5
Stakeholder Interviews
Seven weeks of listening, watching, and mapping how Terra Rossa actually works — synthesized into the themes, quick wins, and the recommended route forward.
The transcript summary and key takeaways from the discovery review call, captured for your records. Everything below is a synthesized snapshot from the Fireflies transcript, not the raw transcript itself.
The core focus is to reduce bottlenecks caused by Cherrelle's central role and to introduce a project management layer that distributes decision-making and execution.
The meeting emphasized embedding AI and digital tools to automate routine tasks, improve information access, and support decision-making.
A strong, autonomous team culture with a need to clarify roles and support smoother communication.
Terra Rossa's work is exceptional. Its system is a person. Every decision, file, confirmation, and client answer routes through Cherrelle — which worked at one scale, and now quietly caps the next one. The good news: the culture is strong, the team is loyal, and the fixes are structural, not personal. Every recommendation in this recap was held to one test: does it move context out of Cherrelle's head and into a system her team can act on?
Design thinking, layered with the Fifth & Found framework: hear from every person doing the work, then watch the work actually happen.
This is the single most important picture from discovery. Toggle between how work moves today and how we design it to move.
Every arrow is a wait. Jai checks with Cherrelle before checking Google Drive. Ronalyn holds tasks until confirmation. Jean Paul consults on pricing and chases status. If Cherrelle is sick, traveling, or heads-down on a shoot — the company pauses.
Friction is what slows them. Flow is when they do their best work. Protect is what we must not optimize away.
I can't be waiting for my forecasting for the year because you're sending me a template — when I could literally just do it with AI.
I tell her all the time — Cherrelle, we need a project manager.
Stuck on one to two items most weeks — waiting on confirmations that arrive when Cherrelle is off the shoot.
There's days where I feel like I'm really productive... and then there's days where I feel like, should I be doing something?
We can't rely on memory — we need notes and citations.
Maybe I just don't have to do as many check-in emails... maybe the updates are more seamless, summaries already built in.
These showed up independently across interviews. When five different people describe the same shape from different angles — that's the system talking.
Every stakeholder routes to Cherrelle before acting. Files, confirmations, pricing, client answers. It's how the trust was built, and it's now the growth ceiling. If she's unavailable, work pauses. The design goal is not less Cherrelle — it's Cherrelle only where she's irreplaceable. And the highest-leverage fix costs nothing: a decision-rights framework — pre-approved templates Ronalyn sends without sign-off, a defined no-confirmation-needed category, and a daily 10-minute mobile approve/reject block for everything else.
Jean Paul named it outright; Jai independently recommended a PM system on May 11; Cherrelle formally pivoted the engagement to project management on May 12. Three arrivals at the same conclusion in 48 hours — plus a founder-defined success bar: glanceable confidence. Nobody should need effort to trust what the tracker says. This is the single highest-leverage structural fix — part human role, part software, part ritual.
One well-maintained calendar answers Alana's #1 question (where was Cherrelle and why — the difference between a business meal, a travel meal, and an on-set expense), powers Ronalyn's scheduling autonomy, and shrinks confirmation loops. Shoot dates, travel, and client work reliably on the calendar means the whole team self-serves context passively. This elevates calendar hygiene from housekeeping to structural infrastructure.
Duplicates, inconsistent naming, no versioning conventions, no knowledge base structure. The Drive holds institutional knowledge without making it findable — so people bypass it and ask Cherrelle instead (see Theme 01).
The NY Women's Forum project lost roughly a week because the client's decision appetite was discovered mid-edit instead of asked at kickoff. One intake question — "how involved do you want to be in the story?" — plus a white-glove default converts this from recurring risk into a service differentiator.
Decisions and priorities travel by email threads and text messages. Jean Paul flagged "lost in translation" across threads; Jamila works by email and text with no shared status surface; tasks reach Ronalyn in Google Chat and get manually re-keyed into the tracker. Everyone performs manual check-ins nobody enjoys — the exact task all of them volunteered to hand to AI.
~200 uncategorized transactions, receipt capture is manual, a forecast template never landed so Cherrelle built her own, and she's openly questioning the retainer's value. The relationship with Alana is strong — the pipes into her are what's broken.
The team is fast, witty, and allergic to friction. Anything with two extra steps gets abandoned. Jean Paul: easy + app-friendly or nothing. Jamila: no new tools for short engagements. Ronalyn: needs visual, hands-on onboarding. Meanwhile AI fluency runs from Cherrelle (stacked prompts) to Jamila (principled skeptic — consent, energy, data mining). Jai: "if the tool requires a manual to learn certain features, that's a lot." Together with Cherrelle's glanceable-confidence bar, that's the evaluation rubric — intuitive-or-nothing. Every tool choice must clear the friction bar and the ethics bar, publicly and credibly.
Low lift, high signal. Each one attacks a named theme and builds trust in the bigger system before we ask anyone to change how they work.
Pre-approved email + proposal templates Ronalyn sends without sign-off, a no-confirmation-needed list, and a daily 10-minute mobile approve/reject block for the rest. Costs nothing. Unlocks everything.
Shoot dates, travel, and client work reliably on Cherrelle's calendar — shared read access for Alana and Ronalyn. Expense classification becomes nearly automatic; no labeling task added to Cherrelle's plate.
Dropbox Sign / PandaDoc replaces the Doc → PDF → email → chase cycle. Removes a known delay from every single shoot, starting with the next one.
Auto-capture, summaries, and action items on client and internal calls. Retires most check-in emails and starts building the notes-and-citations record Alana asked for.
Alana pulls statements directly instead of requesting them; one AI-assisted session clears the ~200-transaction backlog with her verifying. Then receipt auto-capture keeps it at zero.
One naming + versioning convention (Client_Project_Asset_vX_Date) with a 15-minute walkthrough. Kills duplicates going forward, even before the full Drive restructure.
One intake question — "how involved do you want to be in shaping the story?" — with a white-glove, we-decide default. Prevents the next string-out week.
A recurring team touchpoint on the calendar (Jai's #1 ask) and one intake rule: if it's a task, it lives in the task system, not Chat. "Where are we?" always has an answer.
Making room for the work that only humans can do — that's the whole point. AI takes the routing, chasing, and formatting. People keep the judgment, the relationships, and the story.
Three routes were on the table. We're recommending the one that sequences trust before tools.
Decision rights matrix, calendar hygiene standard, nomenclature, Drive restructure, knowledge base architecture — drafted with Cherrelle.
Fireflies live on all calls; e-signature live on deal memos; read-only bank feeds + receipt capture piloted; backlog cleared.
Contract a PM for a few months to implement the PM layer — platform setup, locked cadence, status ritual — and drive change management and adoption. Jean Paul's PM load visibly transfers here.
Proposal engine (Ronalyn), Claude Project (Cherrelle), scheduling AI, client-facing ethics language.
User testing with each stakeholder, visual onboarding, SOPs into the knowledge base, and the fractional PM decision — made from data.
Buy the full AI stack now, roll it out at once. Fast on paper, but it violates the team's adoption physics: friction-averse contractors, one visual learner, one principled skeptic. High abandonment risk.
Bring in a PM on day one and let them build systems from scratch. A PM inheriting an unsearchable Drive and undefined decision rights becomes a second router, not a fix. Route C still hires the fractional PM — at Step 3, once the spine exists for them to run.
Nothing from discovery gets orphaned. Each finding maps to a named Phase 2 workstream — and the biggest fix needs no new tool at all.
| Discovery Finding | Phase 2 Workstream | Alignment |
|---|---|---|
| Approval bottleneck RONALYN · ALANA · JEAN PAUL | Decision-rights framework + async approval workflow | Process design — no new tool required |
| Calendar as context layer ALANA · RONALYN · JAI | Calendar hygiene system + team visibility | Reclaim.ai + Google Calendar sharing |
| Ad hoc task intake RONALYN · JEAN PAUL | Task management migration + intake rules (if it's a task, it lives in the system, not Chat) | PM platform: Asana vs. Monday, pros/cons mapped to needs |
| Proposal drag RONALYN · CHERRELLE | Terra Rossa voice prompt library + proposal engine | Custom GPT or Claude skills — team access |
| Expense context questions ALANA · CHERRELLE | Financial workflow redesign — passive capture over participation | Xero + Dext, read-only bank access, calendar visibility |
| Deal memo signature chase RONALYN · EVERY SHOOT | Production workflow upgrade | E-signature (Dropbox Sign / PandaDoc) |
| Drive sprawl + memory gaps WHOLE TEAM | Knowledge base standards + single source-of-truth index | Claude, Scribe, or Notion evaluated + Drive conventions |
| Client involvement risk JEAN PAUL · CHERRELLE | Client journey map + kickoff expectation tiers | Intake redesign (Phase 2 journey work) |
Mapped to interview pain points, not shiny objects. Each one clears the team's friction bar (easy, app-friendly, minimal steps) and the ethics bar.
| Add-On | Solves | Est. Cost* | Priority |
|---|---|---|---|
| Custom Claude Skill + Terra Rossa Claude Project FOR CHERRELLE · BUILT BY FIFTH & FOUND | A dedicated Claude workspace loaded with brand voice, project history, SOPs, and grant/application patterns — the "Terra Rossa Brain" she asks before she asks anyone. She's already curious about Claude; this is the on-ramp. | Claude Pro ~$20/mo + F&F build (scoped in Phase 2) | NOW |
| Custom GPT — "Proposal Studio" FOR RONALYN · MEETS HER WHERE SHE IS | Ronalyn already lives in ChatGPT. A custom GPT with master templates + winning proposals + voice rules turns her anxiety point into her superpower. Cherrelle approves instead of writes. | ChatGPT Plus ~$20/mo (custom GPTs included) | NOW |
| Fireflies.ai — Paid Plan WHOLE TEAM | Meeting capture, summaries, action items on every call. Retires manual check-in emails; feeds the status ritual automatically. | ~$10–19/seat/mo | NOW |
| Dext (receipt capture) → Xero-side stack CHERRELLE → ALANA HANDOFF | Photo a receipt, it lands captured and pre-categorized — paired with read-only bank feeds so Alana pulls statements instead of chasing them. Ends the backlog permanently. Passive over participatory, exactly as she asked. | ~$25–35/mo | NOW |
| E-Signature — Dropbox Sign or PandaDoc DEAL MEMOS · CREW · EVERY SHOOT | Replaces the Doc → PDF → email → chase cycle on deal memos. Crew signatures are a named delay point; this removes a chase cycle from every single production. | ~$15–25/mo | NOW |
| PM Platform — Asana vs. Monday (pros/cons in Discovery & Audit) PM LAYER · TEAM | One shared, glanceable status surface with AI digests — Jean Paul's "just tell me what's done" vision and Cherrelle's cross-project dashboard, including her independent studio work. Correction from May 12: nobody on the team runs Asana today (Jai used it consulting; Cherrelle found Monday more intuitive elsewhere) — so the recommendation ships as a pros/cons evaluation weighted to working knowledge, judged on the intuitive-or-nothing rubric. Intake rule ships with it: tasks live here, not in Chat. | ~$10–14/user/mo | NEXT |
| Reclaim.ai CHERRELLE + JAI | Elevated from nice-to-have to structural: a defended calendar is the master context layer — it answers Alana's expense questions, powers Ronalyn's scheduling autonomy, and locks Jai's cadence. | ~$8–12/user/mo | NEXT |
| Fifth & Found — Claude Systems Build ONE-TIME BUILD · F&F SCOPED SERVICE | Design, build, and hand off the Terra Rossa Claude environment: custom Claude Project, team-access skills, the voice-true proposal prompt library, and try-first training sessions per person. Includes 30 days of post-launch tuning. | $2,500 one-time (scoped with SOW edits) | NOW |
| AI Agent Builds POTENTIAL SCOPE ADDITION · FLAGGED MAY 12 | Autonomous agents (status digests, intake routing, financial pre-categorization) sit outside the original tool-identification scope. Fifth & Found will price options after observing actual workflows — no agent gets built before the spine it would run on. | Priced post-observation | EVALUATE |
| Notion AI — Knowledge Base Layer TEAM · DECISION POINT | Structured, searchable home for SOPs and institutional knowledge. Phase 2 decision point: Claude vs. Scribe vs. Notion AI vs. a disciplined, restructured Google Drive (the team already lives in Google Suite — lower friction may win). Scribe auto-documents workflows as they happen — strong fit for try-first learners. | ~$10/user/mo add-on | EVALUATE |
| Frame.io + Jump Desk PRODUCTION · ALREADY IN STACK | Keep. Jean Paul's review-and-remote workflow already works — we integrate around it, not over it. | Current spend | KEEP |
*Estimates as of July 2026 — pricing shifts often; we verify current rates before purchase and pilot each tool in the Fifth & Found Test Lab lens: does it perform equitably, is the energy cost defensible, does it clear the team's friction bar.
Every stakeholder spoke of this team with warmth. Systems should protect the relaxed, high-trust style Jai named — not bureaucratize it.
Ronalyn keeps email. Jean Paul keeps executive edit judgment. Alana keeps the financial relationship. We build under people, not over them.
Terra Rossa tells grounded, real stories about people of color and community. Every AI tool adopted must be publicly defensible — Jamila's ethics lens becomes house policy.