The One-Percent Problem
"Our top sales rep closed 340% of quota last year. When we asked her what she did differently, she said: 'I research every prospect for 45 minutes before I call them. I know their priorities, their pain, their recent news. I never pitch — I just solve the problem they already have.' We have 47 other reps who spend 4 minutes on research. The gap isn't talent. It's time." — VP of Sales, B2B SaaS company, 2025
Every sales team has a top 1%. They close more deals, retain more customers, and generate more pipeline from fewer touches than anyone else on the floor. When managers study what makes these performers different, the answer is rarely charisma or experience. It is preparation. The top rep knows her prospect's business before she picks up the phone. She personalises every outreach message to the specific situation her prospect is navigating right now. She follows up at exactly the right moment with exactly the right content. She treats each deal as a unique puzzle rather than a copy-paste of yesterday's pitch.
The other 99% of the team uses the same CRM, the same playbook, and the same email templates. They have access to exactly the same information. What they do not have is the cognitive capacity to process it all. Researching a prospect properly takes 30 to 45 minutes: reading their LinkedIn profile, studying their company's recent news, understanding their industry pressures, identifying the specific pain that overlaps with your product's capabilities, and crafting an opening message that proves you did the homework. Multiply that by 50 prospects a week and the arithmetic is impossible. No human can sustain that depth at that volume.
This is the One-Percent Problem, and it is the reason this chapter exists. The Claude Sales and Marketing Plugins, released in early 2026 alongside the broader Cowork plugin ecosystem, are built around a single organising principle: scale the expertise of your top 1% across the entire team. Every rep gets the research depth of your best rep. Every email is personalised to the depth your best rep would personalise it. Every follow-up is timed and crafted with the judgment of your best closer. Every marketing campaign is planned and analysed with the rigour of your best strategist.
Why This Matters Now
The timing is not accidental. Three converging pressures make AI-native sales and marketing capabilities urgent for any organisation competing in B2B markets.
The first pressure is buyer behaviour. The modern B2B buyer completes 70% of their evaluation before they ever speak to a sales rep. They research vendors online, read peer reviews, consume thought leadership content, and form strong opinions before your rep gets a chance to make their case. By the time a buyer agrees to a discovery call, they have already decided whether you are credible. The quality of your outreach — the research depth, the personalisation, the relevance of your message to their specific situation — is what determines whether you earn that call at all.
The second pressure is data volume. The signals that indicate buying readiness — leadership changes, funding rounds, hiring patterns, contract wins, regulatory shifts, competitive moves — are all public information. They are published on LinkedIn, in trade press, in government filings, in job boards. The constraint is not access to this data. It is the human capacity to monitor it, correlate it, and act on it before the window closes. A VP of Operations at a logistics company in Karachi who just won a major new contract and posted on LinkedIn about scaling challenges is a hot prospect right now. In three months, she will have either solved the problem or committed to a competitor. The intelligence is perishable.
The third pressure is competitive parity. If your competitors adopt AI-native sales tools and you do not, the gap compounds quickly. Their reps will research every prospect at depth. Their outreach will be personalised. Their follow-ups will be timed to external signals. Your reps will still be sending templated emails and hoping for replies. The advantage accrues to the first movers, and it compounds every quarter.
B2B (Business-to-Business): A commercial model where one business sells products or services to another business, rather than directly to individual consumers. B2B sales cycles are typically longer, involve multiple decision-makers, and require more research and personalisation than consumer sales. Examples include a software company selling workflow automation to a logistics firm, or a consulting firm selling advisory services to a bank. Most of the techniques in this chapter apply specifically to B2B contexts, where the buyer expects to be understood before they are sold to.
The Five Questions Your Best Rep Answers
Before every outreach, your top 1% performer answers five questions that most reps never think to ask. These five questions are the foundation of everything the Sales Plugin automates.
Question 1: Who exactly is this person? Not just their job title, but their career trajectory. Where did they work before? How long have they been in their current role? Are they a new hire still proving themselves, or a veteran with established authority? What have they published, posted, or spoken about recently? This context shapes how you approach them — a new VP is motivated differently from a tenured one.
Question 2: What does their company actually do? Not the website boilerplate, but the real business model. What are the actual revenue drivers? What is their market position — growing, stable, or under pressure? How big are they? What does their tech stack look like? A company running legacy systems and hiring rapidly is in a fundamentally different buying posture from a company with a modern stack and stable headcount.
Question 3: What is happening to them right now? This is the question that separates good reps from great ones. Recent news, funding rounds, product launches, leadership changes, market pressures, regulatory changes affecting their sector — these are the timing signals that determine whether a prospect is ready to buy today or in six months. A company that just won a major new contract is under immediate scaling pressure. A company that just completed a round of layoffs is not.
Question 4: What is the specific pain this person is most likely experiencing? Given their role, their company's situation, and your product's capabilities — where is the overlap? This is not a generic "they probably need better software." It is a specific hypothesis: "She is a new VP at a company that just won a big contract, hiring rapidly, and she posted about scaling quality without adding headcount. Our product's core value proposition maps directly to that stated priority."
Question 5: What is the hook? The specific, verifiable thing you can reference in your opening message that proves you did the research. Not "I noticed your company is growing" — that could apply to anyone. Rather: "Your LinkedIn post two weeks ago about maintaining quality as you scale stopped me scrolling — that is exactly the tension we help logistics operators navigate."
The /research command answers all five questions in under four minutes, for every prospect, at any volume. That is the transformation this chapter teaches you to build.
CRM (Customer Relationship Management): The system of record that tracks every interaction between your organisation and its prospects and customers. In sales, the CRM stores contact details, deal stages, communication history, meeting notes, and pipeline value. Common CRM platforms include Salesforce, HubSpot, Pipedrive, and Zoho. The CRM is the data foundation on which everything in this chapter is built — the Sales and Marketing Plugins read from it, write to it, and enrich it. A CRM record that is incomplete or outdated is worse than no record at all, because it gives the rep false confidence.
Pipeline: In sales, the pipeline is the staged progression of deals from initial prospect identification through to closed revenue. A typical pipeline has stages such as Prospect, Qualified, Discovery, Proposal, Negotiation, and Closed-Won (or Closed-Lost). Each stage represents increasing commitment from the buyer and increasing investment from the seller. Pipeline value — the total monetary value of all deals at each stage, weighted by the probability of closing — is the primary metric by which sales teams forecast revenue. When this chapter refers to "pipeline analysis" or "pipeline health," it means assessing whether deals are progressing normally, stalling, or at risk.
The Revenue Operations Framework
Before touching any plugin command, understand the conceptual architecture that organises this entire chapter.
RevOps (Revenue Operations): The discipline of aligning Sales, Marketing, and Customer Success around a shared data foundation, shared process definitions, and shared accountability for revenue outcomes. In the pre-AI era, RevOps was primarily a backward-looking function: building the CRM taxonomy, defining lead scoring models, maintaining the tech stack, producing pipeline reports. In the AI-native era, RevOps becomes a forward-looking orchestration function — the team that builds the AI agents, SKILL.md libraries, and workflow integrations that allow every customer-facing employee to operate at the level of their best colleague.
RevOps exists because the handoff between departments is where revenue is lost. Marketing generates leads that sales does not follow up. Sales closes deals that customer success cannot retain. Customer success identifies upsell opportunities that sales never hears about. Each function optimises independently, and the gaps between them leak money. RevOps closes those gaps by creating shared definitions (what is a qualified lead?), shared data (the CRM that everyone trusts), and shared processes (what happens when a lead crosses the scoring threshold?).
In the AI-native era, RevOps gains a new and more powerful mandate: building the AI agents that automate the handoffs. When a marketing campaign generates a whitepaper download, the CRM Enrichment agent updates the record, the Lead Scoring agent recalculates the score, and if the score crosses the threshold, the Outreach Sequencing agent generates a personalised message for the assigned rep — all within hours, with no manual handoff. This is the revenue engine that the top 1% of organisations are now building. This chapter teaches you to build it.
The four domains this chapter addresses span the full revenue lifecycle:
| Domain | Pre-AI State | AI-Native State |
|---|---|---|
| Prospecting | Manual research; generic outreach; volume beats quality | Automated deep research; hyper-personalised outreach; quality beats volume |
| Lead Scoring | Rule-based models; stale CRM data; gut instinct | Continuous multi-signal scoring; real-time enrichment; predictive qualification |
| CRM Enrichment | Manual data entry; incomplete records; decaying data | Automatic enrichment from web, calls, emails; always-current prospect profiles |
| Marketing | Campaign-level targeting; slow content creation; lagging analytics | Segment-of-one targeting; instant content; real-time performance optimisation |
Consider how this framework applies beyond the UK logistics example that runs through this chapter. A SaaS company in Lahore selling enterprise resource planning (ERP) software to manufacturers across Punjab faces exactly the same four-domain challenge: their best sales rep researches each manufacturer deeply, their average rep sends generic emails, their CRM data decays, and their marketing campaigns target broad segments rather than individual buyers. The domain-specific details differ — the data sources are the Securities and Exchange Commission of Pakistan (SECP) instead of Companies House, the trade press is different, the LinkedIn culture may vary — but the RevOps architecture is identical.
ICP (Ideal Customer Profile): A detailed description of the type of company and buyer persona most likely to become a successful, long-term customer. An ICP includes firmographic criteria (company size, revenue, industry, geography), technographic signals (what technology they use, what they are hiring for), timing signals (events that indicate buying readiness), and persona profiles (the specific roles you sell to, their motivations, fears, and triggers). The ICP is the targeting foundation for everything the Sales Plugin does — research is matched against it, leads are scored against it, and outreach is calibrated to it. A vague ICP produces vague outputs. A precise ICP produces precise, actionable intelligence.
The Plugin Architecture
The Claude Sales and Marketing Plugins ship as a coordinated pair within Anthropic's official knowledge-work-plugins catalogue, released in early 2026 alongside the broader Cowork ecosystem. An Agent Factory extension — sales-revops-marketing@agentfactory-business — adds the three-dimension scoring model, ICP configuration system, jurisdiction overlays, Five Laws framework, and the five RevOps agents that transform the base plugins from competent generic tools into a calibrated revenue engine. Lesson 2 walks through the full installation, verification, and initial configuration of both layers.
This dual-layer design is deliberate: the handoff between marketing-generated leads and sales-worked pipeline is the single most common point of revenue leakage in B2B organisations. When the plugins share a common intelligence layer, the prospect research that marketing's enrichment process builds flows directly into the sales rep's research brief. There is no re-entry, no data loss, and no context switching between systems.
Both plugins connect to a shared configuration file — sales-marketing.local.md — that contains your organisation's ICP definition, messaging framework, brand voice, and competitive intelligence. This is the equivalent of the legal playbook from Chapter 22: the institutional knowledge that makes every output specific to your organisation rather than generic. Without this configuration, the plugins produce competent but generic output. With it, they produce output that sounds like it came from someone who has worked at your company for years.
SKILL.md: The file format introduced in Chapter 5 that packages domain expertise into a structured, reusable artifact an AI agent can load and execute. A SKILL.md file contains a YAML frontmatter header (name, version, description, trigger terms) and a body that encodes the logic, rules, and constraints the agent must follow. In this chapter, the Sales and Marketing Plugins are built on a library of SKILL.md files — one for each command (prospect research, lead scoring, outreach, etc.) plus a global router that directs queries to the right skill. The
sales-marketing.local.mdconfiguration file is the organisation-specific companion that makes the generic skills specific to your business.
Sales Plugin — Core Commands
The Sales Plugin provides eight commands that cover the full sales execution workflow, from initial prospect research through to post-meeting follow-up and pipeline analysis:
| Command | Function |
|---|---|
/research | Deep prospect and account research brief — answers the five questions your best rep answers manually |
/score | Lead scoring against your ICP definition across three dimensions: fit, timing, and engagement |
/enrich | CRM record enrichment from web, LinkedIn, public filings, and connected data sources |
/outreach | Personalised email and message drafting calibrated to research findings and the Five Laws of Outreach |
/sequence | Multi-touch outreach sequence generation with timing, channel mix, and branching logic |
/brief | Pre-call and pre-meeting preparation briefs with discovery questions and objection handling |
/follow-up | Post-interaction follow-up drafting — meeting summaries, next-step proposals, internal deal notes |
/pipeline | Pipeline analysis and deal health assessment — identifies stuck deals, at-risk opportunities, and forecast gaps |
Marketing Plugin — Core Commands
The Marketing Plugin provides six commands that span content creation, campaign planning, and performance analysis:
| Command | Function |
|---|---|
/campaign | Campaign planning and brief generation — audience, channels, budget, timeline, KPIs |
/content | Content creation for any channel and format — articles, social posts, emails, whitepapers, video scripts |
/copy | Ad copy, landing page copy, and subject line generation optimised for conversion |
/persona | ICP and buyer persona development through structured interview |
/analyze | Campaign performance analysis with multi-channel attribution and actionable optimisation recommendations |
/calendar | Content and campaign calendar planning across channels and personas |
B2B Pipeline Stages: The progression from stranger to customer typically follows these stages: Prospect (identified but not contacted), Lead (contacted; some engagement), Marketing Qualified Lead (MQL) (meets marketing's engagement threshold), Sales Qualified Lead (SQL) (meets sales' fit and timing criteria), Opportunity (active deal with defined scope and timeline), Proposal (formal offer extended), Negotiation (terms being finalised), Closed-Won (deal signed) or Closed-Lost (deal lost). Different organisations use different stage names, but the underlying progression is universal. The
/scorecommand maps prospects to these stages based on multi-signal intelligence, and the/pipelinecommand analyses the health of deals at each stage.
What This Chapter Builds
By the end of this chapter, you will have built a complete, operational AI-native revenue engine. The system has three layers:
The first layer is the intelligence layer — the SKILL.md library and sales-marketing.local.md configuration that encodes your organisation's specific expertise, ICP definition, brand voice, and competitive intelligence. This is the institutional knowledge that makes every AI output specific to your business.
The second layer is the execution layer — the 14 plugin commands that sales reps and marketers use directly. Research a prospect. Score a lead. Write an outreach message. Plan a campaign. Analyse performance. These are the tools that turn intelligence into action.
The third layer is the orchestration layer — the five RevOps agents that automate the handoffs between intelligence and execution. The Lead Intelligence agent monitors signals and alerts reps. The CRM Hygiene agent keeps data current. The Outreach Sequencing agent manages multi-touch campaigns. The Marketing Performance agent analyses results. The Revenue Reporting agent produces the leadership dashboard. Together, these agents create a self-sustaining revenue engine that operates continuously, not just when a human remembers to check.
This architecture applies whether you are a 50-person SaaS company in Dubai selling supply chain software across the Gulf Cooperation Council (GCC), a 200-person manufacturer in Faisalabad selling textile machinery across South Asia, or a 500-person logistics firm in London selling warehousing services across the UK and EU. The domain expertise is different. The plugin architecture is the same. The SKILL.md files are different. The RevOps framework is the same.
The next eleven lessons teach you to build each layer, configure it for your specific business, and deploy it as a system that runs continuously — not as a collection of tools you use occasionally.
Try With AI
Prompt 1: Map Your Team's One-Percent Gap
I want to understand the "One-Percent Problem" in my sales team.
Here is how our team currently works:
- Team size: [number of reps]
- Average time spent researching a prospect before outreach: [minutes]
- Average number of prospects contacted per rep per week: [number]
- Our best rep's close rate: [percentage]
- Our average rep's close rate: [percentage]
Based on the One-Percent Problem framework from Chapter 23 — where the gap
is driven by research depth and personalisation capacity, not talent —
analyse my team's situation. Calculate:
1. How many research-hours per week would be needed if every rep matched
the best rep's depth
2. The cognitive capacity gap (hours needed vs. hours available)
3. Which of the five research questions (who, company, timing, pain, hook)
my team is most likely skipping at scale
Then suggest which of the four RevOps domains (prospecting, lead scoring,
CRM enrichment, marketing) would create the most immediate leverage for
my specific team.
What you're learning: This prompt teaches you to apply the One-Percent Problem framework to a real team. You are practising the diagnostic step that precedes any AI deployment — identifying where the cognitive capacity gap is largest and where AI leverage would be highest. The agent's analysis will reveal whether your team's bottleneck is research depth, data quality, personalisation, or something else entirely.
Prompt 2: Define Your RevOps Domains
I am building a RevOps framework for a [industry] company based in
[city/country] with [number] employees. We sell [product/service] to
[target buyer type] at [target company type].
For each of the four RevOps domains — Prospecting, Lead Scoring,
CRM Enrichment, and Marketing — describe:
1. Our likely current state (pre-AI) based on a company of our size
and industry
2. The AI-native target state we should be building toward
3. The single biggest gap between current and target state
4. The first plugin command I should deploy to close that gap
Present this as a table with columns: Domain | Current State | Target State |
Biggest Gap | First Plugin Command.
Also identify: what data sources would I need for CRM enrichment that are
specific to [country]? For example, if I am in Pakistan I would need SECP
filings rather than Companies House. What are the equivalents for my
market?
What you're learning: This prompt teaches you to contextualise the generic RevOps framework for your specific organisation and market. The critical skill here is recognising that while the four-domain architecture is universal, the data sources, regulatory context, and market dynamics are local. A RevOps implementation in Karachi uses SECP data; one in Dubai uses DED commercial licences; one in London uses Companies House. The framework is the same. The inputs are different.
Prompt 3: Audit Your Current Sales Stack
Here is my current sales technology stack:
- CRM: [name and version]
- Email platform: [name]
- LinkedIn: [Sales Navigator? Yes/No]
- Lead scoring: [manual/rule-based/none]
- Content tools: [list]
- Analytics: [list]
- Other tools: [list]
Based on the Claude Sales and Marketing Plugin architecture described
in Chapter 23, analyse my stack for:
1. Which of the 14 plugin commands I could deploy immediately with
my existing tools (via MCP connectors)
2. Which commands require additional data sources or integrations
3. Where I have redundant tools that the plugins would replace
4. Where I have gaps that the plugins would fill
5. The recommended deployment sequence — which 3 commands should I
deploy first and why
Present this as a readiness assessment with a clear priority order.
What you're learning: This prompt teaches you to perform a technology readiness assessment — a critical skill before any AI deployment. You are learning to think about plugins not as standalone tools but as components that integrate with an existing stack via MCP connectors. The agent's analysis will reveal which parts of your current stack become more valuable with AI (your CRM data becomes the foundation for scoring and enrichment) and which parts become redundant (manual lead scoring spreadsheets, generic email templates).