Partner operations · IT professional services · Applied AI

PartnerSignal

A workspace for the partner side of IT services. It finds which reseller partner to call, turns a phone call into a qualified opportunity with AI, brings in the right specialist, and keeps every follow-up on time. I built it to explore how technology channels sell services, and where AI helps without taking the judgement away. It is live, and you can try it in two minutes.

Try the guided demo See what it does 7-step tour · no sign-up · fictional data

Recorded from the live app: ranked day, call capture, specialist handoff, reviewed outreach, pipeline.Open it yourself

The workflow

The everyday work of partner services, each part working end to end.

Selling services through reseller partners comes down to a few repeatable jobs: know who to talk to, understand the need, bring in the right expert, follow up, and keep an honest pipeline. PartnerSignal has a working feature for each.

The workFinding which partners to talk to

Know exactly who to call, and whyA signal feed built from renewals, products reaching end of support and purchase patterns, plus a map of the services each partner has never sold.Signals · Whitespace map

The workUnderstanding the need and qualifying it

Turn a phone call into a qualified opportunityPaste rough notes; the AI assistant returns the customer's problem, the best-fit service and budget, authority, need and timeline, quoting the notes as evidence.Capture a call · BANT

The workWorking with technical specialists

Hand off to the right specialist in one clickRoutes the deal to the least-busy specialist in that service line and writes the handoff brief: context, challenge, scope, open questions.Specialist routing · Handoff brief

The workPersonal outreach and reliable follow-up

Personal outreach, with a person in controlDrafts e-mails that explain which record each line came from. You edit, tick a review checklist, then log it. Every touch sets the next follow-up date.AI drafts · Follow-up cadence

The workKnowing the services, and where AI fits

A playbook for every service lineWhat to listen for, what to ask, what to offer, and which AI use cases to raise, across cybersecurity, cloud, networking, AI, lifecycle and training.Playbook

The workKeeping an accurate pipeline

A pipeline that can't be fakedA deal can't move to Qualified without the evidence. Weekly reports show pipeline, win rate and follow-through, with a CRM-ready export.Pipeline · Reports

7workflows, from prospecting to reporting

18reseller partners across Canada

36prospect signals with deadlines

6service lines, 17 services

0messages ever sent automatically

39automated tests, live on the web

AI with judgement

The AI does the busywork. A person makes the calls.

Call notes are structured by the Claude API when a key is configured, or by a built-in rules engine that returns exactly the same format. Either way, every judgement quotes the notes it came from, can be changed before saving, and nothing is ever sent automatically.

Call notes with evidence highlighted beside the structured qualification — **Capture a call.** The evidence behind each field is highlighted in the original notes.

Outreach draft with the reasons behind each line and a review checklist — **Outreach.** Each line says which record it came from; logging it requires the review checklist.

Decisions you can explain

A priority score that shows its working.

Every conversation gets a 0–100 priority made of four parts anyone can check: service fit, qualification, momentum and urgency. It ranks where attention should go first; it never pretends to be a probability of winning.

Opportunity workspace with the score breakdown, qualification editor and specialist routing — **Opportunity.** The reason for every point, the qualification editor, and one-click specialist routing.

Pipeline board with stage columns and priority bars on each card — **Pipeline.** Drag a card forward and the stage gate checks the evidence first.

Built like real software

A small production service, not a mock-up.

A Python API with a React interface, tested on every change and deployed automatically. It reseeds each morning so the demo always has something due today.

Application

FastAPI + React

Typed API, SQLAlchemy data model, and a TypeScript interface with a guided tour, keyboard support and a mobile layout.

PythonFastAPIReactTypeScript

Claude API, safely

Structured outputs validated against a schema, a fallback engine with the same format, rate limits, and evidence quoted for every judgement.

Claude APIPydantic

Quality

Tested and shipped by CI

39 automated tests, lint and type checks on every push; a Docker image deployed to Render and kept warm during business hours.

pytestVitestDockerGitHub Actions

Two minutes, no sign-up

Take the guided tour of a partner-services day.

Seven steps, from finding the right partner to reporting on the week. All companies and people are fictional.

Try the guided demo Source on GitHub