Future Development Features: 2027

Why it matters

Alumni are people the movement already knows and trusts, and partners have asked whether they can be added to NovaCanopy. The risk is the one the platform has met before: people come onto a platform and don’t know what to do there. So the experience has to start from where an alumnus can plug in, not just from having an account. This entry captures the request. What the experience looks like, and how alumni are onboarded, still has to be designed.

What it looks like

Early idea, to be designed

An alumnus signs in and sees what is currently going on, such as ongoing projects. If a project interests them, they can say they want to be involved, for example by mentoring or by connecting the project with someone. NGLs would be the first group to use it.

Why it matters

Mentorship depends on knowing who is actually available, and today the mentor pool lives outside the platform, in spreadsheets, email threads, and individual consultants' personal contacts. There is no single reliable answer to who is mentoring, in what sector, or whether they're still active. Onboarding a mentor means re-typing their background by hand, and removing one who has stepped away depends on someone remembering to do it. That costs coordinators time they could spend making matches, and it limits mentees to whichever mentors a particular consultant happens to know, rather than the full pool. Holding mentor profiles in NovaCanopy creates one authoritative, current record; generating a profile from the mentor's own LinkedIn removes the manual data entry; and add and remove control keeps the pool accurate as mentors join and leave. It's also the prerequisite for the rest of this epic, since sectoral filtering and mentee shortlisting both assume a maintained mentor pool on the platform.

What it looks like

Invitation

A mentor receives a request to share their LinkedIn profile.

Profile creation

A scraping tool reviews the profile and generates a mentor profile in NovaCanopy.

Why it matters

Today matching is informal: a consultant suggests someone from their own network over WhatsApp or email, so a mentee's options depend on who their consultant knows, pairings go unrecorded, and no one can see how many mentees a mentor has already taken on. Mentee-led shortlisting puts the choice with the person it affects, keeps mentors from being over-subscribed, and leaves a record of who was matched with whom and when.

What it looks like

A mentee logs in, filters the mentor pool by sector, and opens the profiles that look closest to their goals. They select up to three as a shortlist and submit it; the platform confirms the submission and shows the status of each request. Mentors see incoming requests and accept or decline, and the pairing is recorded once a mentor accepts. A mentee whose first choices decline can revise their shortlist from the same screen. Coordinators can see shortlists and confirmed pairings across the cohort, including how many mentees each mentor is carrying.

Why it matters

Coordinators currently scroll one long, undifferentiated list and count drafts by eye to work out where a cohort stands: slow, easy to miscount, and repeated every time someone asks for a status update. Filtering turns that into a single click and makes the list actionable rather than just readable, pulling every draft in one working group to chase, every submission in a term to review, or every entry by one learner to follow up. As the movement grows past 150 participants in 2027 the list only gets longer, so without filtering the time spent finding the right entries rises with every cohort while the time spent acting on them falls. Filtering is also what makes the reminder, download, and export features usable at scale, since each one acts on whatever slice is on screen.

What it looks like

A coordinator opens a submissions list and sees a Filter control in the menu bar. They pick the fields they need, status, working group, term, learner, or date, and the list narrows as each filter is applied, with the active filters shown above the list and a count of matching entries. Filters combine, so "drafts in Working Group Responsible AI this term" is a few clicks rather than a manual scan, and one action clears them. Column headers also sort, so the same list can be ordered by date or learner. Whatever is on screen is what the reminder, download, and export actions act on.

Why it matters

Submission rates are the movement's core delivery metric, and today they rest on informal chasing. A learner whose work sits in draft is followed up over WhatsApp or a personal email, if a coordinator happens to notice, and nothing about that follow-up is recorded. The result is that unfinished work gets lost quietly: the learner doesn't get the nudge, the coordinator can't prove one was sent, and by term end no one can distinguish a learner who was never reminded from one who chose not to submit. Built-in reminders close that gap on both sides. Follow-up becomes one click from the list the coordinator is already working in, and the timestamp and sender stay on the learner's row, so no learner is chased twice by two people, and every reminder is evidence of duty of care the team can show a partner or funder. It's also the lowest-cost lever the platform has on completion, since it acts on work already started, where a single prompt converts a draft into a submission. Past 150 participants in 2027, manual chasing is the first thing to fall away; without reminders in the platform, submission rates fall with it.

What it looks like

A coordinator filters the submissions list to drafts. They select "Send reminder" on a learner's row, and NovaCanopy emails that learner a prompt to complete the submission. The row then shows when the reminder was sent and by whom, visible to anyone else working the same list.

Alternative worth considering

This could instead be an automated procedure, an AI agent that sends a reminder to any submission still in draft at admin-defined intervals.

Why it matters

Each learner's five submissions currently sit in five separate places, so coordinators piece a learner's record together circle by circle and can't tell at a glance what's complete. One linked, sequential record removes that reassembly work, makes gaps immediately visible, and keeps the coordination load flat as the cohort grows.

What it looks like

An admin opens the Community Circle list and sees one row per learner. Expanding the row shows the learner's five submissions in sequence, each marked submitted or still in draft, so the full picture sits in one place rather than across five circles.

Why it matters

Today, learner reflections and submissions live in the module where they were captured: community circle journals in one place, project documentation in another, poll responses somewhere else. The learner never sees the thread connecting them, and neither does the team. This came directly from consultant feedback (Fadi, Aug 2026): learners experience the movement as modular pieces "thrown at them" with no visible connection. A cohesive journal turns fragmented entries into a story, which deepens reflection for the learner, gives consultants a single place to understand someone's growth, and produces exactly the kind of longitudinal narrative the movement needs for storytelling and reporting.

What it looks like

A learner logs in and sees a "My Journey" timeline on their profile: every journal entry, community circle reflection, project milestone, and submission they've made, ordered chronologically from the day they joined. They can scroll their own arc from the first quarter to now, add new reflections directly to the timeline, and see how earlier entries connect to later ones. Consultants with permission can open a learner's profile and read the same continuous story instead of hunting across modules.

Why it matters

Learners currently receive information through multiple channels, including email, WhatsApp groups, classes, working groups, and direct communication. That makes it easy to lose track of what's happening or expected, especially if someone misses a session or doesn't regularly check WhatsApp. Learners have also asked for a clearer journey map so they can independently understand where they are, what they're building toward, and what requires their attention. NovaCanopy should become the place a learner can log into and immediately answer: What is happening? What am I behind on? What do I need to do next?

What it looks like

When a learner logs in, they should see a "Your Journey" or "What's Next" section prominently on their dashboard. This could include a checklist showing things like: the Responsible AI module due Friday, submitting a community journal, a working-group project currently in Phase 2, watching the recording from Tuesday's session, or a presentation due Sept. 30. Items could show statuses such as Completed, Due Soon, Overdue, or Upcoming. For working groups, the learner should also be able to see the current project, objective, stage, and next expected action. If they miss a session, they should be able to see the recording or resources as well as the key action or decision that came out of it.

Why it matters

The Grace Policy is already a live protocol, co-created by paNhari and the Talloires Network Care Team, but only half of it lives on the platform. Attendance and session records are already tracked in NovaCanopy; the grace side (requests, the Care Team log, and follow-ups after three consecutive missed sessions) still runs by hand in a separate rubric, over WhatsApp and email. With the movement growing past 150 participants in 2027, that manual coordination doesn't scale: every grace request tracked by hand is consultant and Care Team time spent on process instead of on people. Building it into NovaCanopy centralizes the process and directly reduces consultant workload, a coordination lift the whole partnership has been working toward. The platform already knows the attendance pattern, so it can surface who may need a check-in and hold the grace record in the same place, while protecting the learner: the system records that grace was requested and granted, never the private reason, unless the learner chooses to share it.

What it looks like

A learner selects "Request grace" on their profile or a specific deliverable, optionally adds context, and submits. They can see the status of their request, and while grace is active, their progress and certification deadlines pause rather than marking them as behind. The Care Team sees, with permission, that a request exists and its status, consistent with the protocol's ask-permission-first principle. This connects to existing attendance tracking and to the Certification epic, where an approved grace request lifts deadlines.

Why it matters

Today the movement's materials live wherever they were created. Resources are shared across many WhatsApp channels and over email, and WhatsApp groups are deleted or archived once a working-group phase ends, so anything shared there becomes hard to find. The result is a recurring coordination burden: every year, group members periodically reach out asking for information from past sessions, with coordinators or working-group coaches re-sharing links by hand each time a learner asks. The Resource Hub puts everything one login away and makes past materials easily navigable, cutting that re-sharing overhead. The analytics extension adds tracking of downloads, opens, and views, so the team learns which materials learners actually use and can shape the curriculum accordingly. Design decision from July 2026: NovaCanopy links to recordings rather than hosting video, keeping platform costs low.

What it looks like

A learner opens the Resource Hub, picks their working group, and sees each session with its recording link, slides, and materials. A learner who missed a session, or wants something from a past phase, finds it without asking anyone. Coordinators add materials once instead of re-sending them. Admins see which resources are used most and least.

Why it matters

Certification is what a learner walks away with. Today eligibility is worked out by hand at the end of each term, which is slow and hard to apply consistently as the movement grows past 150 participants in 2027. Automating it gives learners a credential they can show employers and universities, gives the team a verifiable record third parties can confirm without a manual request, and removes a recurring end-of-term administrative load from consultants.

What it looks like

A learner opens their profile and sees a progress bar for the current term showing what's done and what's outstanding: sessions attended, polls completed, materials read, deliverables submitted. As the term closes they take the gate check, a short multiple-choice check with feedback on any wrong answer and up to five attempts. Once they're at 90 to 100% and have passed, NovaCanopy alerts them that their certificate and badge are ready, and they download and share them. A learner on approved grace sees paused deadlines instead of an overdue flag and completes the outstanding work before moving to the next level. Consultants and admins see the same progress view across their learners.

Why it matters

NovaCanopy has no download option today, so each submission has to be printed page by page: slow, repetitive work that gets worse as submissions accumulate each term. A built-in download turns that into a single action and gives Project Managers and Consultants offline, shareable copies of learner work for review, reporting, and record-keeping.

What it looks like

Every page holding learner submissions carries a Download button offering Word or PDF. Before downloading, the user picks either a single submission or the full list.

Why it matters

The movement's evidence base is only useful if it can be analyzed. Today consultants read responses one at a time on screen, or re-type them into a spreadsheet by hand, which is slow and introduces transcription errors before analysis even starts. As the movement grows past 150 participants in 2027, with community journals, project proposals, and poll responses accumulating each term, that manual approach stops being viable. A one-click export gives consultants the underlying dataset in a form they can actually work with: filter, code, pivot, and chart in Excel, compare across terms and working groups, and pull the numbers needed for partner and funder reporting without a request to the engineering team. It also makes the movement's longitudinal data portable for external evaluation.

What it looks like

A consultant opens the dataset they need in NovaCanopy, community circle journals, community project proposals, poll responses, attendance, or collaborator submissions, sets the filters they want (working group, term, project, date range), and selects "Export." They choose Excel or CSV, and the file downloads with the filtered rows already in it, one row per response and a column per field. They open it in Excel and start analyzing immediately: no re-typing, no clicking through responses one by one, no request to the engineering team. Admins can export across all working groups; consultants see only the learners and projects they're permitted to.

Why it matters

Community and University Collaborators are how the movement scales. The For Youth, By Youth proposal to the Mastercard Foundation commits every returning learner to recruiting three collaborators for their Youth Participatory Action Research (YPAR) project, growing from 150 collaborators in year one toward 450 by year five, which makes collaborators the largest participant group in the movement. The proposal also commits to a tracking and evaluation system through which collaborators and local community organizations share their experiences and lessons learned. Today collaborators have no access to NovaCanopy, so that commitment is met over email or not captured at all, and their perspective is missing from the movement's longitudinal data. In 2027 alone, roughly 75 learners each working with two to three collaborators means 150 to 225 collaborator voices for the year. A collaborator form closes the gap without provisioning full accounts for external partners: they receive a link, respond, and their input lands in NovaCanopy alongside learner data, tied to the same projects and tracked over time.

What it looks like

A learner shares a secure form link with their own collaborators for the relevant project or poll; coordinators can also send links directly when needed. The collaborator opens it, responds without creating an account, and submits. The team sees collaborator responses in NovaCanopy connected to the right project. As an extension, the team will explore collecting collaborator input over WhatsApp, where collaborators already communicate, by integrating WhatsApp with NovaCanopy so responses land in the platform.

Why it matters

Learners currently start a module without knowing what it covers or what they're meant to walk away with, and the curriculum itself sits in consultants' own notes and slides rather than in the platform. That makes the content impossible to evaluate: no one can compare what a module promised against what learners say they got from it. Documenting each module in NovaCanopy sets the objectives out plainly before a learner begins, and gives the team a fixed reference to assess content against poll feedback, the evidence base for the movement's shift toward a more practice-based curriculum. It's also what certification depends on: a credential is only defensible if the objectives behind it are on record.

What it looks like

A learner opens a module (YPAR, for example) and lands on a page giving its overall summary, learning objectives, expected outcomes, and duration, so they know what the module covers before starting it. Below that, the topics are listed in order; expanding a topic title reveals its description and what it sets out to teach. Consultants work from the same structure in their portal, filling in and updating each field as the curriculum changes, and their edits appear on the learner-facing page immediately.

Why it matters

People feel motivated seeing their projects, or themselves, featured for others to see and appreciate. It's also an easy way to account for what's been done, through narrative and pictorial updates.

What it looks like

Learners log in to their home portal and see the rotating carousel of images coming into display, staying for a few minutes before the next one appears. Pictures aren't hard-coded; they're uploaded to the backend control panel, which can delete, replace, or add more.

Why it matters

NovaCanopy currently runs as two apps with separate logins, a staff side and a learner side. Anyone who opens the wrong one hits an error that looks like a broken account, and the team spends support time diagnosing what is simply a wrong-portal problem. In August and September 2026 alone this affected several users, including senior partner staff. With the movement growing past 150 participants in 2027, plus consultants, staff, and partner staff, a single entry point removes a whole category of support requests and makes access management one process instead of two.

What it looks like

Everyone signs in at one address. The system reads the account's assigned role (learner, consultant, staff, admin) and opens the right workspace with the right privileges. Users no longer choose between a learner login and a staff login; admins assign access levels when creating accounts.

Why it matters

Submissions are final once sent, so a learner who spots an error has no way to correct it, and every fix runs through an admin by hand. The volume of these requests makes it a recurring workload, and letting work return to draft keeps learner records accurate without the manual back-and-forth.

What it looks like

A learner opens a submitted piece of work and selects "Request edit," optionally noting what needs changing. On the admin side, every submitted item carries a "Return to draft" action, and the admin sees pending requests alongside it. Once returned, the item moves back into the learner's drafts with its content intact, the learner is notified, and they edit and resubmit as normal. The submission history records that the item was reopened, by whom, and when.

Why it matters

The team invests real time in producing session materials but has no evidence of whether learners use them. Usage data turns that guesswork into a curriculum signal: what gets opened and watched is kept and built on, what goes untouched is reworked or dropped, and it's the same evidence base used to show partners and funders how materials are actually consumed.

What it looks like

An admin opens a resource and sees its usage at a glance: downloads, opens, and recording views. Sorting the list by usage shows which materials learners rely on and which go untouched, so the team can drop or replace what isn't working.

Why it matters

Progress is currently invisible. Learners have no reliable way to see where they stand mid-term, and admins piece participation together from attendance notes and externally shared meeting links, so engagement is inferred rather than known. Automatic tracking makes it visible in real time: a learner sees what's done and what's outstanding, and admins can spot who is drifting early enough to act instead of after the term closes. It's also the evidence layer certification rests on, since eligibility is only defensible if activity and attendance are recorded by the platform. As the movement grows past 150 participants in 2027, that record is what keeps progress monitoring possible without adding consultant workload.

What it looks like

A learner opens their profile and sees a progress bar for the current term, with the milestones behind it listed as done or outstanding: sessions attended, polls completed, materials opened and downloaded, and deliverables submitted. Nothing is self-reported; the bar updates on its own as they use the platform, and their last activity and last seen are recorded automatically. Sessions open inside NovaCanopy through an embedded meeting room rather than an external link, so joining a session marks attendance without anyone taking a register. Admins and consultants open a single view showing every learner's progress side by side, sort or filter to see who is falling behind, and reach out mid-term instead of at the end. A learner on approved grace sees paused deadlines rather than an overdue flag, and the tracker keeps counting their completed activity as normal.

Why it matters

A certificate only means something if it reflects understanding. The gate check confirms a learner absorbed the main points of a term without subjecting them to a formal exam, and because every wrong answer comes back with an explanation, it teaches as it tests. Automatic marking holds that standard consistently across a growing cohort without adding consultant workload, and it gives the certificate a defensible basis when an employer or partner asks what it represents.

What it looks like

A learner opens the gate check at the end of a term and answers a short set of multiple-choice questions. There's no fail state: they have up to five attempts, and every wrong answer comes back with an explanation pointing to the correct one. On each new attempt the system reshuffles the answer options, so an answer that was option B the first time may be option D the next, meaning the learner has to recall the point rather than the position. Once they pass, the gate check clears and their certificate is released for the term.

Why it matters

NovaCanopy handled over 3,000 applications last cycle — 1,445 submitted and more than 1,600 left in draft — and with interest in the movement growing, this build stress-tests for over 10,000. Last cycle, real hours went to work that needed no human judgement: checking whether a video link was public, searching for an applicant's profile online, chasing incomplete submissions. This is the reference build behind the D11 engineering estimate in the 3rd cohort recruitment proposal, and it moves that checking work onto the platform itself. Nothing here is a rebuild of the core form; roughly half the hours go to testing before launch, keeping the form stable while applications are live, and covering the year-end break, all on a platform current learners are using at the same time. Building each piece is never just the code: it carries discussion of what it should do, testing, and user acceptance testing with learners themselves — last cycle five cohort learners tested the platform while engineers observed, and that practice continues here.

What it looks like

Applicant-facing (60 to 72 hours)

Video submissions get automatic detection of private or unreachable links at submission, with embedded playback so reviewers watch inside NovaCanopy. The welcome page is reworked so applicants can preview every step before committing to the attestation letter that stalled most of last cycle's 1,600 abandoned drafts, with the main site and the application form connected behind the scenes into one seamless experience. Applicants connect their LinkedIn profile directly for verification, a pre-form eligibility gate checks year of study and graduation date before the first real question (305 applicants from one university went unreviewed last cycle for exactly this reason), and small additions capture an AI/translation declaration, a cause-area field, and a retested student ID upload.

Automation (40 to 50 hours)

On submission, the system checks every required question is answered, flags duplicate submissions, tests every link, flags text that matches other applications or public sources, and generates a summary of long applications noting missed questions. Reviewers open a queue of only complete, reviewable applications. Merit scoring stays entirely human.

Reviewer-facing (94 to 118 hours)

A shared assignment view replaces the spreadsheet-and-memory system that left two leads scoring 400 applications themselves last cycle. Distressing-content flagging routes disclosures to the TN office or university liaison officer. Bulk email is rebuilt with templates and mail merge for reliable status communication. An analytics and equity dashboard tracks each university against its target range and visualizes equity indicators already captured in the form. A test environment with sample applications lets roughly 30 volunteer reviewers train without touching real applicant data, and a selection integrity and audit trail requires a recorded reason for any zero score and logs every promotion and demotion.

Quality, stability, and support (105 to 160 hours)

A dedicated Nov 24–30 test week covers every applicant and reviewer path on mobile and low-bandwidth connections under stress-test load, since most applicants apply from phones. Technical oversight runs through the October–November build against an Oct 30 sign-off baseline. A Dec 1–15 reserve covers fast bug fixes after the freeze while real applicants are in the form, and a named on-call rotation with automated alerts covers Dec 16 to Jan 4 while the team is otherwise offline.