- We offer certified developers to hire.
- We’ve performed 1500+ Web/App/eCommerce projects.
- Our clientele is 1000+.
- Free quotation on your project.
- We sign NDA for the security of your projects.
- Three months warranty on code developed by us.
Commercial cork flooring projects look straightforward from the outside.
Measure the building. Calculate the required flooring. Order the material. Prepare the subfloor. Install the cork. Finish the surface. Inspect the completed work.
In practice, commercial flooring installation is much more complicated.
A project can involve thousands of square feet, multiple rooms, irregular floor plans, different installation zones, material lead times, adhesives and underlayments, moisture testing, expansion requirements, transitions, stairs, furniture movement, multiple trades and strict completion deadlines.
A small estimating mistake can therefore become an expensive operational problem.
Ordering too little cork can create delays and emergency procurement costs. Ordering too much can tie up working capital and increase leftover inventory. Poorly planned cutting can increase offcuts. An inaccurate labor forecast can undermine the project margin. An unexpected subfloor problem can push the installation schedule beyond the customer’s preferred handover date.
This is where AI development for commercial cork flooring installation can create measurable value.
Artificial intelligence does not replace the flooring contractor’s experience. Instead, it can help turn project information into faster estimates, more consistent decisions and better forecasts.
A well-designed AI system can potentially help a flooring company:
The important point is that AI should be treated as a decision-support system, not as an automatic replacement for an experienced estimator or flooring installer.
The best implementation combines AI with accurate measurements, manufacturer requirements, experienced supervision and disciplined project controls.
AI development in this context means building or integrating software that uses machine learning, computer vision, predictive analytics, optimization algorithms or language models to improve parts of the flooring workflow.
It does not necessarily mean creating an enormous custom AI platform from scratch.
A commercial flooring company could start with a relatively focused system.
For example, an AI estimating application could accept:
The system could then generate an initial project estimate.
An advanced version could analyze the geometry of individual rooms and create an optimized material plan.
A still more advanced system could connect estimation, procurement, scheduling, installation progress and project accounting.
That distinction matters because AI development cost depends heavily on the scope of the system.
A company that needs an AI-assisted estimating tool has a very different development budget from a company attempting to build a complete flooring-management platform.
Cork has several characteristics that make accurate planning particularly important.
Commercial cork flooring can be selected for offices, educational environments, hospitality spaces, retail interiors, healthcare-related environments and other projects where designers want a distinctive flooring material with particular performance and aesthetic characteristics.
But the material is only one component of the total installation.
A commercial project may require planning for:
AI estimation should therefore calculate the complete installation requirement, rather than simply multiplying floor area by a product price.
Traditional estimation often relies on spreadsheets, drawings, manual measurements and the estimator’s experience.
These tools remain useful.
The problem arises when project complexity increases.
Consider a commercial building with:
A simple area calculation may be:
Required area = floor area × waste allowance
But this formula does not tell you how efficiently the actual flooring can be laid.
Two projects with exactly the same floor area can generate very different amounts of waste.
Why?
Because waste depends on geometry, room dimensions, installation direction, pattern, plank or tile format, starting points, transitions, unusable remnants and sequencing.
This is one area where AI-based optimization can become useful.
An AI estimation system can divide the estimating process into several layers.
The system analyzes the actual floor area.
It understands the selected cork product’s dimensions, packaging and installation requirements.
It considers the installation pattern, orientation and practical cutting requirements.
It estimates expected offcuts based on geometry and historical data.
It combines material, labor, preparation, logistics and contingency assumptions.
It identifies uncertainties that could cause the estimate to change.
This produces a much more useful result than a single square-foot price.
Material estimation is probably one of the most attractive starting points for AI implementation.
Suppose a commercial project contains several floor zones.
A conventional calculation might estimate each area separately and add them together.
An AI system could go further.
It could identify:
The system can then estimate material requirements according to the selected product.
A simplified calculation could be:
Net flooring requirement = measured floor area
Then:
Gross requirement = net flooring requirement + estimated installation waste
However, an AI system can estimate waste dynamically rather than applying the same percentage to every project.
Applying a blanket waste percentage is easy.
It is also limited.
Imagine two 1,000-square-foot spaces.
The space is almost rectangular.
The space contains:
Both have 1,000 square feet of floor area.
But their cutting efficiency can be very different.
A fixed percentage might underestimate Project B and overestimate Project A.
An AI model can learn from historical projects to estimate waste based on characteristics such as:
The output can then be reviewed by an experienced estimator.
One advanced application is computer vision.
A system can potentially process floor-plan images and identify relevant geometry.
For example, computer vision may help extract:
If the source file contains reliable vector or BIM information, the system can potentially work from structured geometry rather than relying entirely on image recognition.
This distinction is important.
AI should not be trusted to interpret ambiguous drawings without validation.
A better workflow is:
Drawing → AI extraction → estimator verification → approved quantities
rather than:
Drawing → AI → purchase order
The human review step protects the business from errors caused by unclear drawings or incorrect model interpretation.
A material takeoff system can turn floor plans into structured quantities.
A commercial cork flooring takeoff might include:
| Item | Estimated quantity |
| Cork flooring | Project-specific |
| Adhesive | Project-specific |
| Primer | Project-specific |
| Leveling compound | Project-specific |
| Underlayment | If specified |
| Transitions | Project-specific |
| Edge/trim components | Project-specific |
| Finishing products | Project-specific |
The exact quantities must be based on the selected product and manufacturer requirements.
AI should therefore store product specifications as structured data.
One of the most important parts of AI development is often overlooked.
The AI model itself is only part of the system.
You also need reliable data.
For commercial cork flooring, the product database could contain:
When product information changes, the database must be updated.
Otherwise, even an advanced AI model can generate outdated estimates.
Material quantity is only one component of project economics.
A better AI estimating system calculates several cost categories.
Material cost = quantity × unit price
Labor cost = estimated labor hours × loaded labor rate
This can include relevant tools, equipment and rental expenses.
Transportation, handling and delivery costs may also be relevant.
Subfloor preparation can be a major variable.
Project-specific overhead should be considered where appropriate.
Uncertainty should be reflected separately rather than hidden inside arbitrary pricing.
Labor forecasting can be significantly more useful when based on historical data.
Suppose your company has completed 100 comparable flooring projects.
For each project, you could record:
The AI model can identify relationships between these variables.
For example, it might discover that projects with highly fragmented floor plans consistently require more labor per square foot than large open areas.
That insight can improve future estimates.
A simple productivity metric is:
Productivity = installed area ÷ labor hours
But productivity can be affected by many variables.
AI can analyze historical productivity according to:
This can make estimates more realistic.
Waste reduction is one of the strongest business cases for AI.
Material waste costs money twice.
First, you purchase material that does not become part of the finished floor.
Second, someone must handle, store, transport or dispose of the leftover material.
There can also be indirect environmental and administrative costs.
AI can address waste at several stages.
Optimize quantities.
Optimize cutting layouts.
Track actual consumption.
Record reusable leftovers.
Use historical waste data to improve estimates.
Cutting optimization is an interesting technical use case.
Imagine a project requires many pieces with different dimensions.
Instead of cutting material in an arbitrary order, optimization software can search for arrangements that reduce offcuts.
This is related to cutting-stock optimization and related mathematical optimization problems.
The system can attempt to maximize usable material.
The general objective is:
Minimize waste subject to installation requirements.
The algorithm may consider:
This is an area where optimization algorithms may work alongside machine learning.
Not every leftover piece is automatically waste.
A useful AI system can classify remnants.
For example:
Large, undamaged pieces suitable for planned cuts.
Pieces that may work in future projects.
Pieces too small or unsuitable for practical use.
Damaged or unusable material.
This turns leftover material into an inventory-management problem rather than simply a disposal problem.
Historical waste data is extremely valuable.
For each completed project, record:
Then calculate:
Actual waste rate = waste quantity ÷ purchased quantity × 100
Over time, the AI system can learn which project characteristics correlate with higher waste.
This produces a feedback loop:
Estimate → Install → Measure → Learn → Improve next estimate
That feedback loop is more valuable than a one-time AI deployment.
Commercial flooring companies need to understand not just how much material they need but how much capital a project will consume.
AI can help produce budget scenarios.
For example:
Higher preparation requirements and higher waste.
Most likely material and labor requirements.
Efficient installation and favorable conditions.
This approach can help management understand the range of potential outcomes.
Instead of asking:
“What will this project cost?”
the better question can be:
“What is the expected project cost under different assumptions?”
AI can model variables such as:
This makes the estimate more transparent.
Procurement can become another AI-assisted workflow.
The system can compare supplier information such as:
However, purchasing decisions should not rely exclusively on the cheapest quote.
A lower price with an unreliable delivery schedule could cost more overall.
AI can therefore help evaluate total procurement risk, not merely unit price.
One operational problem is discovering too late that material is insufficient.
AI can monitor:
Required quantity vs purchased quantity vs delivered quantity vs installed quantity
If the projected balance falls below the expected requirement, the system can flag a potential shortage.
This is particularly valuable for phased commercial installations.
A commercial flooring project does not exist in isolation.
Other trades may be working in the same building.
The flooring team may need to coordinate with:
AI can help analyze dependencies and identify schedule conflicts.
An AI system could monitor:
It can then identify projects that are trending behind schedule.
The purpose is not to produce a perfect prediction.
The purpose is to provide an early warning.
Subfloor preparation can significantly affect flooring installation.
A project may encounter:
The precise acceptance criteria depend on the flooring system and manufacturer requirements.
AI can help organize inspection information.
For example, technicians could upload photographs and inspection records.
Computer vision may flag visual anomalies for human review.
But AI should not independently declare a substrate safe for installation unless the system has been appropriately validated and the responsible professional confirms compliance with applicable requirements.
Computer vision can potentially detect visible defects such as:
The camera system can compare images against defined quality criteria.
This can make inspection more consistent.
However, image-based inspection has limitations.
Lighting, camera angle, reflections, dirt and image quality can affect performance.
Therefore, AI inspection should be considered an additional inspection layer rather than an unquestionable authority.
A practical workflow might look like this:
Installation completed → photographs captured → AI pre-screening → flagged areas reviewed → inspector approves/rejects → issue recorded
Every flagged issue can become useful training data.
Over time, the company can discover recurring causes.
For example:
The objective should not be to blame workers.
It should be to identify process improvements.
Commercial construction involves substantial documentation.
An AI assistant can help organize:
Natural-language AI can make this information easier to search.
For example, a project manager could ask:
“Show me all unresolved flooring issues on Level 3.”
The system could retrieve relevant records if the underlying data is properly structured.
Changes are common in commercial projects.
A client might:
AI can compare the original scope against revised information.
It can highlight:
The final commercial change order should still be reviewed and approved by authorized personnel.
Suppose an additional 2,000 square feet is added.
The impact isn’t simply:
2,000 × flooring price
The change may also affect:
AI can help create a structured impact assessment.
There is no single price for AI development for commercial cork flooring.
The budget depends on:
A useful way to plan is to divide the project into levels.
Features might include:
This is generally the simplest starting point.
Add:
Add:
Add:
Each additional layer increases development and maintenance complexity.
Your AI budget should include more than software development.
Consider:
Understanding business processes and requirements.
Cleaning historical estimates and project records.
Creating interfaces for estimators, project managers and installers.
Managing project, material and cost data.
Creating prediction and optimization models.
If visual inspection is required.
Connecting accounting, ERP, CRM, BIM or supplier systems.
Hosting, storage, compute and monitoring.
Validating estimates against real projects.
Protecting commercially sensitive information.
Updating models and software.
Ignoring these costs is one of the easiest ways to underestimate an AI project.
A key strategic question is whether to build custom software or integrate existing technology.
Advantages:
Disadvantages:
Advantages:
Disadvantages:
For many commercial flooring businesses, a hybrid approach can be sensible.
Use established tools for general functions and build custom AI around the company’s unique estimating and operational data.
A sophisticated AI model cannot compensate for poor historical data.
Suppose your company has 500 previous projects but the records contain inconsistent measurements.
One estimator records labor hours.
Another records crew-days.
Another records only total project cost.
Another does not record waste.
Training a model on this information can produce unreliable results.
Data standardization should therefore be one of the first stages of development.
A useful project database could include:
This structure creates the foundation for useful analytics.
Return on investment should not be based on vague claims such as “AI will improve efficiency.”
Instead, identify measurable business metrics.
Potential KPIs include:
Then compare performance before and after implementation.
Suppose a project purchases 10,000 square feet of cork flooring.
If historical waste is 8%, the theoretical waste quantity is:
10,000 × 8% = 800 square feet
If better planning reduces waste to 5%:
10,000 × 5% = 500 square feet
Potentially avoided waste:
800 − 500 = 300 square feet
If the relevant material cost is $X per square foot, the direct material saving is:
300 × $X
Actual savings will depend on whether the reduction reflects genuinely avoided purchases, reusable inventory or simply a different measurement method.
This is why accurate baseline data matters.
Material savings are not the only source of ROI.
Consider a project where the initial estimate predicts:
But actual labor becomes 1,500 hours.
The additional labor can substantially reduce project margin.
If AI improves labor forecasting, the company may price the project more accurately.
In competitive bidding, better estimates can be as valuable as lower material consumption.
An AI system can assign risk indicators to estimates.
For example:
Low risk
Medium risk
High risk
The estimator can then review high-risk projects more carefully.
AI can also help prioritize bids.
A company may receive dozens of opportunities.
Instead of treating every opportunity equally, AI can analyze:
This can help management decide where to invest estimating time.
Large language models can be useful even when they are not directly estimating quantities.
They can assist with:
For example, a project manager could ask:
“Summarize the flooring requirements for all areas scheduled for Phase 2.”
The AI could produce a concise summary from approved project documentation.
Commercial flooring projects may contain long specification documents.
An AI document-processing system can potentially extract:
But extracted information should be verified against the original contract documents before it becomes a binding project instruction.
Building Information Modeling can provide structured building information.
If flooring zones are represented accurately in a BIM environment, AI can use that information for:
The advantage of structured data is that the system does not need to infer everything from photographs or scanned drawings.
A more advanced flooring platform could maintain a digital representation of the project.
The digital model could contain:
This creates a digital project history.
It can become useful for maintenance and future renovation planning.
Commercial flooring companies often have leftover material across multiple jobs.
Without good inventory visibility, usable remnants can be forgotten.
An AI inventory system can track:
When a new project needs a small quantity, the system can search for suitable existing inventory.
This can reduce unnecessary purchasing.
Where product appearance or compatibility depends on production batches, inventory records should preserve batch information.
AI can help identify whether available inventory is suitable for a particular project.
However, product compatibility should always be confirmed according to the applicable manufacturer’s instructions.
Material delivery can affect project productivity.
A flooring crew waiting for material is not installing flooring.
AI can help coordinate:
For large projects, the objective is to get the right material to the right zone at the right time.
A commercial building may have dozens of areas.
Installing everything in the order shown on a floor plan may not be optimal.
AI can evaluate dependencies such as:
The result can be a more practical installation sequence.
Suppose a company has several crews.
Each crew may have different experience levels.
AI can analyze historical productivity and match crews to projects.
Potential factors include:
The final decision should remain with management because productivity history alone does not capture every human factor.
A responsible AI strategy should treat installers as knowledge sources.
Experienced installers know things that may not appear in databases.
For example, they may recognize:
The AI system should capture and complement this expertise rather than trying to eliminate it.
For commercial flooring, human review is particularly important.
A practical workflow is:
AI recommendation → human review → approval → execution → outcome measurement
This is preferable to:
AI recommendation → automatic execution
The human-in-the-loop approach reduces the impact of AI errors.
Do not evaluate your AI system based on whether its outputs “look intelligent.”
Measure it.
For material estimation, track:
Quantity error = predicted quantity − actual required quantity
For labor:
Labor variance = predicted hours − actual hours
For waste:
Waste prediction error = predicted waste − measured waste
For schedules:
Schedule variance = actual duration − predicted duration
These metrics reveal whether the system is actually improving operations.
Do not begin with the entire business.
Choose one workflow.
Material estimation is often a sensible candidate.
Select perhaps 20 to 50 historical projects, depending on data availability.
Recreate their estimates using the proposed system.
Then compare AI results with:
This provides a baseline.
Choose representative commercial cork flooring jobs.
Standardize units, costs and terminology.
Use historical information.
This is essential.
Measure accuracy.
Identify why the system was wrong.
Improve data and rules.
Start with estimator review.
If an AI system is evaluated only against the same projects used to develop it, the results can look better than real-world performance.
A robust test uses unseen projects.
This helps determine whether the model generalizes.
The question is not:
“Can AI reproduce old estimates?”
The question is:
“Can AI make useful predictions for a project it has never seen before?”
The timeline depends on scope.
A simple prototype can potentially be developed much faster than a fully integrated platform.
A typical project sequence might be:
Map the existing workflow.
Collect and clean historical projects.
Build a small estimating workflow.
Test with real projects.
Improve accuracy and usability.
Connect business systems.
Expand to scheduling, inventory and quality.
Avoid setting unrealistic deadlines before understanding the data and integration requirements.
A commercial flooring AI platform might contain:
User interface
Estimators and managers interact with the application.
Application backend
Handles projects, users, calculations and workflows.
Database
Stores products, projects, costs and outcomes.
AI layer
Handles prediction, classification and natural-language functions.
Optimization engine
Handles cutting and scheduling problems.
Computer vision
Processes images where required.
Integration layer
Connects external systems.
Analytics layer
Displays KPIs and trends.
This modular approach makes future expansion easier.
Cloud deployment can provide:
On-premises deployment may provide greater control for certain organizations with specific security or infrastructure requirements.
The correct architecture depends on:
Commercial flooring project data can include confidential information.
Examples include:
AI development should therefore include appropriate:
Employees should understand what information can and cannot be entered into third-party AI services.
Generative AI can produce plausible but incorrect information.
For example, it might invent:
This is unacceptable if the information influences purchasing or installation.
The system should therefore use approved product databases and retrieval mechanisms.
For critical specifications, the AI should point users to the authoritative source rather than improvising.
Create rules for:
Define who is responsible when AI produces an incorrect recommendation.
AI governance is not merely an enterprise IT concern.
It matters whenever software influences real-world construction decisions.
A company may immediately request computer vision, robotics and digital twins.
That can consume budget without addressing the most valuable business problem.
Start with the workflow that has measurable economic value.
Without quality data, predictive AI becomes difficult.
An inaccurate automated estimate is more dangerous than a manually reviewed estimate.
Human expertise remains essential.
The number of AI features does not determine ROI.
For many commercial flooring businesses, material estimation is an attractive starting point because it has:
A company can build an initial system around:
Floor area + product + geometry + installation method → quantity estimate
Then progressively add:
waste prediction → labor prediction → procurement → scheduling → inventory
Replace fragmented spreadsheets with structured data.
Use historical data to improve estimates.
Add geometry and cutting analysis.
Connect supplier and inventory information.
Add scheduling and productivity forecasting.
Add computer vision and inspection analytics.
Integrate all project workflows.
This staged approach reduces implementation risk.
Imagine a hypothetical commercial project with:
The conventional approach may use:
Total area × fixed waste percentage
The AI-assisted approach could instead:
The important difference is that AI transforms estimation from a single arithmetic calculation into a data-driven workflow.
A flooring company should track at least four quantities:
Purchased material
How much was ordered?
Installed material
How much became part of the floor?
Reusable leftover
How much remains usable?
Discarded material
How much is genuinely waste?
This distinction prevents the company from claiming waste reduction simply because leftovers were moved into inventory.
Reducing unnecessary material consumption can support broader sustainability objectives.
Less material waste can mean:
However, sustainability claims should be based on measured outcomes.
Avoid claiming a specific environmental benefit unless it can be substantiated.
AI can also help create clearer proposals.
A proposal can summarize:
But commercial terms should always be reviewed before sending.
Natural-language tools can assist with routine communications such as:
This can save administrative time.
The system should not send commercially significant commitments automatically without appropriate approval.
One of the most valuable uses of AI may happen after installation.
Compare:
Estimate vs actual
For:
Then ask:
Why was the estimate wrong?
This produces organizational learning.
A mature system works like this:
Bid
↓
Estimate
↓
Purchase
↓
Install
↓
Measure
↓
Compare actual vs predicted
↓
Identify variance
↓
Update rules/model
↓
Improve next estimate
This is the foundation of a genuinely intelligent flooring operation.
AI may calculate quickly.
It does not necessarily understand every project nuance.
An experienced estimator may know that a particular building:
Some of that information can eventually be captured as data.
Until then, human judgment remains essential.
AI becomes more attractive when a flooring company has:
A company doing only a few small projects may receive better ROI from improving basic digital processes first.
If your business still relies on:
then the first investment should probably be data and process standardization.
AI works best when it has a reliable operational foundation.
If you decide to outsource development, evaluate providers based on:
Do not choose a provider solely because it promises the lowest development price.
The most important question is whether the provider understands your business problem.
Ask:
Good answers should be specific rather than based on generic AI marketing language.
Do not look only at development cost.
Consider:
Total cost of ownership = development + infrastructure + integrations + maintenance + support + model operations + internal training
A platform that appears inexpensive initially can become costly if every modification requires expensive developer involvement.
Build maintainability into the original architecture.
Successful implementation requires more than software.
Estimators should learn:
Installers and supervisors should understand how inspection information feeds the system.
Management should understand KPI dashboards.
Introduce AI gradually.
Collect data and document processes.
Standardize historical estimates.
Build prototype.
Pilot with selected projects.
Expand after measurable improvements are demonstrated.
The exact schedule will depend on development scope and business resources.
The people who use the system every day should influence its development.
Ask estimators:
Ask installers:
This creates a system based on actual operational needs.
The future of AI-assisted commercial flooring may include:
But not every company needs all of these capabilities.
The best system is the one that solves the company’s highest-value problems.
If you are planning AI development for commercial cork flooring installation, use this framework:
Find out where money and time are being lost.
Standardize estimates, costs, material quantities and project outcomes.
Material estimation is often a strong candidate.
Test it on historical and new projects.
Track material, labor, waste, time and margin.
Introduce waste and cutting optimization.
Connect estimation with purchasing.
Introduce scheduling and productivity forecasting.
Use visual AI for appropriate inspection tasks.
AI should support professional decision-making rather than replace it.
It is the development or integration of artificial intelligence tools that help a flooring company estimate quantities, forecast costs, predict waste, optimize cutting, manage projects, monitor inventory and analyze installation performance.
AI can improve quantity estimation when it receives reliable floor-plan, product and project data. However, estimates should be reviewed by qualified personnel before material is purchased.
Potentially. AI can analyze project geometry, historical waste patterns and cutting requirements to improve material planning and reduce unnecessary offcuts and over-ordering.
It can combine material quantities, supplier pricing, labor assumptions, preparation requirements, logistics and historical project data to produce a structured estimate.
Not always. A company may benefit from integrating existing estimating, document-processing or analytics tools before investing in fully custom AI.
There is no universal price. A focused estimating assistant is considerably less complex than a complete platform involving computer vision, BIM, inventory, scheduling and enterprise integrations.
Computer vision and document-processing technology can potentially extract geometry and information from floor plans. Results should be validated, particularly when drawings are unclear or incomplete.
Yes, if sufficient historical project data is available. The model can learn relationships between floor complexity, project size, product type, crew characteristics and actual labor hours.
It can classify and track leftover material, identify reusable inventory and potentially match existing remnants with future project requirements.
No. AI is most valuable as a decision-support tool. Experienced estimators should review important quantities, assumptions and commercial decisions.
Compare measurable baseline performance against post-implementation performance. Useful metrics include estimating time, material variance, waste, labor variance, rework, schedule performance and gross margin.
Collect historical floor areas, product information, purchased quantities, installed quantities, waste, labor hours, project duration, costs, floor-plan characteristics and final project outcomes.
AI development for commercial cork flooring installation is not fundamentally about adding an impressive chatbot to a flooring business.
The real opportunity is much more practical.
AI can help connect the information already generated throughout a flooring project and turn it into better decisions.
The strongest applications include material estimation, cost forecasting, waste prediction, cutting optimization, labor forecasting, procurement planning, inventory management, project scheduling and quality inspection.
For a commercial flooring company, the business case should begin with measurable problems.
If excessive material is being purchased, improve estimation.
If offcuts are accumulating, improve cutting and inventory optimization.
If estimates frequently miss labor requirements, build a productivity model.
If projects are regularly delayed, develop predictive scheduling.
If inspection is inconsistent, investigate computer vision.
The most effective AI strategy is therefore not:
“Where can we put AI?”
It is:
“Which recurring business problem costs us the most money, time or margin, and can reliable data help us solve it?”
Start with that question.
Build a focused pilot.
Measure actual results.
Keep experienced flooring professionals involved.
Then expand the system only when the evidence supports doing so.
That approach can turn AI from a technology experiment into a practical operating capability for commercial cork flooring installation.
Get cost, timeline & technical suggestions within 24 hours.
