When will AquaForge Generative Design be available?

Generative Design is targeted for a 2027 launch, following the Regulatory Intelligence module that is available now. Engineers can join the early-access list to be notified first and to influence the roadmap.

Coming 2027

Generative Design for Water Treatment.
Explore every treatment chain.

Q: What is generative design for water treatment?

Generative design for water treatment is AI-driven process design. Instead of an engineer manually configuring one treatment chain at a time, the software explores thousands of possible configurations, simulates the most viable ones across cost, footprint, energy, and compliance, and presents the best alternatives with the reasoning behind each. The engineer reviews, questions, and decides.

Q: Does generative design replace the process engineer?

No. AquaForge augments engineers by surfacing and explaining options faster than any manual process can, but every recommendation requires human review and approval. The engineer retains full responsibility for the design.

Q: How accurate are the cost and performance estimates?

Outputs are at preliminary-design level, appropriate for comparing alternatives and setting direction rather than for detailed P&ID design. Process models are built on validated physics-based simulation and independently characterized equipment data rather than optimistic vendor specifications.

AI-powered water treatment design software that generates and optimizes 10+ treatment chains in hours — and explains every decision.

Get notified at launch → See what’s available now

Optimized treatment chains per design session

Faster simulation than existing design tools

Of operating cost is disposal — optimized, not ignored

Of recommendations explained and traceable

The cost isn’t just in the treatment. It’s in the disposal.

Most people outside the industry think water treatment is about cleaning water. Engineers know the real economics are elsewhere.

Disposal drives capital cost

Disposing of captured contaminants in an environmentally compliant way accounts for an estimated 40–50% of project capital costs. A chain that removes well but disposes expensively is not optimal.

Disposal dominates operating cost

Over 70% of ongoing operating expense is tied to contaminant disposal. Optimizing treatment performance alone leaves the largest cost lever untouched.

Generative Design optimizes across the full cost picture at once — treatment performance, disposal economics, footprint, energy, and compliance — not one dimension in isolation.

How generative design works

Three steps from your constraints to design-ready treatment chains.

Feed the constraints

Bring in regulatory requirements from Regulatory Intelligence (or enter them manually) plus your input water characteristics — as ranges, with correlations between parameters. The system understands uncertainty natively.

AI explores the solution space

Patent-pending simulation technology evaluates treatment chain options across compliance, cost (CAPEX and OPEX including disposal), footprint, and performance — up to 100x faster than existing tools.

Review and refine

Get 10+ ranked alternatives with trade-off analysis. Ask why any chain was chosen, give instructions (“prefer Xylem for ozonation”), and the AI re-optimizes around your direction.

Built for how design engineers actually work

New plant or plant upgrade — the design problem is the same: explore more options without running out of time. Switch between them.

Design a compliant treatment chain from a blank sheet

Trigger

You’ve won an RFP for a new drinking water or wastewater plant. You have regulatory requirements and raw water quality data, and weeks — not months — to settle on a design direction.

The Pain

Manual design explores only a handful of options. Time pressure pushes engineers toward familiar approaches rather than optimal ones. The full design cycle runs 6 to 24 months.

How AquaForge Helps

The system screens thousands of treatment chain permutations, then simulates the most viable candidates across cost, footprint, energy, and compliance — producing 10+ ranked alternatives with full process flow diagrams.

Outcome

A defensible design direction in hours instead of months, backed by explainable trade-offs you can take to the client.

Greenfield design — new-build treatment chains

Upgrade an existing plant with minimal disruption

Trigger

A plant with conventional treatment must now meet new requirements — for example, EU quaternary limits for micropollutants. You need to know what to add, what to modify, and what to keep.

The Pain

65% of treatment projects are modernizations, yet most design tools were built greenfield-first. Modeling an existing plant and isolating incremental cost is slow and error-prone by hand.

How AquaForge Helps

Define or upload the existing treatment chain, set a “minimal modification” constraint, and the system generates upgrade paths with incremental cost separated from total cost and a disruption assessment for each.

Outcome

Targeted upgrade options that respect the infrastructure you already have — compliant, costed, and ready to compare.

Brownfield modernization — built in from day one

Descriptive, not prescriptive

Generative Design explains its reasoning. You steer. Every answer traces back to the data behind it.

YOU

Why did you recommend ozonation over UV for this step?

For the PFAS range (12–45 ng/L) against the applicable discharge limits, ozonation reaches ~94% removal versus ~67% for UV. If footprint were your priority, UV needs about 40% less floor space — so the answer depends on which constraint you weight.

YOU

Input BOD is 150–280 mg/L. At peak load, what are my margins?

At 280 mg/L the chain still meets every limit, with the secondary stage running at ~85% of rated capacity. Below 200 mg/L you have headroom to reduce aeration energy.

YOU

We prefer Veolia membranes for this application.

Adjusting. The chain holds compliance with that membrane family and CAPEX shifts by a small margin. Here’s the updated trade-off chart and the two alternatives it changes.

Every recommendation is traceable. Every decision is yours.

Work backwards from your compliance targets

Most simulation tools only run forwards: given inputs, what comes out? Generative Design also runs in reverse — given your discharge limits, how much influent variability can the chain absorb? Move the sliders to see the idea.

Target discharge limit 10 mg/L

Treatment chain robustness Standard

Design safety factor 1.3×

Max influent you can handle

0 mg/L

Illustrative model of the reverse-simulation concept — not a design output. The product computes this from validated process models.

How we’re different

Generative design isn’t new. An optimizer that explains itself and handles real-world uncertainty is.

Dimension	Traditional tools	AquaForge Generative Design
Interaction	Configure parameters, click generate, read the output	Conversational — ask “why,” give instructions, iterate
Philosophy	Prescriptive — tells you what to do, not why	Descriptive — explains the reasoning, you decide
Input handling	Single average or typical values	Ranges with correlations between parameters
Design space	Engineer pre-decides much of it, narrowing the options	Open exploration — the AI doesn’t inherit human bias
Speed	Minutes to hours per single simulation	Up to 100x faster, even with range-based inputs
Reverse simulation	Not available	Given outputs, what input margins do you have?

Comparison is against traditional design tools as a category.

Starts where Regulatory Intelligence finishes

Compliance gap analysis

→

Design constraint set

→

Optimized treatment chains

Your compliance gap analysis becomes the design constraint set — no re-entry, no spreadsheet handoffs, full traceability from regulation to treatment chain. And it’s available today.

Explore Regulatory Intelligence →

Frequently asked questions

What engineers ask before joining the early-access list.

What is generative design for water treatment?

It’s AI-driven process design: instead of an engineer manually configuring one treatment chain at a time, the software explores thousands of possible configurations, simulates the most viable ones across cost, footprint, energy, and compliance, and presents the best alternatives with the reasoning behind each. The engineer reviews, questions, and decides.

When will Generative Design be available?

Generative Design is targeted for a 2027 launch, following the Regulatory Intelligence module that is available now. Join the early-access list to be notified first and to influence the roadmap.

How is data quality assured in generative design? 

Data quality is maintained through input validation, engineering rule checks, standardized unit normalization, simulation verification, and human engineering review.

Does it replace the process engineer?

No. AquaForge augments engineers — it surfaces and explains options faster than any manual process can, but every recommendation requires human review and approval. The engineer retains full responsibility for the design.

Can AI be trusted for critical infrastructure projects?

Generative AI should be treated as an engineering support tool rather than an autonomous decision-maker. There should always be an “engineer in the loop” with an AI assistant. This hybrid approach improves efficiency while maintaining safety and accountability.

Can it design upgrades to an existing plant, not just new ones?

Yes. Brownfield modernization is built in from day one because most treatment projects are upgrades, not new builds. You define or upload the existing chain, set a minimal-modification constraint, and the system generates upgrade paths with incremental cost and disruption clearly separated.

Do I need Regulatory Intelligence to use it?

No. Regulatory Intelligence feeds requirements and gap analysis directly into Generative Design with no re-entry, which is the smoothest path — but you can also enter requirements manually if you don’t use it.

How accurate are the cost and performance estimates?

Outputs are at preliminary-design level — appropriate for comparing alternatives and setting direction, not for detailed P&ID design. Process models are built on validated physics-based simulation and independently characterized equipment data rather than optimistic vendor specs.