Best AI skills for B2B teams
The best AI skills for B2B teams: 10 Claude and Codex skills for call debriefs, lead routing, pipeline hygiene, churn risk, QBRs, RFPs, and SaaS metrics.
By Type Team
What are the best AI skills for B2B teams?
The best AI skills for B2B teams are Call Debrief, Lead Triage Router, Pipeline Hygiene Audit, and Churn Risk Scanner. Together they keep call context in the CRM, score and route inbound leads, flag bad pipeline data before the forecast call, and rank renewals by ARR at risk.
An AI skill is a packaged set of instructions, often with scripts and sample data, that Claude or Codex follows the same way every time it runs.
B2B go-to-market teams repeat the same rules-heavy work every week: debriefing calls, routing inbound, cleaning deals before the forecast, prepping QBRs, answering RFPs, and reporting MRR and NRR. That work feeds the forecast, the board deck, and renewal decisions, yet its context is scattered across call recordings, the CRM, and Slack. Skills make it consistent, and most of the ones below show their math.
Every skill in this guide is published in the Type Skills Library with its full files, and each one can be installed into a shared Space for your B2B team, where Claude or Codex works from your CRM, product data, support tools, and the rest of your stack.
| Skill | Best for | What you get |
|---|---|---|
| Call Debrief | AEs, SDRs, and CSMs right after a customer call or when a new recording lands | A debrief under 200 words with direct quotes, a proposed CRM update, and a drafted follow-up |
| Lead Triage Router | RevOps and SDR managers triaging inbound form fills, trials, and downloads daily or hourly | A tiered, routed lead list with per-rep load, the fit and intent math behind each score, and data gaps |
| Pipeline Hygiene Audit | Sales managers and RevOps before a weekly forecast call, QBR, or board pipeline review | Open pipeline value with data problems, duplicate account clusters, and a ranked fix list per owner |
| Churn Risk Scanner | CSMs and CS leaders in a weekly or monthly renewal review, and before the renewal forecast | Your book ranked by weighted ARR at risk, with top risk factors and pasteable evidence per account |
| QBR Deck Builder | CSMs and account managers preparing a QBR, EBR, renewal, or account handover | Deck-ready sections on adoption, goal attainment, support health, risks, and ranked expansion |
| RFP Response Builder | Sales, solutions, and proposal teams the day an RFP lands, for bid/no-bid calls | Coverage score, mandatory blockers, points at risk, best-match answers with owners, and stale answers |
| SaaS Metrics Pack | Finance and RevOps for monthly MRR reporting, board packs, investor updates, and diligence | An MRR waterfall, NRR, GRR, logo churn, quick ratio, ARPA, CAC, CAC payback, and a tie-out check |
| Failed Payment Recovery | Finance and billing ops, triggered by a payment-failure webhook or an hourly schedule | Classified failures, scheduled retries and up to four emails per invoice, a touch ledger, and a weekly summary |
| GEO / AI Search Audit | B2B marketing and demand gen teams checking if AI answer engines can read your site | A draft report with a prioritized fix list, plus a draft llms.txt and Organization JSON-LD if missing |
| Cash Runway Forecaster | Finance leads and founders in the weekly treasury cycle, before board meetings or a raise | Weekly closing balances for 13 weeks, runway in weeks, the first covenant breach, and scenarios |
Which AI skills should B2B go-to-market teams use?
These 10 skills cover the path from first call and inbound lead through renewal, board reporting, and cash.
The list follows the revenue lifecycle. New business comes first (calls, inbound leads, pipeline), then customer success (churn risk and QBRs), then deal support (RFPs), then the finance and marketing work GTM teams share with other functions (SaaS metrics, failed payments, AI search readiness, and cash runway). Not sure where to begin? Start with the first three or four.
1. Call Debrief
Call Debrief turns a sales or customer call recording into a short structured debrief, a CRM record that matches what was actually said, and a drafted follow-up, without writing to the CRM until someone approves.
Call context is usually the first thing lost between the recording, the CRM, and Slack. Call Debrief quotes the buyer on budget, timeline, authority, and competitors, says plainly when a call ended without a next step, and flags calls with no matching CRM record, which usually means the deal is missing from the pipeline. It skips short, internal, and already-debriefed calls.
- Best for
- AEs, SDRs, and CSMs right after a customer call or when a new recording lands
- Works from
- A Granola, Fathom, Gong, Fireflies, or pasted transcript, plus HubSpot, Attio, or Salesforce if connected
- You get
- A debrief under 200 words with direct quotes, a proposed CRM update, and a drafted follow-up
2. Lead Triage Router
Lead Triage Router scores inbound leads on ICP fit and intent with a rubric you control, assigns a tier and an owner from your territory rules, and shows reps exactly why each lead reached them.
Speed-to-lead falls apart once inbound volume outgrows a person reading every form fill. SDRs get a ranked, routed list with the reasoning attached, and disqualified leads (personal email domains, companies below your size floor) go to nurture with the reason noted. It does not assign owners in the CRM or enroll anyone in a sequence.
- Best for
- RevOps and SDR managers triaging inbound form fills, trials, and downloads daily or hourly
- Works from
- CSV exports of leads and routing rules, pulled from HubSpot, Salesforce, or Attio plus product analytics
- You get
- A tiered, routed lead list with per-rep load, the fit and intent math behind each score, and data gaps
3. Pipeline Hygiene Audit
Pipeline Hygiene Audit checks a CRM deal export for stale deals, past close dates, missing amounts or owners, skipped stages, stuck deals, and duplicate accounts, then gives each owner a prioritized list to fix.
A forecast is only as good as the records under it. The audit leads with how much open pipeline carries a data problem and weights each finding by deal size, so a $500k deal with a past close date outranks a $5k deal with the same problem. It judges the record, not the deal, and never changes anything in the CRM.
- Best for
- Sales managers and RevOps before a weekly forecast call, QBR, or board pipeline review
- Works from
- CSV exports of deals, stage order, and optional stage history from Salesforce, HubSpot, or Attio
- You get
- Open pipeline value with data problems, duplicate account clusters, and a ranked fix list per owner
4. Churn Risk Scanner
Churn Risk Scanner scores renewal risk for every account from usage decline, license utilization, support burden, CSAT, champion departure, and invoice history, then ranks the whole book by weighted ARR at risk.
The signals that usually precede a non-renewal (falling usage, a departed champion, late invoices) live in different systems. Ranking by ARR at risk separates a large account with a small problem from a small account with a serious one, and missing data shows as NO DATA instead of being scored as a happy customer. It is correlational, not predictive, and writes nothing back.
- Best for
- CSMs and CS leaders in a weekly or monthly renewal review, and before the renewal forecast
- Works from
- Five CSV exports: accounts, monthly usage, support tickets, contacts, and invoices
- You get
- Your book ranked by weighted ARR at risk, with top risk factors and pasteable evidence per account
5. QBR Deck Builder
QBR Deck Builder assembles one customer's quarterly business review from usage, goal, support, and commercial data into deck-ready sections covering adoption trend, value delivered against stated goals, open risks, and ranked expansion opportunities.
The slide that earns the renewal is the customer's own goals with progress shown, and this skill calculates attainment from baseline to target for each one. It also makes account handovers cleaner, because the state of the account is written down with evidence instead of living in one CSM's head. Output is text or JSON, not slides.
- Best for
- CSMs and account managers preparing a QBR, EBR, renewal, or account handover
- Works from
- CSV exports of usage, goals, support tickets, commercial terms, and your product catalog, one account per run
- You get
- Deck-ready sections on adoption, goal attainment, support health, risks, and ranked expansion
6. RFP Response Builder
RFP Response Builder sorts an RFP's requirements into mandatory, weighted, and optional, matches each one against your answer library, and reports what you can already answer, what will cost points, and what nobody can answer yet.
The expensive RFP mistake is finding a mandatory gap the week the response is due. Run this the day the RFP arrives: the blocker count is your bid risk, and the ranked gap list is your SME assignment list. Matching is keyword-based and the skill does not write the response, so keep answer tags rich and have a person tailor every answer.
- Best for
- Sales, solutions, and proposal teams the day an RFP lands, for bid/no-bid calls
- Works from
- CSV exports of the RFP requirements and your answer library (Loopio, Responsive, Notion, Confluence, Drive)
- You get
- Coverage score, mandatory blockers, points at risk, best-match answers with owners, and stale answers
7. SaaS Metrics Pack
SaaS Metrics Pack turns a subscription export into the full MRR waterfall (new, expansion, contraction, churn, reactivation) plus NRR, GRR, logo churn, quick ratio, ARPA, CAC, and CAC payback, and fails if the numbers do not tie.
When a board member or investor asks how you calculated churn, you need the per-customer classification behind the number. The pack reconciles opening MRR plus movements to closing MRR, can check it against your billing system's closing figure, and refuses to report if it does not tie. Its NRR excludes reactivations and its CAC is blended across all new logos.
- Best for
- Finance and RevOps for monthly MRR reporting, board packs, investor updates, and diligence
- Works from
- One subscription CSV with full history, exported from Stripe, Chargebee, Recurly, Paddle, or the GL
- You get
- An MRR waterfall, NRR, GRR, logo churn, quick ratio, ARPA, CAC, CAC payback, and a tie-out check
8. Failed Payment Recovery
Failed Payment Recovery finds failed charges and past-due invoices, classifies each decline, and runs a retry and outreach ladder with approval gates, recording every touch so no customer is contacted twice.
Failed payments are churn nobody chose. By default the skill holds every retry and email until someone approves its first draft (then runs on its own unless set to approve every send), caps outreach at four touches per invoice, keeps quiet hours of 9pm to 8am in the customer's time zone, and never offers a discount without approval. Failures older than its 14-day ladder go to a person as one summary, and with a read-only billing connection it only reports what could be recovered.
- Best for
- Finance and billing ops, triggered by a payment-failure webhook or an hourly schedule
- Works from
- A connected billing system (Stripe, Recharge, or another via Custom API) and a way to send customer email
- You get
- Classified failures, scheduled retries and up to four emails per invoice, a touch ledger, and a weekly summary
9. GEO / AI Search Audit
GEO / AI Search Audit checks whether ChatGPT search, Perplexity, Claude, Google AI Overviews, and Copilot can reach, read, and quote your website, then turns the findings into a prioritized fix list.
When a buyer asks an answer engine to compare tools in your category, your site has to be crawlable and quotable to show up. The audit scores AI crawler access in robots.txt, passage citability, structured data, llms.txt, and whether pages render without JavaScript. It reads only public pages (five per site by default), and its scores measure readiness, not actual AI citations.
- Best for
- B2B marketing and demand gen teams checking if AI answer engines can read your site
- Works from
- Your website URL, or a profile file listing your site and competitor URLs for a side-by-side benchmark
- You get
- A draft report with a prioritized fix list, plus a draft llms.txt and Organization JSON-LD if missing
10. Cash Runway Forecaster
Cash Runway Forecaster builds a 13-week direct cash forecast from AR, AP, payroll, and recurring commitments, showing weekly closing balances, runway in weeks, and the first week cash drops below your minimum.
For a B2B company, runway depends on when customers actually pay their invoices, not on an average burn rate. This forecast schedules real receivables and payables, tests scenarios such as slower collections or stretched payables, and reconciles every input row so nothing is silently dropped. It only counts new sales you enter as expected billings.
- Best for
- Finance leads and founders in the weekly treasury cycle, before board meetings or a raise
- Works from
- Cleared opening bank balance plus CSV exports of AR ageing, AP ageing, payroll, and recurring commitments
- You get
- Weekly closing balances for 13 weeks, runway in weeks, the first covenant breach, and scenarios
How should B2B teams choose which AI skills to start with?
Start with recurring work that has a clear right answer and a clear owner, run the skill on its bundled sample data, then point it at your own CRM export.
The best first skill is the one attached to a meeting you already hold. Pipeline Hygiene Audit fits the weekly forecast call, Churn Risk Scanner fits the renewal review, and Lead Triage Router fits the daily inbound queue. Each has a clear right answer (a stale deal is stale, a missing owner is missing), which makes the output easy to check.
Eight of the ten ship a script, a data contract, sample data, and an expected output, and tell Claude or Codex to run the sample first and stop if it does not match, so RevOps can check the output before any real pipeline data is involved. Call Debrief and Failed Payment Recovery are instruction-only and work through connected tools. Thresholds such as stale-deal days, tier cutoffs, and risk weights live in config files, so tuning them is a reviewable edit, not a meeting.
Pay attention to which skills act and which only report. Eight of the ten produce reports, drafts, and work lists and change nothing in your systems. Call Debrief asks before each CRM update by default, and Failed Payment Recovery holds every card retry and customer email until someone approves its first draft.
- Pick work that repeats on a schedule: the forecast call, the renewal review, the inbound queue.
- Give each skill one owner, usually the person who runs that meeting today.
- Run the bundled sample first and compare it with the expected output.
- Tune thresholds in config, then rerun on last month's data before trusting them.
- Keep approval on for any skill that writes to the CRM or the billing system.
How do B2B teams run these skills together in Type?
In Type, a B2B team connects its CRM, call recorder, billing, and support tools once, installs skills from the library into a shared Space, and runs them in Slack, in the Type apps, or on a schedule.
Teams can create a Space per function, such as sales, customer success, and finance, and assign each Space only the tools it needs, with team and individual permissions. Claude or Codex, on the subscriptions you already pay for, can then pull the CRM, billing, or support export a skill expects instead of someone building the CSV by hand.
Recurring runs can become automations. For example, a team might schedule Lead Triage Router every morning, Pipeline Hygiene Audit before its Monday forecast call, and Churn Risk Scanner once a month. Results land in the Space or in Slack, where the whole team sees the same debriefs, flags, and fix lists, and when one person improves a skill, everyone gets the improvement.
See how B2B teams set this up on the Type for B2B teams page, or browse more skills in the Type Skills Library.
Frequently asked questions
What is an AI skill for B2B teams?
An AI skill is a reusable, packaged set of instructions, often with scripts and sample data, that Claude or Codex follows the same way every time. For B2B teams, skills cover recurring go-to-market work such as call debriefs, lead routing, pipeline hygiene, churn risk, QBR prep, and RFPs. When one person improves a skill in Type, the whole team gets the improvement.
How is an AI skill different from a prompt?
A prompt is a one-off request, and the result depends on who wrote it. A skill packages the instructions, rules, and configuration, often with a script and sample data, so the work is done the same way for every rep. Skills like Pipeline Hygiene Audit also show their arithmetic, so a manager can defend each flag in the forecast call.
Do these AI skills work with Claude and Codex?
Yes. In Type, your team can run these skills with Claude or Codex, using the subscriptions you already pay for. The skill packages are plain SKILL.md folders, so they also work in other agent environments that read SKILL.md, such as Claude Code and Codex.
Will these AI skills change anything in our CRM?
Most of them do not. Pipeline Hygiene Audit, Lead Triage Router, Churn Risk Scanner, QBR Deck Builder, and RFP Response Builder work from exported data and return reports without writing back to the CRM. Call Debrief can update CRM records once it confirms write access, but by default it asks before each update and only drafts follow-ups. Failed Payment Recovery acts in your billing system instead, retrying cards and emailing customers after someone approves its first draft.
Can we customize these skills for our ICP, territories, and sales process?
Yes. Lead Triage Router keeps its ICP weights, tiers, and disqualifiers in a config file and its territory rules in a routing sheet. Pipeline Hygiene Audit and Churn Risk Scanner keep their thresholds and factor weights in config too. Change them to match your sales motion, then rerun on a recent week or month of data before trusting the results.
Where can I find more AI skills for B2B teams?
The Type Skills Library at type.com/library/skills publishes every skill in this guide with its full files, which for most include scripts, data contracts, and sample data. Each skill has its own page, and you can install it into Type or use its SKILL.md folder in another agent environment.