Not a pitch deck. A working plan - the systems I'd stand up, the accounts I'd prioritize, and the thinking behind every step.
Built To Scan Multiple Markets At Once.
A Four-Part Scoring Model For Every Account.
Finding the right accounts shouldn't depend on someone manually searching city by city, and deciding who to prioritize shouldn't be a gut call either. Here's how I'd build both.
Fit
Does this account's size, structure, and day-to-day operation actually match where this product creates real value?
Reach
Is there a clear path to the right person, or does the work start with finding out who that even is?
Pain
Is there a specific, visible problem right now - not a guess, something you can point to?
Timing
Is there a signal suggesting this is the moment, or is patience the better play?
An account scoring high across all four gets the fastest, most direct path. One that's a strong fit but early on timing still gets attention - just a different kind, built around staying visible until the timing catches up.
The AI Prospecting Engine
This is the core system behind that idea - a working engine that runs the discovery, research, and scoring automatically, so the right accounts surface on their own instead of waiting to be found.
Pull businesses matching the target profile across every relevant market. Pull what's publicly available for each one. Run every account through the framework above, automatically. Surface anything that scores high, with the reasoning attached. Land a daily summary sorted by priority, so the day starts with the right accounts already in front of the team.
The people this reaches aren't looking for a platform with more features - they're looking for their evenings back. That means the language has to be concrete, not abstract.
Draft fast, edit slower.
Use tools to produce several message directions quickly - then every version gets a real edit pass before it goes anywhere.
Test small before scaling.
New messaging goes to a limited slice first. If it underperforms, that's worth knowing on 200 sends, not 20,000.
Match the channel to the person.
Senior roles get a lighter, relationship-first touch. Operational roles respond better to something more direct.
Personalization isn't all-or-nothing.
A small set of top accounts earn fully custom messaging. A broader tier gets persona-based messaging with one real, specific detail - enough to stop it from reading like a mass send.
Getting tone right starts before any sequence gets written - studying how existing content already talks to this audience, noting actual word choices, not just messaging points. A small set of drafts gets a gut check before a full sequence goes out, and reply rate by variant in the first week tells its own story - underperformance is usually a tone miss, not a targeting miss.
The risk with AI-assisted drafting is defaulting to a template instead of a real account. The fix is specific input - real research feeding the first draft - and a dedicated review step checking tone and specificity before anyone sees it. If it reads like it could've been sent to any company, it gets sent back automatically.
Fast systems only help if they're accurate and they don't fall over.
One source of truth.
If two systems are both tracking the same account, they need a clear rule for which one owns which piece of information, plus a regular check that catches drift before it becomes a leadership-meeting problem.
Built to fail safely.
Before sending anything to a large list, test on a small slice first. If something breaks mid-send, know exactly what already went out versus what didn't, pause cleanly, fix the cause, and resume from the last clean point.
Reading the pipeline, not just the funnel.
Track how long things sit at each stage and flag anything stalling past a threshold. A stall between two early stages usually signals a definition mismatch, not a lead-quality problem - and a high-scoring account stalling as much as a low-scoring one means the scoring itself needs sharpening.
What Leadership Needs To See
One question answered clearly: is this working well enough to justify the investment, and if not, what's changing.
What The Team Needs To See
Something more actionable: which accounts are responding right now, so effort goes where it's working.
A function that doesn't exist yet isn't a blank page - it's a chance to build something trusted from the start, instead of retrofitting trust onto something built in isolation.
Listen First
Build The Logic In The Open
Pilot Small
Scale What Holds Up
Tie It Back To Revenue
Before any scoring model exists, the first conversations are with sales, RevOps, and whoever's closest to customer success - what's working, where leads get lost, what people wish existed. Every scoring decision gets shown to the people using it before it's finalized, using real accounts. The first sequence runs against a narrow segment, and results get reported honestly, including what isn't working. Once a pilot shows signal, it scales, and documentation gets built alongside it. Every layer ties back to pipeline and how it converts compared to other channels.
With Sales
Bringing data into every conversation, not just plans, and asking for a read on account prioritization before anything's finalized.
With RevOps
A clear rule for which system owns which piece of shared account information, plus a regular check for drift.
With Customer Success
The same signals that flag a strong-fit target account are early indicators worth sharing forward into expansion and retention conversations.
On competing priorities
The deciding factors are always revenue impact and time sensitivity, kept visible so people can see the tradeoff, not just receive a decision.
Moving fast only helps if the people around it are moving with it. When a pace gap shows up, naming it plainly beats pushing through it or letting it build into frustration.
The events that work best in this market aren't the biggest ones - they're the ones where the right people are actually in the room.
Regional Operator Roundtables
Small, twelve to twenty operators, run close to where they work. Lower cost, higher trust, easier to fill without a paid list.
Roadshow Circuits Tied to Compliance Calendars
Timing a multi-city circuit around known industry moments means showing up exactly when the audience is already thinking about the problem.
Co-Hosted Sessions With Adjacent Partners
Sharing a room with a business that already has trust with the same audience solves the attendance problem before it starts.
Filling The Room.
Ranked list of likely attendees, scored on fit and signal before invitations go out.
In The Room.
Briefed reps, fast in-room capture, real reasons each attendee matters.
After The Room.
Every conversation routes into the same scoring and sequencing system the next day.
Scanning public industry directories and association listings city by city, cross-referencing against the same fit-and-signal scoring used for outbound, and surfacing a ranked list of likely attendees before invitations go out.
Every event leaves a pattern worth studying - who actually attended, who became pipeline, how long that took. That history becomes the input for the next event's targeting, not a report that gets filed away.
Reps get a briefed list of likely attendees with a real reason each one matters. During the event, a fast way to log who was actually talked to matters more than a form nobody fills out honestly. Afterward, the fastest follow-up wins - whatever came up in the room routes straight into the same scoring and sequencing system as every other lead.
This is the thinking. The next step is building it against real accounts, real signals, and real feedback from the people who'd actually use it.