The inputs are scattered.
A drawing holds geometry. A specification holds requirements. A spreadsheet holds assumptions. Someone has to connect them.
THE INVESTMENT THESIS / VERTICAL AI
Engineering capacity.
Without another handoff.
Shibi is a CAD-aware AI engineering coworker: connecting drawings, documents and tools to help teams carry out project work.
Already piloted across sprinklers, plumbing, ductwork and electrical tasks. The sprinkler toolkit is being completed, with broader task coverage guided by evidence from real workflows.
Early pilots · Engineer-reviewed outputs · Existing workflows first
SHIBI / ENGINEERING WORKFLOWCONCEPT ILLUSTRATION
ONE PROJECT, CONNECTED CONTEXTFrom drawing to deliverable.
Engineering judgment stays human.
Inspect inputs, assumptions and outputs before approval.
THE CUSTOMEREngineering teams & specialist contractors
THE PRODUCTA CAD-aware AI engineering coworker
THE PILOTSSprinklers, plumbing, ductwork & electrical
01 / THE WORK THAT GETS IN THE WAY
The project is digital.
The handoffs are still manual.
Engineers move between AutoCAD, Excel, specifications and calculation tools. Each revision can mean finding the right inputs again, repeating calculations and rebuilding a proposal.
The work repeats.
Calculations and proposal preparation consume attention even when much of the underlying process is familiar.
The answer needs a trail.
Engineering teams need to inspect how a result was produced. A plausible response alone is not a usable deliverable.
The buyer’s reason to care: more capacity for project work, with review kept inside the engineering team. That is the value proposition pilots must prove.
02 / A COWORKER THAT DOES THE WORK
From project inputs
to something worth reviewing.
Adoption starts with the files and tasks a team already has. The goal is to reduce repetitive execution, without requiring a wholesale move away from its CAD and spreadsheet workflow.
- 01
Understand the project
Bring drawings, specifications and schedules into a shared project context. Surface missing inputs before they become downstream rework.
Drawing context + explicit assumptions - 02
Do the engineering work
Use CAD-aware AI to coordinate the task and deterministic tools for hydraulic calculations. Prepare proposal material from the project’s inputs.
Calculations + draft deliverables - 03
Make review practical
Keep the inputs and assumptions alongside the result. The engineer checks the work, resolves exceptions and owns the final decision.
Reviewable work + human approval
Workflow illustration. Project suitability, supported inputs and engineering checks are established during a pilot. Professional review remains required.
03 / THE TECHNICAL THESIS
AI understands the task.
Engineering tools do the math.
The product advantage comes from connecting project context, task execution and inspectable results. Each layer has a different job.
CAD + project documents
Connect the task to the drawing, specification and assumptions that define the project.
AI + deterministic calculations
Use AI to interpret and coordinate. Use dedicated engineering tools for numerical work, rather than asking a language model to invent the answer.
Evidence + engineering judgment
Make the basis of the work inspectable. Engineers resolve exceptions and approve outputs; Shibi does not replace professional responsibility.
WHAT CAN COMPOUND
Deeper workflow integration, broader tested task coverage and a stronger evaluation process. Defensibility must be earned in repeated use, not asserted with a model label.
04 / ONE COWORKER, MULTIPLE DISCIPLINES
Connect the project.
Deepen the tools.
The product direction is broad engineering assistance grounded in CAD and project documents. Pilots across mechanical, electrical and plumbing work help establish which tasks are useful, repeatable and ready for deeper tooling.
Across MEP workflows
Sprinkler, plumbing, ductwork and electrical tasks have informed the shared coworker approach.
Fire sprinklers
The sprinkler toolkit is being completed, including calculation and proposal workflows with engineer review.
Deeper task coverage
Broaden task coverage only as repeat usage, output quality and customer demand justify it.
Pilot snapshot: 12 engineers across 3 firms using free pilots, week ending October 20, 2026. This dated snapshot does not establish current weekly active usage or paying customers. Multi-discipline pilots do not establish complete, validated toolkits.
MARKET CONTEXT / PUBLISHED ESTIMATE
A broader software market
around the built environment.
Autodesk's October 7, 2025 Investor Day presentation estimates a $51B FY29 application TAM across architecture, engineering, construction and operations (AECO): $20B Design, $11B Make and $20B Operate. Design is included in the total.
This is Autodesk's published estimate for the broader AECO lifecycle, based on Cambashi and rounded. It is context for the opportunity, not an independently established Shibi TAM, current revenue or guaranteed addressable spend. Shibi's reachable market still depends on supported workflows, adoption and willingness to pay.
Autodesk Investor Day 2025, pp. 5–6BUSINESS MODEL HYPOTHESIS
Land on a useful task.
Grow with the team.
The commercial hypothesis is recurring software revenue from engineering teams and specialist contractors, with expansion through additional users and workflows. Pricing and packaging should follow demonstrated customer value.
Founder-led pilots are the starting point. Repeat usage and willingness to pay come before a scaled sales motion.
05 / WHY NOW
The missing connection
is becoming buildable.
AI makes unstructured project documents more usable as software inputs. Deterministic engineering tools provide the calculation layer. Bringing them together creates a path from assistance to execution, in the workflow engineers already know.
WHY THIS TEAM
Technical depth.
Close to the work.
Technical co-founder
CTO, Shibi · NYU Mechanical Engineering PhDBrings technical depth and hands-on product development to the engineering workflow.
Batuhan Turanlı
CEO, Shibi · CCO, Norm TeknikBrings fire-protection industry experience and a commercial perspective on how engineering teams work.
06 / THE NEXT PROOF POINTS
Build the business
one verified milestone at a time.
Early pilots are a starting point, not a finished traction story. These are the questions that should govern product investment and growth.
- 01
Prove repeatable workflow value
Measure completed tasks, review effort and repeat usage against the team’s existing process.
- 02
Convert value into paid adoption
Validate willingness to pay, onboarding effort and retention before scaling customer acquisition.
- 03
Expand with evidence
Add adjacent tasks when quality, customer demand and delivery economics support the next step.
LONG-TERM DIRECTION
Keep the project connected
as the work changes.
Our ambition is workflow assistance that connects drawings, documents, calculations, revisions and procurement: carrying project context forward as teams move from design decisions to deliverables and purchasing.
This is a future integration direction, not a claim that all these workflows are automated today. Engineers retain responsibility for engineering decisions, safety checks and approvals; procurement commitments remain with the authorized people.
LET’S LOOK AT THE WORK
See the product.
Discuss the evidence.
For investors interested in vertical AI and engineering software: talk with the founders about the product, pilot learnings and the next milestones.