Claude Skills for Sales Teams: 8 Workflows That Move Revenue (2026)

Claude skills for sales are reusable, packaged workflows that Claude loads on demand when a rep starts a task like drafting a prospect brief or a discovery recap. Anthropic shipped Skills in October 2025 as folders containing a SKILL.md file, optional scripts, and reference documents that Claude reads only when the work matches. For a sales org, that means one firm-standard version of each play — always on tone, always on template, always attached to the right data.

Sales teams already lean on generative AI. HubSpot's 2025 State of AI in Sales found that 81% of sales professionals now use AI in some form, and reps who use it save an average of two hours a day on non-selling tasks. Gartner's 2025 GenAI in sales research puts adoption of AI-assisted seller workflows at more than 60% of B2B sales teams by 2026. The question is no longer whether reps will use AI — it is whether the outputs match how your team actually sells.

This guide walks through eight concrete Claude skills that map to real revenue moments: prospect research, first-touch personalization, discovery recap to CRM, MEDDIC and BANT qualifiers, follow-up sequences, deal review talking points, competitive battle cards, and renewal risk narratives. You will see what belongs in each SKILL.md, when Claude fires it, and how Layer3 Labs deploys them for AEs, SDRs, and RevOps leaders.

What are Claude skills for sales teams?

Claude skills for sales teams are packaged capabilities — instructions plus scripts plus reference files — that Claude activates when a rep's task matches the skill's trigger description. Each skill is a folder containing a SKILL.md file with YAML frontmatter that names when to activate it, plus any Python, Markdown, or template files the workflow needs.

Claude reads only the frontmatter until a task matches. That means a rep can have 30 sales skills installed without slowing anything down. When the AE pastes call notes and says "recap this discovery," Claude loads the recap skill, follows the template, and returns CRM-ready output every time.

Skills are portable across Claude.ai, Claude Code, Claude for Small Business, the Anthropic API, and Enterprise. That matters for sales: your AE using Claude in a browser and your RevOps engineer calling Claude from a script both get the same output shape.

• A SKILL.md file with a short description of when Claude should fire the skill.
• Reference material — your ICP definition, MEDDIC rubric, pricing sheet, competitor cheat sheet.
• Optional scripts — a Python helper that hits your CRM API or scrapes a public LinkedIn profile URL.
• Templates — the exact email, recap, or brief shape your team ships.
• Version control — skills live in a Git repo so RevOps can review changes before they hit the field.

Skill 1: Prospect research brief for AEs and SDRs

The prospect research skill produces a one-page account brief in under two minutes. A rep drops in the target company URL and any known contact, and Claude fires the skill because the task matches its trigger — "build a pre-call brief on a prospect account."

The skill loads your ICP definition, your qualifying question bank, and a scoring rubric. It returns firmographic basics, recent trigger events (funding, hiring surges, leadership changes, product launches), the likely buying committee, and three tailored discovery questions grounded in the account's public signals.

One failure mode to design around: reps will paste in noisy LinkedIn scrapes with duplicated bullet points. Add a "dedupe and prioritize" step in SKILL.md so the brief never runs longer than one page. That constraint is what keeps the skill useful during the 90 seconds before a call.

• Trigger: "research this account before a first meeting" or a pasted company URL.
• Reference files: ICP definition, industry-specific pain points, discovery question bank.
• Output: firmographics, three trigger events with dates, buying committee guess, three custom discovery questions.
• Guardrail: never invent a stat — cite the source URL for every claim or drop the claim.
• Best for: SDRs building daily prospecting queues and AEs prepping first meetings.

Skill 2: First-touch email personalization from LinkedIn and site data

The first-touch personalization skill writes an outbound email that references something specific to the prospect — not "I saw you were hiring." The rep pastes a LinkedIn profile URL and a company blog post or press release, and Claude fires the skill.

Personalization at the individual level lifts response rates. Gartner's 2025 research on B2B buyer behavior shows that outbound sequences with a specific and verifiable personal reference in line one see materially higher reply rates than templated equivalents. The trick is making "specific" repeatable across a whole SDR team.

The skill enforces one hard rule: line one must reference something the prospect actually said, wrote, or built in the last 90 days. If the source material is older than 90 days, the skill flags the rep and asks for a fresher signal before drafting.

• Trigger: paste of a LinkedIn URL plus a recent post, article, or product mention.
• Reference files: your firm voice guide, opener bank, three offer variants, subject line library.
• Output: subject line, three-line email, a follow-up variant, and the source URL used for line one.
• Guardrail: refuse to draft if the recency signal is older than 90 days.
• Best for: SDRs running high-volume cold sequences and AEs re-opening cold accounts.

Skill 3: Discovery-call recap to CRM notes

The recap-to-CRM skill turns a raw call transcript into structured CRM notes in your team's exact shape. The AE pastes a Gong, Fireflies, or Zoom transcript, and Claude fires the skill because the task matches "summarize a sales call for CRM."

The output is not a generic summary. The skill reads your CRM field map — the exact stage, pain, next step, and champion fields your ops team requires — and writes a paste-ready block for each. It also flags any required field the transcript did not surface, so the AE knows what to ask next call.

This is the single most requested Layer3 skill build. AEs hate manual CRM hygiene, RevOps hates missing fields at the QBR, and this skill fixes both at once. HubSpot data shows sellers using AI-assisted recap tools save roughly two hours per day of admin time — hours that go back to selling.

• Trigger: paste of a call transcript or "recap yesterday's discovery with Acme."
• Reference files: CRM field map, stage definitions, next-step vocabulary, deal desk requirements.
• Output: summary paragraph, per-field CRM block, next-step list with owner and date, gaps to fill.
• Guardrail: never invent a pain quote — cite the transcript timestamp for every extracted claim.
• Best for: AEs after every first-call and stage-progression meeting.

Skill 4: MEDDIC and BANT qualifier scoring

The qualifier skill scores an open opportunity against your chosen framework — MEDDIC, MEDDPICC, or BANT — and returns a clear go, coach, or disqualify recommendation. The AE points the skill at the deal record or pastes a recent conversation, and Claude fires the skill.

The skill loads your team's explicit rubric for each letter. What does "Metrics" actually mean in your sale? What counts as a real "Champion" versus a coach? Without a rubric, MEDDIC scoring is vibes. With a rubric baked into a skill, every AE grades the same way and the forecast tightens.

The output includes the score, the two weakest letters, and a coaching prompt for the AE: "Ask this exact question on the next call to move Champion from a 2 to a 4." That is where the productivity lift shows up — not in the score itself, but in the next action it forces.

• Trigger: "qualify this deal" or paste of deal notes and CRM export.
• Reference files: your MEDDIC or BANT rubric with definitions and example evidence.
• Output: letter-by-letter score, weakest links, next-call coaching questions.
• Guardrail: any score above 4/5 must cite specific evidence from the notes.
• Best for: AEs prepping deal reviews and sales managers doing pipeline inspection.

Skill 5: Follow-up sequence drafting

The follow-up skill writes a three- to five-touch sequence after any real sales event — a demo, a proposal, a stalled reply, a competitor mention. The AE pastes the last exchange and names the event, and Claude fires the skill.

The skill loads your firm's cadence rules (day spacing, channel mix, breakup timing) and the tone standards your leadership actually approved. It writes each step with a distinct angle — value reminder, third-party proof, question reframe, respectful breakup — so reps stop sending five "just circling back" emails in a row.

The important design choice is refusing to send anything the rep did not review. The skill drafts, but a human presses go. That guardrail keeps outbound compliant with CAN-SPAM and the 2025 FTC guidance on AI-generated commercial mail.

• Trigger: paste of the last email exchange plus a named event ("demo yesterday").
• Reference files: cadence spacing rules, tone guide, opener and CTA banks, breakup templates.
• Output: three to five steps with subject lines, channel, send day, and rationale.
• Guardrail: never auto-send — always deliver drafts for human review.
• Best for: AEs after every demo, POC, and proposal.

Skills 6 and 7: Deal-review prep and competitive battle cards

The deal-review skill produces the exact talking-point sheet your sales manager expects for weekly forecast calls. The AE names the deal, and Claude fires the skill, pulls the current CRM state, and writes a one-page brief: stage, dollar value, MEDDIC gaps, risks, asks for the manager.

The battle-card skill answers the mid-call question every AE dreads — "how do you compare to [competitor]?" The rep asks Claude for the card, and the skill loads your approved positioning against each named competitor: three strengths, two weaknesses, three trap-setting questions, and the price posture.

Both skills solve the same core problem: institutional knowledge trapped in someone's head. Battle cards written six months ago in a Google Doc go stale fast. A skill lets RevOps ship a single update — new competitor pricing, a new feature parity claim — and every rep gets the new version on their next call.

• Deal-review trigger: "prep me for pipeline review on Acme" — output is a one-page brief.
• Battle-card trigger: "how do we position against [competitor]" — output is a three-block card.
• Reference files: approved competitive positioning doc, pricing sheet, feature parity matrix.
• Guardrail: never quote a competitor price or feature claim without a source URL.
• Best for: AEs mid-call and sales managers running weekly pipeline reviews.

Skill 8: Renewal risk narrative for customer success and RevOps

The renewal-risk skill builds a written health-check narrative for any customer approaching renewal. The CSM or account AE points the skill at the account, and Claude fires the skill because the task matches "assess renewal risk."

The skill reads product usage exports, support ticket counts, and the most recent QBR notes. It returns a green, yellow, or red flag with the top three risk drivers and the recommended play — a save motion, an executive alignment call, or a proactive expansion pitch.

For RevOps leaders, this is where AI moves from anecdote to forecast quality. Instead of subjective health scores in a spreadsheet, every CSM writes the narrative the same way, and roll-ups actually compare across the book of business. Layer3 Labs typically ships this skill alongside a CRM integration so the narrative writes back to the account record for the QBR deck.

• Trigger: "assess renewal risk on Acme" or CSM opens the renewal checklist.
• Reference files: usage benchmarks by tier, save-play library, expansion criteria.
• Output: red/yellow/green rating, top three risk drivers, recommended play, executive-summary paragraph.
• Guardrail: flag every claim that lacks a data source rather than making one up.
• Best for: CSMs, account AEs, and RevOps leaders on any book above 20 accounts.

How Layer3 Labs deploys Claude skills for sales teams

Layer3 Labs deploys Claude skills for sales teams in a four-step engagement built for small and midsize B2B sales orgs. Step one is a workflow audit — we sit with two AEs, one SDR, and one RevOps leader and inventory the workflows that eat the most non-selling time.

Step two is skill design. We write the SKILL.md, source the reference material from your Gong library, CRM, and playbook, and stand up a private Git repo your RevOps team owns. Step three is a two-week pilot with three reps, tracked against your baseline metrics — hours saved, reply rates, CRM completeness, forecast accuracy.

Step four is rollout. We train the full team, wire the skills to your CRM through the /ai-crm-integration pattern, and set an ownership cadence so the skills stay current. Layer3 Labs was founded to make AI implementation practical for SMBs — no six-month platform builds, no vendor lock-in.

• Workflow audit: two AEs, one SDR, one RevOps leader, one 90-minute session.
• Skill design: SKILL.md files, reference material, Git repo owned by your team.
• Two-week pilot: three reps, measured lift against baseline.
• CRM wiring: HubSpot, Salesforce, or Pipedrive through the Layer3 integration pattern.
• Ownership cadence: quarterly review, single named owner in RevOps.