PriorityIQ — Product Design Spec

Author: Ella Nalbandyan, Campana & Schott

Date: April 9, 2026

Status: Ready for Build Day

Build Day Team: TBD (1 builder + 4 non-builder roles)

1. Problem Statement

Portfolio prioritization sounds simple until you're actually doing it. Here's what makes it hard in practice:

1. Portfolio managers spend 4–6 hours per cycle manually scoring requests in Excel — and then redo it when someone moves the goalposts after the meeting. There's no single source of truth, so every cycle starts from scratch.
2. Leadership overrides the ranked list with gut-feel decisions. This happens constantly, and it kills trust in the process. The people who built the scoring framework stop taking it seriously because they know it doesn't actually drive decisions.
3. Requestors don't get a clear answer on why their item wasn't prioritized. So they re-submit it through a different channel, or escalate, or just show up frustrated at the next planning meeting. Backlog inflation is often a communication problem, not a volume problem.
4. Consultants rebuild this framework on every engagement. Same criteria debates, same Excel template, same two days of setup work — every single time.

The real issue underneath all of this: prioritization is treated like a one-time output from a meeting instead of an ongoing, documented process. Without something systematic holding it together, every ranking is just a reflection of whoever had the most political capital that week. Clients don't need a better spreadsheet. They need a process that's defensible — one where anyone can see how a decision was made and why.

2. Product Concept

One-sentence pitch

PriorityIQ scores portfolio intake requests against a defined set of weighted criteria, ranks them automatically, and gives every stakeholder a plain-English explanation of where their item landed and why.

How it works

• The portfolio manager sets up scoring criteria once at the start of the engagement — things like strategic alignment, estimated ROI, effort, risk, dependencies — and assigns weights. Takes about 10 minutes. Doesn't change unless they want it to.

• Requestors submit items through a short structured intake form. No paragraph text boxes. Just the fields that actually matter for scoring.

• PriorityIQ runs each submission through the scoring engine, flags anything that's missing, and produces a ranked list with a one-sentence rationale per item explaining the score.

• The portfolio manager reviews, can override with a logged reason, and publishes. Everything is auditable — who changed what, and when.

• Stakeholders get a read-only view of the ranked backlog with the scoring logic visible. The 'why wasn't my thing prioritized' conversation becomes a 30-second check instead of a 30-minute meeting.

The strategic play

The criteria framework you build on engagement one doesn't disappear. It becomes the starting point for engagement two. Over time, C&S builds a library of proven scoring models by industry — pharma, life sciences, medtech — that competitors can't replicate without the same engagement history behind them. Post-engagement, clients pay a monthly subscription to run prioritization self-serve between cycles, so they're not calling a consultant every time a new request comes in.

3. Target Buyer

• Primary buyer: VP or Director of Portfolio Management, or Head of Digital PMO. They own the backlog, own the stakeholder relationships, and are the ones who have to explain to someone's frustrated exec why their project didn't make the cut. The ranking process is their problem to defend.

• What they do today: A manually maintained Excel scoring sheet they prep for days before a quarterly planning meeting, then spend hours defending, then spend weeks cleaning up the fallout from. All in, it's 3–5 FTE days per cycle — not counting the relationship damage.

• Why they would pay: PriorityIQ cuts the prep time, removes the subjectivity, and gives them a documented audit trail. That last part is career protection. One cycle where they don't have to walk an exec through a 2-hour prioritization debate pays for months of subscription.

• Secondary beneficiary: The people submitting requests — project managers, workstream leads. They get visibility into why their item ranked where it did and what would need to change to move it up. Fewer re-submissions, fewer escalations, less noise.

4. Architecture

Input Layer

Two entry points: a structured intake form for requestors to submit new items (title, description, business sponsor, estimated effort, strategic objective, dependencies), and a criteria config panel where the portfolio manager sets up and weights the scoring dimensions. The UI is a clean web app — shareable via link, no login required for the demo.

Transform Layer (Core IP)

Each intake submission gets passed to Claude with a system prompt that acts as an anchor — same behavior, same scoring logic, every time, regardless of how the request was written. The prompt embeds the portfolio manager's criteria and weights and asks the model to score each dimension 1–5 with a one-sentence rationale. Temperature is set low (0.2) so scores are consistent and repeatable. A lightweight RAG layer pulls in the client's actual strategic objectives so alignment scores are grounded in real priorities, not generic language. The weighted composite score is calculated in code — not by the model — so the math is always transparent and auditable.

Display Layer

Three views: a ranked backlog table with composite scores, dimension breakdowns, and AI rationale per item; a criteria dashboard showing how each scoring dimension is weighted and contributing to rankings; and a stakeholder read-only view — a clean, shareable link with the ranked list and plain-English reasoning, no edit access. All views update in real time.

Tech Stack (Hackathon)

• Frontend: React + Tailwind CSS, single-page app

• Backend: Node.js / Express, deployed to Railway or Render

• AI: Claude API (claude-sonnet-4) with structured prompt templates

• Database: SQLite — fast to set up, zero infra overhead

• Auth: None for the demo — shareable link model

5. Hackathon Execution Plan

Timeline

Critical risk and fallback

The scoring engine is the biggest risk — getting Claude to return clean, structured scores with rationale in a consistent format can take longer than expected to tune. If it's not working cleanly by the 2:30 mark, hardcode 3–4 pre-scored example items with realistic rationale and demo the ranked view and stakeholder link as a working prototype. The pitch still holds — judges care about the concept and the UI, not whether the API call fires live on stage.

Role allocation

6. Defensibility & Competitive Moat

Why a client would pay — and keep paying

• Switching cost: The scoring criteria and historical rankings built in PriorityIQ become institutional memory. Leaving means losing the audit trail and re-litigating every prioritization decision from scratch with a new tool.

• Data flywheel: Every scored item helps calibrate what 'high priority' actually looks like for that client. After 2–3 cycles, the model's initial scores reflect real decision patterns — something a new tool or a new consultant can't replicate quickly.

• Process legitimacy: Once a team has gone through one cycle with an auditable, defensible output, going back to gut-feel Excel is politically impossible. The tool doesn't just support the process — it becomes the process.

Pricing model

Per-engagement license: $8,000–$15,000 for a 6-month term, bundled into the SOW. Anchored to the value it replaces: 3–5 FTE days per prioritization cycle × 4 cycles × ~$1,500/day blended rate = $18,000–$30,000 in consultant time displaced. PriorityIQ captures roughly 40–50% of that. Post-engagement SaaS: $2,500/month for self-serve access between engagements. One avoided re-engagement pays for a year.

Competitive positioning

Closest analogue: Productboard — but built for consulting portfolio management instead of product roadmapping, with AI scoring replacing manual effort.

7. Success Criteria

Hackathon demo (April 17)

• ☐ A new intake item is submitted live during the pitch

• ☐ Claude scores it and returns a structured result within 10 seconds

• ☐ The ranked backlog updates in real time

• ☐ The AI rationale is specific and readable — not generic filler

• ☐ The stakeholder read-only link opens cleanly on a phone or second screen

• ☐ The criteria weighting panel is shown and explained

"Would a client pay for this?" test

• Does it solve a pain the buyer feels every single quarter? Yes — prioritization cycles are recurring and universally frustrating.

• Is the ROI math obvious in 30 seconds? Yes — consultant time displaced vs. subscription cost is a straightforward comparison.

• Would the buyer be embarrassed to show their current process next to this? Yes — an Excel sheet with gut-feel scores next to an auditable AI ranking is an uncomfortable contrast.

• Does it get better with use? Yes — the scoring calibrates to the client's actual decision patterns over time.

• Can a non-technical buyer follow it in a 4-minute demo? Yes — the UI is simple and the output is plain English.

8. Open Questions

1. How opinionated should the default criteria be? Do we ship with a pre-built pharma/life sciences scoring template, or make the portfolio manager configure from scratch? Pre-built is faster to demo but might feel too generic to a buyer who thinks their situation is unique.
2. What do we do when leadership overrides the ranking? Do we require a written reason? Optional feels toothless. Required creates friction. This is a UX and change management call that needs a real answer before anyone would actually pay for this.
3. What's the minimum viable intake form? Too many fields and people stop filling it out. Too few and the scoring is meaningless. What's the 5-field version that still produces a defensible score?
4. Does post-engagement SaaS cannibalize re-engagement revenue? If a client can run prioritization self-serve for $2,500/month, do they ever call us back? We need to be deliberate about what stays in the tool vs. what still requires a consultant in the room.
5. What's the data governance answer? Portfolio items contain sensitive strategic information. Before any client conversation, we need to know where data lives, how long it's retained, and whether it touches a public model. This will come up immediately.