Salesforce-native managed package

AI that runs your follow-up —
inside your own Salesforce org.

WinCadence scores every lead — and the account behind it — builds a cadence for that specific person, drafts each touch for your rep to approve, sends and tracks it, and reads the reply to decide what happens next. On your objects, your permissions, your reports — nothing to export, nothing to reconcile.

Book a 30-min walkthrough Watch the 72-second overview
  • Installs as a managed package
  • Your data stays in your org
  • Every touch written back to Salesforce
Managed package Installs into your org — Lead, Contact, Account stay yours.
Permission-aware Permission sets, sharing and record ownership are enforced server-side.
Human in the loop AI drafts. Your rep approves. Nothing sends blind.
Fully auditable Every score, send, open and reply is a record you can report on.

The gap

Your pipeline doesn't leak at the demo. It leaks at touch three.

  • Nobody knows which lead is worth the hour. Reps work the list top-down, and the best-fit prospect sits at row 240.
  • Follow-up quality collapses after the first email. Touch one is personalised. Touch four is a template with a first name in it.
  • Replies land in a mailbox, not in the CRM. An interested buyer answers on a Friday, and the manager finds out on Tuesday.

Bolting a separate sequencing tool onto Salesforce fixes none of that — it just moves the activity somewhere your reports can't see it.

How it works

Three moves, all of them inside Salesforce.

From a raw lead list to a logged, measurable cadence — without a rep opening a second tool.

01

Score

One click on a Lead or Contact runs enrichment: the profile is found and identity-verified against the person and their company, then scored 0–10 with the reasoning, the positive signals, the missing context and a recommended next move.

  • The same one click on an Account scores the company against your ICP
  • Every run is saved as a snapshot, so you can see what changed since last time
  • Wrong-namesake matches are rejected, not written to the record
  • Scheduled re-scoring keeps a whole segment current
02

Run the cadence

Enrolling a prospect generates a sequence built for that person — channels, number of touchpoints and pacing come from your campaign, the angle comes from their profile. Each step arrives with a subject line and a draft.

  • Reps edit, approve, complete or skip — the draft is a starting point, not a send
  • Steps are date-paced across the campaign window and never drift outside it
  • Overdue steps chase the record owner by email and in-app notification
03

Read the response

Sends are tracked to delivered, opened, clicked or bounced. Replies come back into Salesforce, attach to the step that caused them, and are read for intent.

  • Positive reply → task for the owner, remaining steps stood down
  • Not interested → enrollment closed, the rep's queue stays clean
  • Neutral → the next touch is rescheduled instead of firing on top of them

72 seconds

The short version, before you give us 30 minutes.

Want to see the product itself? →

What you actually get

Five systems that only work because they share one record.

01

Three ways a prospect gets in — search, a note from the stand, or a form

A cadence only helps once someone is a record in your org. WinCadence covers the step before that: on accounts you already have, and at the events where you meet people who aren't in the CRM yet.

  • Buying committee search on the Account — find the people who work there by title, seniority, location and keywords. Company, domain and size come from the Account, so the rep only has to describe the person.
  • Imported as Contacts under that Account — not Leads somebody has to convert back onto the same company later. Anyone you already have is flagged before import, so a repeat search doesn't leave you deduplicating.
  • A meeting note on a phone — typed, or spoken — the rep says what was said while their hands are still busy, and the transcript lands in the same box they would have typed into, editable before anything is saved. Anything they typed themselves is never overwritten. The recording isn't kept; the text on the Lead is.
  • A photo of the card is enough on its own — when there was no time to write, the card alone starts the Lead. A card is a weaker source than a conversation and WinCadence says so, lowering its own confidence and leaving the rest blank rather than filling it in. The rep confirms the fields and only then does the Lead exist — nothing is created from an unchecked read, because character errors in an email address are silent.
  • Your own Lead fields, answered by the conversation — your admin picks which fields the rep is asked for — source, campaign, segment, industry, your own custom ones — and the note is read for them too, so the field the conversation actually answered arrives already filled in and marked as read from the note. Picklists offer your org's live values and a value that isn't on your list is dropped, not invented. Answers carry to the next Lead, so the show is set once in the morning rather than twenty times a day.
  • What they came looking for, in their words — one line at the top of the Lead description, which is the line somebody sorting a stack of badges after the show is actually looking for.
  • A public form for stands and events — published on your Experience Cloud site for guest users, fields chosen by your admin, the show name filled in and hidden. Every page saves on its own, so somebody who walks away halfway still leaves you their details.
  • Straight into a campaign — anything captured can be enrolled as it is created, so scoring and the first touch happen while the conversation is warm.
02

Scoring & enrichment for the person and the company

A number on its own gets ignored. Every score ships with the evidence: what the model found, what it couldn't find, how confident it is, and the one action it would take next. It runs on two levels — the prospect on a Lead or Contact, and the Account they belong to.

  • 0–10 fit score on the prospect, with a written insight on the record
  • 0–10 ICP score on the Account — firmographics researched across the open web, not just one profile page: size, industry, HQ, founding year, specialities
  • Two confidences, not one — Company Match and Person Match are scored and explained separately, because they are different mistakes: the wrong company re-segments the record, the wrong person costs one call. Set the floor below which a match counts as no result at all
  • The employer on your record is taken as given — the research describes the company you have, not the one the person moved to, so a job change stops quietly rewriting the industry, the segment and who else to call there. The move itself shows up where it belongs: in Person Match
  • Industries from your picklist — give us your taxonomy and a value outside it can't come back, so the field stays reportable
  • Snapshot history on both — every run kept, with a diff against the previous one
  • The whole career, not just the current title — role history, education and languages on the person's timeline, with what changed since the last run called out (a new employer is usually the reason to call)
  • Buying signals on the Account — funding, tenders, new units, leadership moves, expansion, M&A and hiring, each dated and linked to where it was found
  • A task only when one is warranted — the owner is nudged when the account scores 7+ and the signal is a dated, pre-purchase event. Routine news stays on the card instead of in somebody's task list
  • Nothing written to the standard object — insights live on their own records, so your Lead, Contact and Account fields stay yours
  • Scheduled refresh for a segment of prospects you define, so scores don't go stale
  • Refresh the facts without re-running the verdict — re-reading where somebody works is a cheaper action than asking for a new score, and priced that way. The previous score stays put
  • A bad minute at the AI provider doesn't cost you a record — transient failures are retried, and if the score still can't be produced, the profile you already paid for is kept rather than thrown away with it
03

A cadence per prospect, not a template per campaign

You set the guardrails — channels, touchpoints, duration, tone, objective, language. The sequence is then generated for the person in front of you, and stays inside the window you defined.

  • Multi-channel steps — email, LinkedIn, call, task, whatever your campaign allows
  • Drafts with subject lines, editable in the timeline before anything goes out
  • Bulk enrollment from field criteria, for Leads and Contacts alike
  • Parallel campaigns handled cleanly — pause the others, or run them side by side
  • Start over, cancel or hand back a prospect without touching the other cadences
04

Sending, tracking and reply triage in one loop

Send from the timeline through Salesforce itself or your own email provider. What happens to that message then updates the same step — and the reply decides the next move.

  • Delivery states — delivered, opened, clicked, bounced, unsubscribed, complained
  • Replies routed back into Salesforce and matched to the step by thread
  • Intent read on arrival — interested, not interested, or "not now"
  • Automatic stand-down so nobody gets touch five after saying yes
  • Owner tasks that say why they exist: a reply, a score jump, or a rep's own hand-off
05

A manager view that matches what the rep sees

One home page answers the two questions that matter: is the cadence being run, and is it producing pipeline. Every number is clickable down to the rows behind it.

  • Campaign performance — steps done, open, skipped, overdue, per campaign
  • Pipeline impact — opportunities created, won and lost amounts from the campaign
  • Needs attention — the prospects who replied or spiked, before they cool
  • Scoring coverage — how much of the portfolio is scored and the top data gaps, counted across every Lead and Contact you can see rather than only the ones already on a campaign. A prospect nobody has enriched counts as unscored, which is the honest answer; if the org is large enough that the number is partial, the card says so instead of showing a total that isn't one
  • Data scoping by role — reps see their own prospects, managers see the campaign

Typical business impact

What teams tell us changes first.

The mechanism is simple: reps spend their hours on the leads worth the hour, and no conversation dies because nobody remembered to send touch four.

30%
Lead conversion
40%
Shorter sales cycle
25%
Team productivity

Put your own numbers in

Your assumptions, your arithmetic — nothing is sent anywhere.

— extra opportunities per quarter
— added pipeline per quarter
Pressure-test this with us

Modelled at a 30% relative lift in lead-to-opportunity conversion, 13 working weeks per quarter. It is an illustration, not a forecast.

Reported ranges depend on industry, lead quality, list hygiene and how consistently the team runs the cadence. We would rather show you the mechanism than promise you a number.

Industries

Built for long, crowded, follow-up-heavy pipelines.

The cadence engine is the same everywhere. What changes is the buying committee, the compliance bar, and the timing that makes follow-up land.

SaaS & Technology

Trial signups and product-qualified leads arrive faster than any rep can triage. Usage and firmographic signals are scored together, so only the accounts worth a human touch enter a cadence — and expansion stops competing with new business for the same rep hour.

Financial Services & Insurance

Outreach here has to survive an audit. Cadences run on approved messaging, every touch is written back with a timestamp, and drafts stay in review before anything reaches a client — advisers keep the relationship, compliance keeps the record.

Manufacturing & Industrial

Quotes go quiet for months while engineering, plant operations and procurement argue internally. A multi-contact cadence stays alive across the whole committee, paced for a long cycle, so the deal is still warm when the budget finally unlocks.

Professional Services & Consulting

Pipeline is relationships, and relationships go cold quietly. Scoring surfaces the dormant accounts most likely to buy again, and outreach is drafted in the partner's own tone — presence without the mass-email feel that damages a referral network.

Healthcare & Life Sciences

Clinical buyers, procurement and value-analysis committees all move on different clocks. Pacing respects contact rules and quiet periods, follow-up is documented, and reps see which stakeholder went quiet before the evaluation stalls. Account research can run on a healthcare profile, which tells a provider from a supplier and then asks the questions that fit — facility type, ownership, network, beds and departments for one; what they sell and how they reach the market for the other

Logistics & Transportation

Rate requests are won on speed. Inbound leads are scored on lane fit and volume the moment they land, the first touch goes out while the shipper is still comparing, and monitoring flags every quote that hasn't had a reply worth chasing.

Construction & Real Estate

Deals are driven by dates, not moods — bid deadlines, permits, lease expiries, project phases. Cadences are timed against those milestones, so follow-up arrives when the decision is open instead of three weeks after the award.

Staffing & Recruitment

Two pipelines, one team, constant risk of double-messaging. Client and candidate cadences run in parallel with shared visibility, so a hiring manager and the people you're placing never get conflicting outreach from the same desk on the same day.

Also in use across
  • Energy & Utilities
  • Telecommunications
  • Wholesale & Distribution
  • Education & EdTech
  • Media & Advertising
  • Business Process Outsourcing
Don't see yours? Ask us how it maps →

Security & control

The questions your admin will ask, answered up front.

Your records never leave your org

WinCadence is a managed package running in your Salesforce. Prospect records, cadence steps, emails and replies are your data, in your org, under your retention policy.

You own the prompts and the model

Prompts are editable in Setup and the model is configurable. You decide what context is sent for scoring or drafting, and you can change it without a redeploy.

Sharing is enforced, not decorated

Cadence steps follow the prospect's owner. A rep sees their own prospects in reports, search and the API — not just on the page. Scoping is applied server-side, so nothing a browser sends can widen it.

Authenticated integrations only

Backend calls are authenticated per client. Inbound delivery webhooks are verified by HMAC signature with a freshness window — an unsigned or stale call is rejected.

Roles that match your team

Admin, campaign manager and sales user permission sets ship with the package. Campaign setup, bulk enrollment and org-wide numbers stay with the people who should have them. Field-level security is respected too: a field a rep isn't allowed to see is simply absent from the form, instead of breaking the save.

Consumption you can see

AI usage is metered and visible in Setup: balance, estimated consumption and a history of external calls, priced per action — and the page lists only the actions your plan actually includes, so you are not reading a price list for things you can't run. No surprise bill at the end of the quarter. What your plan doesn't include is refused by our servers before a credit is spent, so a part of the package you don't use can't quietly cost you anything.

Before you book

Straight answers.

How long does it take to get running?

The package installs in minutes. A working setup — settings, permission sets, the components on your Lead and Contact pages, email sending and reply capture — is a guided checklist in Setup, and most teams complete it in an afternoon with an admin in the room. We usually do the first campaign with you.

Does this replace Salesforce Sales Engagement / High Velocity Sales?

It covers the same ground for most teams — sequences, tracked sends, task queues — and adds the part those tools leave to you: deciding who deserves the touch and what to actually say to them. It runs on your standard objects, so you are not choosing a parallel activity model.

Will the AI email my prospects without a human seeing it?

No. Steps arrive as drafts on the timeline. A rep reads, edits and sends — or skips. The AI is generating and triaging, not autonomously mailing your market.

What happens to a reply?

It comes back into Salesforce, attaches to the step that triggered it and is read for intent. Interested creates a task for the record owner and stands the cadence down. Not interested closes the enrollment. "Not now" pushes the next touch out instead of sending on top of them.

Can we use our own email infrastructure?

Yes. Send natively through Salesforce, or through your own provider when you need higher volume and richer delivery events. Either way, delivery states and replies land on the same cadence step.

Does it work on Contacts, or only Leads?

Both, including in the same campaign. Enrollment, scoring, the timeline and bulk adding by field criteria all work against Leads and Contacts. Which one you get depends on where the person came from: buying committee search on an Account creates Contacts on that Account, because the company is already a record and converting them back onto it later is wasted work. Meeting notes and the event form create Leads, since at that point you usually don't know the company yet.

What does a rep actually do at a trade show?

Talk, then hold up the phone. They can speak the note instead of typing it, or just photograph the card when there was no time for either. WinCadence reads it, drafts the Lead, and asks for the handful of fields your admin decided matter — with your org's own picklist values, and with anything the conversation already answered filled in and labelled as read from the note.

The rep confirms before the Lead exists, because a misread character in an email address is silent. The answers that describe the stand rather than the person — show, campaign, source — carry to the next Lead, so they are set once in the morning.

How do we know the AI matched the right person?

Because it tells you, in two numbers rather than one. Company Match says how sure it is about the employer and what that employer does; Person Match says how sure it is about the individual there. Both come with a one-line reason, and both are kept on the record, so "we know where to call, not yet who" reads as exactly that instead of averaging into a mediocre-looking lead.

You can set a floor below which a match is treated as no result at all, and the employer already on your record is taken as given — so somebody changing jobs no longer rewrites the industry and segment of the record you had.

Can it score companies, not just individual people?

Yes — Accounts get their own enrichment and their own 0–10 ICP score, driven by the company rather than a person. The research runs across the open web, not a single profile page, so you get size, industry, HQ, founding year, specialities and website alongside the score, the positive signals, the missing context and a recommended next move.

Every run is kept as its own snapshot with a diff against the previous one, so an account that grew, moved or changed focus is visible rather than silently overwritten — and nothing is written onto the standard Account fields.

How is it priced?

Per org, with AI usage metered as credits so you can see exactly what scoring and drafting consume. The package comes in four parts — prospect insights, company insights, lead capture and the cadence engine — so you can start with the ones you need and add the rest later without a reinstall. We will size it against your seat count and lead volume on the call rather than guess at it here.

Can we trial it in a sandbox first?

Yes — that is the normal path. Install in a sandbox or developer org, run one campaign end to end with real-shaped data, then decide about production.

Next step

Book a 30-minute walkthrough.

No deck. We open a Salesforce org, score a lead that looks like yours, generate the cadence and show you what a reply does to it.

  • Minutes 1–10 — your pipeline, your list, where follow-up actually breaks
  • Minutes 10–25 — the product on a live record, at your pace
  • Minutes 25–30 — install path, effort and cost, honestly

You will speak to someone who has implemented this, not an SDR reading a script.

We reply within one business day. No sequence, no drip — by submitting you agree to be contacted about WinCadence.