendum arrives with a revised layout, generic software stores a PDF and notifies that a file uploaded. A system of action parses content and surfaces downstream consequences: which trades are affected, which leveled quotes are now stale, which internal scope sheets need immediate review. Chaos becomes a structured, prioritized task list.

By reducing manual document comparison, spreadsheet formatting, and hunting for buried exclusions, the system acts as a second set of eyes on Layer 1 (what is explicitly stated across active documents) so humans can spend judgment on Layer 2: what is missing. That is Piper's role in this workflow: amplify Layer 1 completeness so estimators can read between the drawings under deadline pressure.

Standardizing tribal knowledge

Teams standardize tribal knowledge by embedding historical lessons, company-specific checklists, and senior insights directly into the digital precon workflow, so junior staff benefit from decades of pattern recognition at the moment of decision.

Departments routinely span junior staff pricing a first commercial project and directors with thirty-five years of field experience. Historically, the knowledge that a specific sub always excludes cold-weather protection, or that a soil condition requires specialized shoring, lived in the veteran's head. When seniors retire or leave, that memory disappears.

Storing knowledge in a passive searchable database is not enough. If a junior estimator does not know to search for a risk, they will never find it. Knowledge must appear contextually when relevant: while leveling masonry, surface historical cost data, remind the user to verify parapet coping attachment after a prior failure, and flag a bidder who historically excludes scaffolding.

Making the best estimator's judgment a persistent part of how the whole team works raises baseline excellence. Advanced systems do not replace the human estimator. They synthesize baseline facts so professionals can interrogate design intent and build a risk-adjusted budget that protects margin while remaining competitive. Final gate: final bid review QA/QC.

Sources

• AACE International, Recommended Practice 18R-97, Cost Estimate Classification System
• AACE International, Recommended Practice 56R-08, Cost Estimate Classification System