Sanitary sewer is the one design where gravity makes most of the decisions before you do. Wastewater runs downhill at a minimum slope or it doesn’t run at all, and the ground either cooperates or it doesn’t. The judgment in that is real but small. The rest is a day of reach-by-reach hydraulics and a writeup, which is exactly the kind of work I didn’t want the professional engineer in charge spending a day on.
So I handed the production to an agent. And because the same agent also designs individual septic systems, which are governed by soil instead of slope, it’s a clean look at something I’ll come back to at the end: what actually compounds when you build the second workflow, and what doesn’t.
What a sanitary design actually involves
Start with the flow. Count the units or the population, run them through the jurisdiction’s per-capita or per-bedroom basis (75 to 150 GPD depending on who adopted what), and apply a peak hour factor, usually 1.5 to 3.0 times average. Then, if the system is older or the climate is wet, add infiltration and inflow, because old clay sewers leak and a smoke or dye test will tell you how much (500 to 1,500 GPD per mile per day is a working assumption for aging clay). Get the design flow wrong and every pipe downstream is sized against the wrong number.
Then you build the gravity model. This is where sewer is its own animal. You’re not holding a pressure; you’re running partially-full open-channel flow down a pipe on a slope, reach by reach, with Manning’s equation. And you check three things at every manhole, not one:
- The slope has to sit between the jurisdiction’s minimum and a sane maximum. California wants 0.5% minimum on a gravity main (Title 22, Table 1-WR); some states run 0.3%. Too flat and solids settle; too steep and the flow outruns the solids and you get the same problem at the bottom.
- The depth of flow over the pipe diameter (the d/D ratio) has to stay below about 0.75. A sanitary main isn’t supposed to run full. If d/D climbs past that, you’ve under-sized the pipe or under-sloped it.
- The velocity has to clear a cleansing threshold so the pipe scours itself instead of silting up, and the hydraulic grade line can’t surcharge above the invert at any manhole.
Then the fork that defines the whole design: gravity or pressure. If the outfall is downhill or level, gravity wins, and it wins clean, because there’s no pump, no power, and nothing to maintain for the life of the system. But if the outfall is uphill, or the site is so flat that holding minimum slope buries the pipe deeper and deeper until the trench is absurd, you switch to a lift station and a force main: a pump sized to peak hour flow (not average), a force main held to 3 to 6 ft/s, total dynamic head worked out from static lift plus friction, air and vacuum valves at the high points, and a discharge pressure kept under the safe limit so you’re not setting up a burst. That’s a pump you’ll own and run and maintain forever, so the call between the two is genuine engineering.
Septic is the same question (where does the wastewater go, and can the receiving system take it) with a completely different governor. Here it’s the soil, not the slope. Design flow is still bedroom-based (75 GPD per bedroom, more for high use, plus an allowance for the disposal), but the absorption field is sized off the soil’s measured percolation rate: field area equals design flow divided by the loading rate the soil can accept. And the soil gets a veto. If the perc test comes back slow (north of 60 minutes per inch) or the soil is clay, an in-ground trench field won’t work, and you move to a mound or a drip system. Sizing a trench field in clay because the bedroom count said you could is how a system fails two years in.
Then, either way, you write it up: flow basis, the hydraulic or soil calc, the profile or the field layout, the code citations, the appendices. Format to firm standards. QC it. And the professional engineer in charge reviews the analysis, owns the judgment, and stamps it.
If you’ve designed a collection system or laid out a leach field, none of that is news. If you manage the people who do, you already know it’s the better part of a day, and almost none of it is the judgment.
What it used to take us
A full sanitary design at Calichi, the flow basis, the gravity model run reach by reach, the gravity-versus-pumping call, and the written report, ran the better part of an engineer’s day. A septic layout with the soil work and the field design ran the same shape.
I’m not going to put a stopwatch number on this one. Fire flow is the workflow we measured to the hour, 16 to 24 hours of engineer time down to about 2, and I’ll only claim a measured number where it’s true. This isn’t that workflow, and it doesn’t even share fire flow’s hydraulic engine, so borrowing the figure would be dishonest. What I can tell you honestly is the shape of it: the flow basis takes minutes, the gravity-versus-pressure call and the soil read are real judgment and they deserve an engineer, and everything else is Manning’s on every reach, d/D and surcharge at every manhole, and the writeup, done by the most expensive person in the room. The honest way to find your firm’s number is to measure it on your real projects, which is exactly what the Discovery does.
How the agent does it now
The engineer emails the agent the way they’d brief a junior: sanitary for this project, here’s the address, the units and occupancy, the existing sewer plans, the outfall, and the topo; or for septic, here’s the lot, the bedroom count, and the perc data. From there:
- It resolves the jurisdiction from the address and loads that agency’s standards: minimum and maximum slope, minimum diameter, depth limits, and for septic the state and county soil-absorption tables.
- It computes design flow from the per-capita or per-bedroom basis, the peak hour factor, and an I&I allowance for older or wet-climate systems, and shows the formula and the citation.
- It builds the gravity model from the project files, with inverts, distances, and pipe sizes pulled from the as-builts and the CAD utility layer rather than guessed, then runs Manning reach by reach and checks slope, d/D, velocity, and surcharge at every manhole.
- Where gravity fails (a surcharged manhole, or a depth that’s gotten absurd), it evaluates the lift-station and force-main alternative and sizes it: pump to peak flow, force main to velocity, total dynamic head, air and vacuum valves, discharge pressure under the limit.
- For septic, it sizes the absorption field off the measured perc rate, not the bedroom count alone, and flags slow perc or clay and switches the method to mound or drip instead of sizing a trench field that will fail.
- It drafts the report in firm format with the profile or field layout and the appendices, and routes the whole package through a separate QC agent before a person sees it.
Then the finished package lands in the project folder, and the engineer picks it up the same day instead of clearing one for it.
The QC gate
Same architecture as everywhere else in the system, and I wrote a whole piece on why the independent check exists, so here I’ll just say what it catches on this workflow. On the collection side the QC pass flags a reach sloped below the jurisdiction minimum, a d/D ratio over the limit (a main running too full), a velocity under the cleansing threshold (solids will settle), a surcharged manhole, and a force main discharge pressure over the safe ceiling. On the septic side it flags a trench field sized in soil that can’t accept it, and a loading rate that doesn’t reconcile with the perc rate on record. It independently recomputes Manning’s full-pipe capacity and the d/D ratio on every segment, and it won’t ship if the confidence on the sizing stays low. This is sewage backing up into buildings and leach fields failing under houses, so the bar to pass is set high on purpose.
The point isn’t that the first agent is perfect. It’s that a check that runs the same way every time catches the boring mistakes, so the engineer spends attention on the ones that need judgment.
What stays human
Everything that needs the professional engineer in charge. The engineer sets the flow basis and the I&I assumption, makes the gravity-versus-pressure call, reads whether a site’s soil can support on-site disposal at all or whether this should have been a sewer connection in the first place, weighs the unusual hydraulic condition, and applies the stamp. The agent does the production pattern-work. The engineer does the engineering. That division is the whole design, and it’s why this supplements your engineers instead of replacing them.
Where it applies
The collection-system side shows up on subdivisions, campuses, mixed-use, commercial, and industrial sites that tie into a public main or a treatment plant. The septic side shows up on rural and large-lot residential where there’s no main to tie into. Both cross jurisdictions, and that’s where the manual version gets slow, because the slope minimums, the diameter and depth rules, and the soil-absorption tables all change at the county or state line. An agent that resolves the framework from the project location and applies the right standards turns the slowest part into the fastest.
We run it everywhere we practice, and the code lookup resolves anywhere in the country.
What this means for your firm
Here’s the honest version of the compounding story, because this workflow is the one that complicates it. The hydraulic engine here, Manning’s partially-full gravity flow, is genuinely different from the pressurized EPANET models the fire flow and domestic water agents run. So this one did not ride the same pipeline the way domestic water rode fire flow. I had to build a different physics engine for it.
What it rode was everything around the physics: the jurisdiction resolver, the CAD and plan reader, the report builder, and above all the independent QC architecture. That’s the part worth understanding before you buy anything. You don’t need one universal model that does every kind of engineering. You need shared scaffolding that each new physics engine plugs into, so the second and third workflows cost you the engine and not the whole machine around it. The scaffolding is the asset. The engines slot in.
That’s what the first two gates of how I deploy are for. An AI Readiness Audit to find where the highest-value automation sits across your divisions, then a Strategic AI Discovery that puts my engineers and yours on your actual workflows and produces the hours-saved picture by workflow, so you see exactly where the time goes back. Every gate is a stop or go. You only go further once you’ve watched the last step work on your own projects.
If your engineers are spending a day each on sewer hydraulics and the writeup, or on septic layouts and soil sizing, the first step is the AI Readiness Audit.