Proposal
A teacher sees something worth remembering. Then the day keeps moving.
The discovery call on September 1 narrowed this product, and the narrower version is the better one. It is the version this proposal quotes.
Describe a project in advance, select frameworks, let AI generate a checklist of standards, then have faculty assess against it across four role dashboards on three platforms.
Select a student, record a voice note after a conference, let AI say what it demonstrated, approve it, and the record updates. One role, one device, one workflow that has to be excellent.
It should be primarily a data-capture tool, not a data-management tool from the teacher's perspective.
Prairie School · discovery call, 9/1/26That single sentence reshaped the estimate. A capture tool has one screen that must be perfect and a handful that must be clear. A management platform has four audiences, three form factors, and a permissions model. We have scoped, and priced, the capture tool.
Everything the RFP asked for that is not in the launch release appears later in this document by name, so the narrowing is a decision you made with us rather than something discovered halfway through a build.
Under five seconds from opening the app to recording. Nothing filed until a human says so.
The roster is the home screen. Search and recents, not a dashboard. Two taps to recording.
During or right after a conference. One hand, minimal typing, no file management.
Transcribes, writes a plain-language summary, and proposes two or three standards with a reason for each.
Confirm, correct, add, remove. Nothing enters the record until a teacher says so. Trust is won or lost here.
Evidence attaches to the student, the project, and the standards. Exposure is tracked over time.
This is the reason to hire a design and engineering partner rather than license a standards tracker, and it is the first thing that disappears if a phase gets compressed. We would rather protect it out loud.
A student already performing at grade level in math is not necessarily demonstrating growth by continuing to get 100%. The goal is to show how the student is pushing beyond what they already know.
Prairie School · discovery call, 9/1/26During Phase 1 we develop several visual directions for representing a learner's growth over time, present them for a decision, and design the one Prairie selects into a finished visual.
What a learner knows is a shape. New evidence pushes the boundary outward in a direction. A year becomes a shape that visibly grew.
Standards as nodes, evidence as the lines between them. Shows the cross-disciplinary connection that project-based learning claims and rarely proves.
Organized by domain, with ring depth as repeated exposure over time. Legible on day one and it handles exposure natively.
One project as a path, with photos and voice notes surfacing as artifacts in sequence. The view closest to a story.
Wherever the RFP showed a box with a note count, the product shows a frequency and recency mark instead.
Breadth is how much of the map has been touched. Depth is how often and how recently. Teachers need both.
In September a learner has almost no evidence. If the product is not beautiful when nearly empty, teachers judge it in week one and never come back.
Nothing should read as deficit. Absence of evidence is "not yet visited," never "failing."
Everything below is what it takes to deliver that, based on what we know today. Items marked with a dot are where the product succeeds or fails.
Teachers sign in with the Google Workspace account the school already gives them, so there are no new passwords to manage. Two kinds of users at launch: teachers, and a school administrator who manages settings.
One place where the school is configured: which standards frameworks are turned on, which teachers have accounts, and which classes exist. Built so a second school could be added later without rebuilding the product.
The roster: each student, their grade, the classes or blocks they belong to, and which teachers work with them, including specialists. Where the roster comes from is a decision we make together during design.
The full text of every framework Prairie uses, loaded in a structured form with its hierarchy, grade bands, and codes intact: Common Core ELA and Math, NGSS, Illinois SEL, Illinois Learning Standards, National Core Arts, and Prairie's own learner outcomes.
Look up any standard across the frameworks the school has turned on, browse by framework, and see which students have evidence against it. A lookup tool, deliberately simple.
Pick a student, then talk, snap a photo, upload a file, or type a quick note. Designed to be used one-handed and to start in under five seconds, so it is faster than a notebook. Behind it sits secure storage and playback for audio, photos, and documents.
Voice notes become text automatically. The product is taught the names of Prairie's students and Prairie's own vocabulary so they come through correctly, a teacher can fix any word, and the original recording is kept alongside the text.
From the transcript, the AI writes a plain-language summary of what the student demonstrated, then proposes two or three standards it believes the observation shows, each with a reason. It can also point out a standard the teacher may not have considered. When it is not confident, it says so rather than guessing.
Nothing is filed until the teacher says so. The teacher checks the transcript and summary, keeps or drops each suggested standard, adds any the AI missed, optionally tags a project, and files it. The product records what the AI proposed and what the teacher changed. This also covers the moments when things go wrong: poor audio, a wrong suggestion, no suggestion, or the AI service being briefly unavailable.
Everything filed about a student in one place: a timeline of observations in the teacher's own words, with voice, photo, file, and typed notes together, and a view of each standard as a history of when it showed up rather than a box that is checked or not.
A light way to group observations: what projects are running, roughly where they are, which students are on them, and which standards are showing up most. Not a project-management system.
The showpiece: one visual that shows a learner's growth over time as something other than a bar filling up. The direction is chosen during design, then built so that it draws itself from the real record, can be viewed as of any point in the year, and is worth looking at even in September when it is nearly empty.
The servers, database, secure file storage, and background processing that run the AI steps, plus the tools we use to deploy updates, monitor the system, and keep backups.
Teachers see only the students they work with. Data is encrypted, changes are logged, and there are clear rules for how long audio and photos are kept and how records are corrected or deleted. Agreements with the transcription and AI providers prohibit them from training on or keeping Prairie's data. Prairie's data can be exported at any time.
The product meets WCAG 2.1 AA, including a non-visual equivalent of the growth visualization so that it does not depend on color alone.
The capture screens are built for a phone. The reading and reviewing screens, meaning a student's record, standards, projects, and school setup, also get proper layouts for a laptop or desktop.
Delivered as an installable web app: teachers open a link, add it to their home screen, and it behaves like an app, with no app store involved.
Each of these is a real part of the eventual product. Listing them here means they can be planned on purpose later rather than rediscovered mid-build.
We design the product first, prototype it, put it in a teacher's hands, and then price the build against real decisions.
Understand how teachers work today, settle the standards foundation, define what growth looks like, design the workflow, and build a clickable prototype that runs on a phone.
Prairie ends this phase with something to react to, test with teachers, and show to a board.
Production development of everything in section 04: the capture workflow, the AI pipeline, the standards library, the record, the growth visualization, and the infrastructure underneath.
Then iterate with Prairie teachers using it in real classrooms.
Steps 1.3 and 1.6 together are a third of the phase. That is deliberate, and the highlighted rows below are the ones we would protect if the phase ever needed to compress.
Half a day with two or three Prairie teachers, watching how they currently document student learning: the handwritten notes, the Docs and Sheets, the photos, the folders. We are looking for where the time actually goes, because the product's central promise is capturing an observation faster than a teacher can write it down. We then map what a teacher, a student, a parent, and an administrator is thinking and trying to accomplish at each stage.
The product's usefulness depends on how well it can match a real classroom moment to real standards. We gather the frameworks Prairie uses, assess what form that data is actually available in, and define how multiple frameworks coexist in one interface. This step also determines how a future school would select its own frameworks, which is what makes the product sellable beyond Prairie.
We develop and present visual directions for representing a learner's growth over time, and work with Prairie to choose one. This is the part of the product that cannot be bought off a shelf, and it is what turns evidence of learning into something a family will actually keep.
The screens and flows for the launch product: select a student, capture, review the transcription and summary, approve or correct the suggested standards, and watch the record update. This includes the states that decide whether the product feels trustworthy, meaning what happens when the audio is poor, when the AI suggests the wrong standard, when it suggests nothing, and when a teacher has no signal.
We assemble the designs into an interactive prototype that runs on a phone. It is not connected to real data, but it is complete enough to hand to a Prairie teacher and watch them use it, and complete enough to demonstrate to a board, a funder, or a prospective partner school.
Take the direction Prairie picked in step 1.3 to a finished visual, with the supporting interface around it. This is the showpiece for the Phase 2 conversation and the thing that makes the product recognizably Prairie's rather than generic.
Running in parallel throughout: specifying what the AI is being asked to do at each step, what a teacher confirms, what the system stores, and how it behaves when it is unsure. This becomes the technical roadmap and the basis for the Phase 2 estimate.
Eight workstreams, estimated before design. Below them, the three variables that move the number most, and two things we would rather name than bury.
Google Workspace sign-in, the two user types, school configuration, roster, classes, and the multi-tenant structure that lets a second school exist later.
Acquiring, structuring, loading, and maintaining every framework, plus the lookup interface on top of it. The single largest scope variable in the engagement.
The under-five-seconds capture screen, secure audio and photo and document storage, playback, and the transcription service tuned to Prairie's names and vocabulary.
Summarization, standards matching against thousands of candidates, reasons for each suggestion, confidence handling, the missed-standard nudge, and every failure state behind it.
The approval surface, the audit of what the AI proposed against what the teacher changed, the student timeline, standards histories, and light project grouping.
Turning the design from Phase 1 into a real visualization that draws itself from live data, scrubs to any point in the year, and has a non-visual equivalent.
Servers, database, background processing, deployment, monitoring, backups, encryption, audit logging, retention and deletion rules, and data export.
WCAG 2.1 AA conformance, laptop layouts for the reading screens, installable web app packaging, testing, and release to the pilot teachers.
Some frameworks are published in structured, machine-readable form. Others exist only as documents on a website, or require a license from a data provider. Frameworks are also revised over time, so the library needs a plan for staying current.
The launch plan assumes a working connection when a teacher records. Much of Prairie's day happens outside, and your grounds include places without signal. Offline is not a small addition: the workflow changes to capture now and review later, and recordings have to be stored and synchronized safely.
The plan assumes an installable web app, which is the faster path and works on any phone. There are real reasons to consider native: more reliable audio recording when a call comes in or the screen locks, better offline capability, and app store distribution.
Art+Logic has been building custom software since 1991. What has changed recently is how much of the production work collapses, and being honest about which half that is.
A half day in a classroom is a half day. Client checkpoints, teacher sessions, and the conversations where decisions actually get made take exactly as long as they always did, and we budget them at full weight.
Choosing the information architecture, picking the growth direction, knowing when something is actually good. Our tools help us think faster, but they do not think for us. We discount this lightly and deliberately.
Drawing frames, generating states and variants, building prototype code, writing documentation. This is where modern tooling genuinely changes the arithmetic, and it is a large share of a traditional estimate.
The practical result for Prairie is a design phase measured in weeks rather than months, a prototype in teachers' hands sooner, and more rounds of iteration inside the same budget. The rate is Art+Logic's standard $125 per hour. What has changed is the scope a given number of hours now covers.
Grouped by how much they change. The first three are worth answering before or during week one.
Common Core, NGSS, Illinois, and Prairie's own. What form this data is in is the biggest single lever on both phases.
One document will teach us more about the growth model than an hour of discussion will.
A real design fork. Both are reasonable, and the answer changes how the standards and setup work is structured.
The RFP mentions 100 to 150 users, but teacher count is what actually sizes the launch release.
The RFP asked for it, the call did not mention it. We would rather decide it than silently include or silently drop it.
Each is a different amount of work and a different ongoing maintenance story.
The privacy assumption says after, the workflow description says during or immediately after. It affects both the design and the consent posture.
It drives the vocabulary and tone of the growth visualization even before a student-facing view exists.
The growth artifact should look like it came from Prairie, not from a software company.
We have assumed it. Worth confirming, since color-encoded and shape-encoded growth data needs non-visual equivalents designed alongside it.
Art+Logic bills at $125 per hour against a monthly burn-rate ceiling agreed in advance. Prairie can pause or stop at any point.
A working session to walk through the scope, confirm what is in and out of the launch release, and answer the three questions in section 10.
Contract terms and a monthly burn-rate ceiling for Phase 1, sized to Prairie's budget cycle.
Schedule the teacher observation session, begin the standards data investigation, and start the growth model conversation.
A prototype in teachers' hands, a growth visualization Alyssa can show a board, and a Phase 2 number we can commit to.
For more than three decades we have delivered business-critical custom software for organizations that cannot afford to get it wrong. We understand that technology must serve the mission, not dictate it.
We deploy multidisciplinary teams of senior software architects, UX strategists and cloud engineers who prioritize understanding your workflows before writing a single line of code. We leverage modern AI tooling to accelerate research and production work, but every architectural decision, every design judgment and every security review remains under the supervision of experienced humans.
Our delivery philosophy is simple: transparent, collaborative, and predictable.
Custom software"For over 35 years, Art+Logic has specialized in building custom software others said couldn't be done. With deep expertise in AI workflows, systems integration and elegant UX, we're uniquely positioned to help Prairie turn two minutes of a teacher talking into a lasting record of a child's growth."
Almost every assessment product on the market can tell a parent that their child completed 90 percent of something. Very few can show a family where their child's edge actually moved this year. That second thing is what Prairie is asking for, it is harder, and it is the reason this is worth designing properly before it is built.
Phase 1 starts within two weeks of a signed agreement. We are ready when you are.
Prairie School of DuPage × Art+Logic · Confidential · September 2026