About SignalDesk

Inspiration

SignalDesk was inspired by a very specific workflow problem: small service businesses often do not lose opportunities because they cannot find leads, but because the context around those leads gets scattered across tabs, notes, inboxes, spreadsheets, and memory. By the time someone is ready to act, the moment is often gone. That problem showed up especially clearly in founder-led sales, where the same person is doing research, qualification, follow-up, and delivery. :contentReference[oaicite:0]{index=0}

While looking at marketing and consulting workflows, I found the same underlying failure mode across both: research was happening, but it was not reliably turning into the next action. Existing tools handled storage, sending, or recordkeeping, but not the narrow layer between “I found something important” and “I know what to do next, when to do it, and what happened last time.” That became the core insight behind the project. :contentReference[oaicite:1]{index=1}

The project thesis became:

$$ \text{capture context} \rightarrow \text{suggest next action} \rightarrow \text{time follow-up} \rightarrow \text{log outcome} \rightarrow \text{keep momentum} $$

Instead of building another CRM, I wanted to build a browser-first follow-up system that fits how founder-led service sales actually happen. :contentReference[oaicite:2]{index=2}

What I Built

I built SignalDesk, a Chrome browser extension designed to turn early-stage sales research into timely next actions without forcing heavy CRM behavior. The extension lets users capture pages, snippets, contacts, trigger signals, and notes while browsing, attach them to account dossiers, and then manage follow-up through reminders, outcome logging, and a daily action queue. :contentReference[oaicite:3]{index=3}

The core product experience is intentionally narrow:

quick capture while browsing
dossier-based organization
an action queue as the home screen
reminders and browser notifications
one-tap outcome logging
optional AI summaries and next-action suggestions

A key design decision was that the home screen is the action queue, not the capture inbox. That reflects the project’s main belief: action matters more than storage.

I also designed the product to support two adjacent user types with the same core workflow but different overlays:

Marketing users, who need trigger freshness and prioritization
Consulting users, who need qualification guardrails and disqualification confidence

That meant keeping one shared architecture while letting prompts and fields vary by profile.

How I Built It

Technically, I built the project as a Chrome Manifest V3 extension with a side panel UI, context menu / keyboard shortcut capture, and a background service worker for reminders, notifications, and scheduled tasks. Data is stored locally using IndexedDB, while lightweight settings use chrome.storage.local. :contentReference[oaicite:6]{index=6}

The architecture was chosen around a few product constraints:

The project had to feel fast enough for daily use.
It had to work without requiring a cloud account in V1.
AI needed to be optional, assistive, and never the only way the product worked.
Trust had to be preserved through visible prompts and explicit user confirmation.

The AI layer was designed carefully. It can summarize dossier context, suggest next actions, suggest reminder timing, draft outreach copy, and propose dossier assignment or merges — but it does not send messages, silently merge records, auto-disqualify prospects, or invisibly change states. The idea was to keep the human in control at every important decision point.

What I Learned

The biggest lesson from building SignalDesk was that the hard part is not adding AI — it is designing behavior. A lot of software in this space becomes a prettier backlog: users save information, but they do not return to act on it. This project taught me that the real test is not whether capture works, but whether the system becomes a daily habit.

I also learned how important scope discipline is. It was tempting to drift toward CRM features, outreach automation, analytics, or collaboration, but those features would have diluted the product’s wedge. Building this extension forced me to think more clearly about what the product is not:

not a CRM replacement
not an autonomous SDR
not a sending engine
not a general note-taking tool

Another important lesson was that trust is a product feature. In a workflow like sales follow-up, silent automation can easily feel wrong. That is why so many core rules in the project are about user control: no silent merges, no automatic sending, no hidden state changes, and no destructive automation without confirmation. :contentReference[oaicite:11]{index=11}

Challenges I Faced

One major challenge was avoiding feature sprawl. The problem space naturally pulls toward building “just a little CRM,” but that would have weakened the product. I had to constantly narrow the scope back to the core loop: capture, decide, remind, log, and continue.

Another challenge was balancing usefulness with trust in the AI features. If AI suggestions are too weak, they do not help. If they are too automatic, they reduce confidence. The solution was to make AI assistive, editable, and fully inspectable rather than authoritative. :contentReference[oaicite:13]{index=13}

A third challenge was designing for both marketing and consulting without making the product too broad. These segments share the same structural problem, but they fail in different ways: marketing struggles with noise and stale signals, while consulting struggles with scarcity and under-qualification. The challenge was to support both without breaking the core workflow. :contentReference[oaicite:14]{index=14}

Finally, there was the technical challenge of building an extension that feels lightweight while supporting a timeline-heavy workflow. Features like dossiers, reminders, outcomes, notifications, local persistence, and queue logic required an architecture that was more structured than a simple clipper, but still lean enough for a V1 browser extension. :contentReference[oaicite:15]{index=15}

Why This Project Matters

SignalDesk matters because it focuses on a neglected but high-value problem: the gap between research and action in founder-led service sales. Instead of trying to automate the whole sales process, it improves one small but critical loop. If a user captures something meaningful and acts on it while the signal is still warm, that can directly affect revenue.

In that sense, the success of the project can be described simply:

$$ \text{Value} \neq \text{number of things saved} $$

$$ \text{Value} = \text{timely actions taken on the right opportunities} $$

That idea shaped both the product strategy and the engineering choices behind the extension.