Scout
An AI sales agent that researches, scores and follows up — so reps only talk to people ready to buy.
- AI sales agent
- 21-Day Launch
- Web app, CRM integration, Email + calendar
Approvals
4 drafts waiting · 2 strong fits
Priya Nair
Halden Dental Group
Why now
- Multi-site practice with 14 clinics1
- Opening 3 clinics this autumn2
Three new clinics, one front desk?
Hi Priya — congratulations on the three clinics opening this autumn. That's a lot of new patient enquiries to route. Since your team already books through Calendly…
3 sources·Follow-up Thu 10:00
What Scout does
Scout watches inbound leads, researches each company, scores fit against your ideal customer, drafts a personal first message, updates the CRM and schedules follow-ups. Reps approve, edit and send.
Key decisions
The rep sends every email.
Scout researches, scores and drafts, and a person approves each message. Sending on someone’s behalf is a trust line, so the first version proves the drafts are good before anything goes out alone.
Every claim shows its source.
Reps act on research they can check. Reasoning they can’t see gets ignored, however good it is.
Inbound leads first, no bulk sequences.
Outbound list building is a different product with different risks. Leaving it out kept the build on the part that proves the idea.
Small sales teams spend their best hours on the worst leads.
Inbound leads arrive with a name and an email. Before a rep can decide whether to call, they open LinkedIn, the company website, the CRM and a news search. Most leads turn out to be a poor fit — and the good ones wait.
- Research takes 10–20 minutes per lead, done by hand
- Good-fit leads wait hours or days for a first reply
- Follow-ups slip because nobody owns the reminder
- CRM data is incomplete, so reporting can't be trusted
The founder-led sales lead
Head of sales at a 10–50 person B2B company
Owns the pipeline, still sells personally, and manages two or three reps without a sales-ops team.
What they need to get done
- Know which new leads deserve attention today
- Reply to good-fit leads while they are still interested
- Keep the CRM accurate without nagging the team
- Never let a promising conversation go quiet
The assumption to prove
Reps will trust AI research and drafts if they can see the sources and approve before anything is sent.
Keep the human as the sender. Scout does the research, scoring and drafting; the rep reviews a queue and sends with one click. Trust is earned through visible evidence, not hidden automation.
In the first version
- Inbound lead capture from forms and CRM
- Company research with cited sources
- Fit scoring against an editable ideal-customer profile
- Personal first-message drafts
- CRM record updates and follow-up scheduling
Deliberately left out
- Fully autonomous sending
- Outbound list building and bulk sequences
- Call recording and coaching
- Multi-currency forecasting
The core journey
Lead arrives
Form fill or new CRM contact triggers Scout
Research
Company site, public profiles and news, with sources
Score
Fit against ideal-customer rules, with reasons
Draft
Personal first message referencing real context
Approve
Rep edits and sends from the review queue
Follow up
CRM updated, reminder or meeting scheduled
Design exploration
We explored three ways to present the agent's work before settling on a review queue. The question underneath every option was the same: how much should a rep have to read before they trust a lead?
Sales
Setup
Agent active
Researching 3 new leads
Review queue
Tuesday · 12 leads researched since 08:00
Researched today
186 since 10:00
Strong fits
53 to answer today
Drafts ready
9Avg. 112 words
Follow-ups due
52 overdue
Priya Nair
Halden Dental Group
Why it's a fit
- Multi-site practice with 14 clinics1
- Opening 3 clinics this autumn2
- Already runs HubSpot and Calendly3
Budget not confirmed — ask on the first call
Draft
118 wordsHi Priya — congratulations on the three clinics opening this autumn. That's a lot of new patient enquiries to route…
Sales
Setup
Agent active
Researching 3 new leads
Review queue
Tuesday · 12 leads researched since 08:00
Researched today
186 since 10:00
Strong fits
53 to answer today
Drafts ready
9Avg. 112 words
Follow-ups due
52 overdue
Priya Nair
Halden Dental Group
Why it's a fit
- Multi-site practice with 14 clinics1
- Opening 3 clinics this autumn2
- Already runs HubSpot and Calendly3
Budget not confirmed — ask on the first call
Draft
118 wordsHi Priya — congratulations on the three clinics opening this autumn. That's a lot of new patient enquiries to route…
Chat assistant or work queue?
Work queue
Reps don't want to ask questions — they want a prioritised list of what to do next. Chat stays available inside each lead.
Show the score as a number or a verdict?
Verdict with reasons
A number invites arguments; 'Strong fit — 3 reasons' invites action and makes the logic reviewable.
Auto-send or approve?
Approve, one click
Sending on someone's behalf is a trust line. The MVP proves value first; autonomy can be earned later.
Scout’s design system
Calm, dense and evidence-forward. Teal signals the agent's work; amber is reserved for things that need a human.
Scout Teal
Agent activity, primary actions
Mint
Evidence highlights
Amber
Needs your review
Deep Green
Text and navigation
Mist
Workspace background
Typography
System UI stack at compact sizes; tabular numerals for scores and dates; mono for source citations.
Components
- Lead queue row
- Fit verdict badge
- Source citation chip
- Draft editor
- Agent timeline
- CRM sync status
Principles
- Show the evidence
- Humans send
- One next action per lead
Key screens
Sales
Setup
Agent active
Researching 3 new leads
Review queue
Tuesday · 12 leads researched since 08:00
Researched today
186 since 10:00
Strong fits
53 to answer today
Drafts ready
9Avg. 112 words
Follow-ups due
52 overdue
Priya Nair
Halden Dental Group
Why it's a fit
- Multi-site practice with 14 clinics1
- Opening 3 clinics this autumn2
- Already runs HubSpot and Calendly3
Budget not confirmed — ask on the first call
Draft
118 wordsHi Priya — congratulations on the three clinics opening this autumn. That's a lot of new patient enquiries to route…
Sales
Setup
Agent active
Researching 3 new leads
Review queue /Halden Dental Group
Priya Nair
Operations Director · Halden Dental Group · Demo request, 9 min ago
Company research
- Industry
- Dental care 1
- Size
- About 180 people 1
- Locations
- 14 clinics, 3 opening 2
- Tools
- HubSpot, Calendly 3
- Based in
- Leeds, UK 1
Sources
- 1haldendental.example/about
- 2haldendental.example/news
- 3careers.haldendental.example
Scout only states facts it can link to a source. Anything else is marked for review.
Fit rules
4 of 5 matched- 5+ locations14 clinics
- Uses a CRMHubSpot
- Growing this year3 openings, 6 roles
- UK or IrelandLeeds
- Budget signalNot found
Agent activity
- 08:12Lead received from demo request form
- 08:12Read 4 pages, cited 3 sources
- 08:14Scored against fit rules v3 — strong fit
- 08:15Drafted first message (118 words)
- 08:15HubSpot contact and company updated
First message
WarmShorterFormalTopriya.nair@haldendental.example
SubjectThree new clinics, one front desk?
Hi Priya,
Congratulations on the three clinics opening this autumn 2 — that's a lot of new patient enquiries to route.
Since your team already books through Calendly, we could plug straight into your existing setup.
Would a 20-minute look next Thursday be useful?
Follow up Thu 10:00 if no reply
Approvals
4 drafts waiting · 2 strong fits
Priya Nair
Halden Dental Group
Why now
- Multi-site practice with 14 clinics1
- Opening 3 clinics this autumn2
Three new clinics, one front desk?
Hi Priya — congratulations on the three clinics opening this autumn. That's a lot of new patient enquiries to route. Since your team already books through Calendly…
3 sources·Follow-up Thu 10:00
How it works underneath
An event-driven agent: each new lead starts a job that calls research and scoring tools, stores evidence, and waits in a queue for human approval before any external action.
- External
- Interface
- Service
- AI
- Data
Read the connections as a list
- Web forms / CRMtoAPI + job queue
- API + job queuetoResearch & scoring agent
- Research & scoring agenttoWeb search, enrichment
- Research & scoring agenttoPostgres + evidence store
- Scout web apptoAPI + job queue
- API + job queuetoHubSpot / Pipedrive
- API + job queuetoEmail + calendar
- Web forms / CRM
- Scout web app
- API + job queue
- Research & scoring agent
- HubSpot / Pipedrive
- Email + calendar
- Web search, enrichment
- Postgres + evidence store
- Web forms / CRMtoAPI + job queue
- API + job queuetoResearch & scoring agent
- Research & scoring agenttoWeb search, enrichment
- Research & scoring agenttoPostgres + evidence store
- Scout web apptoAPI + job queue
- API + job queuetoHubSpot / Pipedrive
- API + job queuetoEmail + calendar
Likely stack
- Next.js + TypeScript
- Postgres
- Background job queue
- LLM with tool calling
- HubSpot / Pipedrive API
- Gmail / Outlook API
A 21-day build plan
Scope and ICP rules
Define what 'good fit' means in plain rules, pick the CRM, map the fields Scout may write.
Design the queue
Wireframes for queue, dossier and approval; clickable prototype tested with a rep-style walkthrough.
Agent + integrations
Research tools, scoring prompts with evaluation examples, CRM sync and email drafting.
Review loop
Approval queue, follow-up scheduling, audit log and failure handling.
Test and launch
Run against historical leads, tune scoring, deploy and hand over.
Across devices
Research and editing happen at a desk; approvals happen everywhere. The mobile layout collapses the dossier into a swipeable card focused on one decision.
- Queue becomes a single column with verdict first
- Draft editing uses a full-screen sheet
- Sources collapse into an expandable list
Approvals
4 drafts waiting · 2 strong fits
Priya Nair
Halden Dental Group
Why now
- Multi-site practice with 14 clinics1
- Opening 3 clinics this autumn2
Three new clinics, one front desk?
Hi Priya — congratulations on the three clinics opening this autumn. That's a lot of new patient enquiries to route. Since your team already books through Calendly…
3 sources·Follow-up Thu 10:00
An inbox of anonymous leads, turned into a ranked, researched, ready-to-send queue — with the rep still in charge of every message.
Sales
Setup
Agent active
Researching 3 new leads
Review queue /Halden Dental Group
Priya Nair
Operations Director · Halden Dental Group · Demo request, 9 min ago
Company research
- Industry
- Dental care 1
- Size
- About 180 people 1
- Locations
- 14 clinics, 3 opening 2
- Tools
- HubSpot, Calendly 3
- Based in
- Leeds, UK 1
Sources
- 1haldendental.example/about
- 2haldendental.example/news
- 3careers.haldendental.example
Scout only states facts it can link to a source. Anything else is marked for review.
Fit rules
4 of 5 matched- 5+ locations14 clinics
- Uses a CRMHubSpot
- Growing this year3 openings, 6 roles
- UK or IrelandLeeds
- Budget signalNot found
Agent activity
- 08:12Lead received from demo request form
- 08:12Read 4 pages, cited 3 sources
- 08:14Scored against fit rules v3 — strong fit
- 08:15Drafted first message (118 words)
- 08:15HubSpot contact and company updated
First message
WarmShorterFormalTopriya.nair@haldendental.example
SubjectThree new clinics, one front desk?
Hi Priya,
Congratulations on the three clinics opening this autumn 2 — that's a lot of new patient enquiries to route.
Since your team already books through Calendly, we could plug straight into your existing setup.
Would a 20-minute look next Thursday be useful?
Follow up Thu 10:00 if no reply
Approvals
4 drafts waiting · 2 strong fits
Priya Nair
Halden Dental Group
Why now
- Multi-site practice with 14 clinics1
- Opening 3 clinics this autumn2
Three new clinics, one front desk?
Hi Priya — congratulations on the three clinics opening this autumn. That's a lot of new patient enquiries to route. Since your team already books through Calendly…
3 sources·Follow-up Thu 10:00
Leads waiting while your team does the research?
Scout is one shape an AI agent can take. Tell us how your sales process runs today, and we’ll scope the first version around it.
21-Day Launch
$3,700
For fuller products: several roles, AI agents that take actions, heavier integrations or a mobile MVP.
- Up to 5 core workflows
- Up to 20 screens
- Multiple roles & permissions
Your idea gets its own written scope and fixed price before anything starts. Compare packages
In the first version
- Lead intake from one CRM and web forms
- Research agent with cited sources
- Editable fit rules and verdicts
- Draft review and one-click send
- CRM updates and follow-up reminders
- Admin settings and activity log
Later, once it's proven
- Outbound prospecting
- Multi-step sequences
- Call intelligence
- Team performance analytics