Buyer Personas & Propensity

The framework — strategy first

Personas & Propensity — the human layer inside the ICP.

Why personas come after ICP and before positioning.

ICP TELLS YOU THE ACCOUNT. PERSONAS TELL YOU THE HUMAN.

The Ideal Customer Profile (ICP) work in the prior section identifies which accounts deserve your investment. But accounts don’t sign contracts — humans do. The persona work names the humans, captures what each one wants and fears, and produces the one-page profile every downstream prompt reads from. Without personas, your positioning is generic, your campaigns target job titles, and your sales conversations sound the same to every role on the buying committee. With personas, every channel speaks to a specific human with specific stakes.

Propensity-to-buy is the layer most teams never add. It answers the question: of all the personas inside our ICP accounts, which ones are showing buying readiness this quarter? The propensity signal is what turns persona work from documentation into a live targeting layer.

The seven dimensions of a usable persona.

A persona that fits on a slide is decoration. A persona that’s usable across every downstream prompt captures seven dimensions, and each one feeds a specific operational decision elsewhere in the playbook.

DIMENSION	WHAT TO CAPTURE	WHERE IT’S USED
Goals	What this role is trying to achieve this year — the outcome that determines their bonus, their next promotion, the conversation they want to have with the board.	Hero messaging on landing pages, sales discovery, email subject lines.
Priorities	The 3–5 things they actually spend time on each week, ranked. Different from goals — goals are what they aim for; priorities are what they do.	Campaign sequencing, content calendar, what to lead with in a demo.
Challenges	The friction inside their organization that gets in the way of their goals. Not the product’s problem space — the human’s problem space.	Pain-led ad copy, customer-story selection, problem framing in webinars.
Objections	The recurring reasons this persona walks away. Cost, return-on-investment (ROI) uncertainty, risk aversion, vendor-transition fatigue, complexity, time constraints, internal politics.	Sales objection-handling, FAQ pages, pricing-page disarmament copy.
Metrics	The numbers this role gets measured on. The Chief Financial Officer (CFO) watches cost containment, fiscal responsibility, ROI, risk. The Chief Revenue Officer (CRO) watches market share, sales growth, retention.	Proof-point selection, dashboard mockups, case-study metric callouts.
Watering holes	Where this persona reads, listens, attends, and posts. Specific publications, specific podcasts, specific conferences, specific Slack and LinkedIn communities.	Paid media targeting, PR pitching, event selection, sponsorship choices.
Trigger events	The organizational moments that move this persona from passive to active — a new exec hired, a quarter missed, a regulatory deadline, a competitor announcement, a layoff.	Intent-data lists, Sales Development Representative (SDR) cadence selection, Account-Based Marketing tier-1 trigger watch — AND the propensity-to-buy signals below.

Per-persona deep dives — worked examples for the four buying-committee roles.

Most B2B SaaS buying committees have four canonical roles. Below, the seven dimensions filled in for each one as a worked example. Adapt to your specific category — the dimensions stay the same; the contents change.

Economic Buyer — the CFO / VP Ops / CRO

DIMENSION	FOR THE ECONOMIC BUYER
Goals	Hit the number this year. Deliver the cost-savings or margin-expansion commitment made to the board. Defend the budget line at the next review.
Priorities	Cost containment, ROI defensibility, risk mitigation, board-level financial reporting, capital allocation.
Challenges	Too many vendor requests for budget; can’t separate must-fund from nice-to-fund. Existing tools underperform but switching costs are real. The team below them keeps requesting tools without ROI math.
Objections	“Too expensive.” “ROI is unclear.” “Switching is too risky.” “Doesn’t perform at our scale.”
Metrics	ROI, payback period, customer-acquisition cost, gross margin, free cash flow, opex as % of revenue.
Watering holes	WSJ CFO Journal, The CFO Show, McKinsey/Bain CFO research, CFO Leadership Council, peer roundtables, LinkedIn (career moves + earnings cycles).
Trigger events	Quarter missed; budget cycle opening; new CEO hired; M&A activity; major vendor renewal coming up; recent earnings call where leadership committed to a margin target.

Champion — the Director / VP of the function buying the product

DIMENSION	FOR THE CHAMPION
Goals	Deliver on the operational mandate. Get promoted. Avoid the visible failure that ends the career.
Priorities	Team productivity, process maturity, removing the manual work that consumes their Fridays, building a function that compounds without their daily attention.
Challenges	Team is short-staffed. Tools don’t talk to each other. Reporting to the Economic Buyer requires data they don’t have. Internal politics around whose budget owns the new tool.
Objections	“Will my team adopt it?” “Will it break the existing workflow?” “Is the implementation effort worth the outcome?” “What if it fails — who carries that?”
Metrics	Team output per headcount, time-to-value for new tools, retention of their direct reports, the operational KPIs their function reports.
Watering holes	Function-specific publications and podcasts (Marketing Brew for marketing leaders, the Operations Room for ops, the Modern CFO for finance). Peer Slack and Discord communities. LinkedIn (long-form posts from function peers).
Trigger events	Recent promotion (they want to ship something visible in the first 90 days). New executive above them (they want to look good for the new boss). Team scaling up or down. Operational incident that exposed a gap.

Technical Evaluator — IT, Security, Engineering leadership

DIMENSION	FOR THE TECHNICAL EVALUATOR
Goals	Don’t get blamed for the next security incident. Don’t approve a tool that creates technical debt. Keep the stack consolidating, not expanding.
Priorities	Security posture, compliance, integration depth, data privacy, vendor consolidation.
Challenges	Too many vendor security reviews; quality is uneven. Lines of business buy tools without IT’s input. The team can’t catch every privacy/compliance edge case in a 30-day review.
Objections	“What’s your SOC 2 (Service Organization Control 2) status?” “Do you integrate with our identity provider (IdP)?” “Where does the data live?” “What’s your sub-processor list?” “How do you handle data deletion under the General Data Protection Regulation (GDPR)?”
Metrics	Time-to-security-review-complete, number of security incidents traced to vendors, vendor consolidation ratio, integration coverage.
Watering holes	Dark Reading, The Chief Information Security Officer (CISO) Series, Risky Business podcast, RSA Conference, regional CISO peer dinners, the CHIME network for healthcare-specific.
Trigger events	Recent security incident industry-wide. Compliance framework update (new SOC 2 Type II requirement, new GDPR enforcement action). Annual security review window. New CISO hired.

End User — the person doing the work daily

DIMENSION	FOR THE END USER
Goals	Get the work done. Spend less time on the parts that drain. Look competent to their manager.
Priorities	Daily workflow, mobile experience (if frontline), training time, ease of learning new tools.
Challenges	Tools they didn’t choose. Frequent context-switching between systems. Training they didn’t have time for. The system always tells them they did it wrong.
Objections	“Another login.” “I don’t have time to learn this.” “The old way worked fine.” “Will I get blamed when this breaks?”
Metrics	Their own operational KPIs (orders processed, shifts covered, calls handled). Time-on-tool. Error rate.
Watering holes	Practitioner blogs and YouTube channels (not corporate publications). Hands-on workshops, certification programs. TikTok, Instagram, YouTube tutorials. Manager-focused Facebook groups.
Trigger events	New manager (they need to prove themselves). Software change driving frustration with current tool. Personal upskilling moment.

Worked example — the four roles inside a fintech buying committee

To show how the four roles map to a specific category, here’s the buying committee for a B2B SaaS selling a regulatory-reporting platform to mid-market US commercial banks ($1B–$25B in assets). Same framework, vertical-specific contents.

ROLE	THE HUMAN INSIDE A MID-MARKET BANK	WHAT THEY CARE ABOUT
Economic Buyer	Chief Risk Officer (CRO) or Chief Compliance Officer (CCO).	Audit-readiness, regulator relationships, defensibility of the control environment, total cost of compliance as a margin line.
Champion	Head of Regulatory Reporting — typically reports into the CRO or CFO.	Reducing the manual rework cycle behind each filing, shrinking the close-to-submission window, hitting filing dates without weekend escalations.
Technical Evaluator	Head of Data Engineering or Director of IT.	Integration with the core banking system, data lineage controls, model-risk management governance, single-sign-on (SSO) and data-residency posture.
End User	Senior Compliance Analyst or Risk Reporting Manager.	Daily reconciliation workflow, exception handling, audit trail per data point, the ability to explain a number to an examiner the next day.

The named competitive set in this category: Wolters Kluwer OneSum, AxiomSL (now Adenza), Workiva, Moody’s Analytics. The trigger events that move this committee from passive to active: a new Federal Reserve or Office of the Comptroller of the Currency (OCC) examination finding; a regulatory-reporting framework update (Federal Financial Institutions Examination Council, FFIEC, call-report revisions); a new CRO hired; a missed filing; a merger announcement that suddenly doubles the entity scope. The recurring phrase in customer calls: “we need to be able to defend every number in front of an examiner.” That phrase becomes the language that downstream prompts read from.

Propensity-to-buy — the signal layer that turns personas live.

THE PROPENSITY LAYER — PER-PERSONA BUYING-READINESS SIGNALS

A persona profile tells you who the buyer is. A propensity signal tells you when they’re actually shopping. Most buying committees are passive 90% of the time and active 10% of the time — the propensity-to-buy layer surfaces the active moments before competitors notice.

Propensity signals are per-persona. The CFO shows different signals than the Director of Ops than the IT Evaluator. The discipline is naming the 3–5 highest-leverage signals per role and instrumenting the watch for them.

High-propensity signals by persona

PERSONA	HIGH-PROPENSITY SIGNALS	WHERE TO WATCH
Economic Buyer (CFO/VP Ops)	Earnings call mentions cost pressure or margin commitment; quarterly budget review on the calendar; recent hire of a CFO; M&A activity announced; missed quarter.	Public earnings transcripts, LinkedIn for hiring signals, news monitoring, intent-data vendors flagging budget-cycle research.
Champion (Director/VP function)	Recent promotion to the role (within 6 months); function hiring spree (2+ open reqs); attendance at industry conferences; LinkedIn post about a function pain point that maps to your product.	LinkedIn for promotions and posts, job boards for function hiring, conference attendee lists, your own first-party intent (case-study downloads, demo requests, podcast appearances).
Technical Evaluator (IT/Security)	Recent SOC 2 / ISO renewal announcement; new CISO hired; security incident industry-wide; compliance framework update relevant to their industry.	Compliance announcement databases, security incident trackers (Have I Been Pwned, KrebsOnSecurity), regulatory body updates (the Department of Health and Human Services, HHS, for the Health Insurance Portability and Accountability Act, HIPAA; the Securities and Exchange Commission, SEC, for financial).
End User (frontline / IC)	High volume of negative reviews of the incumbent tool on G2 or Glassdoor; YouTube tutorial volume on the incumbent dropping; community Slack/Discord complaint patterns.	G2 / Capterra review streams, Glassdoor mentions of specific tools, social listening on practitioner communities.

Three rules. Single-signal propensity is noise. One signal in isolation rarely predicts buying; three independent signals firing within 60 days is the real predictor. Propensity is account-level + persona-level. An account is high-propensity when MULTIPLE personas inside it are firing signals simultaneously. Propensity decays. A signal fired 90 days ago is mostly noise today; the watch needs to be continuous, not annual.

The propensity score — how to compute it

PROPENSITY SCORING — THE 100-POINT MODEL

Score each account on a 100-point scale derived from the persona signals firing in the last 90 days. Use the bands below to route accounts to the right outreach motion.

80–100 = Hot. 3+ high-propensity signals firing across 2+ personas. Route to sales with a same-week outreach. ABM Tier-1 treatment.
60–79 = Warm. 2 signals firing or 1 high-value signal (e.g., a confirmed budget allocation). Route to nurture sequence designed to convert to discovery within 30 days.
40–59 = Watching. 1 signal firing. Surface in dashboards; no immediate outreach. Watch for second signal.
Under 40 = Cold. No fresh signals. Keep in the broad marketing universe; don’t allocate sales capacity.

The buyer process map — from trigger to advocate.

The persona work and the propensity score lead to the buyer process map: the journey from the trigger event to the closed deal to the long-term advocate. Five canonical stages, each with a persona transition and a propensity-to-buy threshold.

STAGE	WHAT’S HAPPENING	ACTIVE PERSONAS	MARKETING OWNS
1. Unaware	The buyer has the problem but doesn’t know your category solves it. Propensity score < 40.	End User feels the pain. Champion may have surfaced it. CFO and IT not engaged.	Brand-led awareness, problem-framing content, organic reach to the End User community.
2. Aware	The buyer knows the category exists. Propensity score 40–59.	Champion is researching. End User is reading. CFO might be Cc’d on emails.	Educational content (webinars, comparison guides), nurture sequences, LinkedIn organic for the Champion role.
3. Considering	The buyer is actively evaluating you against alternatives. Propensity score 60–79.	Champion is leading. IT Evaluator is reviewing. CFO is watching from a distance.	Demos, case studies, comparison content, review-site investment, security documentation pre-staged.
4. Decision	The buyer is finalizing the decision against you vs. 1–2 others. Propensity score 80+.	Champion is selling internally. CFO is reviewing the business case. IT is completing the security review. End User pilots happening.	References, ROI defense memos, security documentation finalized, executive sponsor exchanges.
5. Customer → Advocate	Closed-won, deployed, expanding, referenceable. Propensity score on adjacent purchases is rising.	End User is the daily voice. Champion is the public reference. CFO sees the ROI report.	Customer marketing, expansion campaigns, reference cultivation, community programs, advocacy content.

Refresh cadence and ownership.

Personas drift — not weekly, but meaningfully every 12–18 months. A CFO’s priorities in 2024 are not the same as in 2026. Propensity signals drift faster — what worked as a signal six months ago may have been gamed by competitors by now. The discipline:

Personas refresh annually. Re-interview five customers per role. Update the seven dimensions. Re-validate the watering holes — they shift fastest.
Propensity signals refresh quarterly. Audit the signals against actual closed-won deals from the last quarter. Which signals correlated with wins? Which were noise? Adjust the score weights.
The Persona Researcher Agent watches both. The agent runs the refresh on cadence and flags drift to the named human owner (typically the Product Marketing Manager, PMM, Lead or Head of Marketing).

CAPTURE IN YOUR OPERATOR BRIEF — SECTION 3

The brief-builder cards below collect the four personas (Economic Buyer, Champion, Technical Evaluator, End User) plus the propensity signal layer. Once Section 3 is populated, every prompt here — and every downstream prompt in the playbook — uses your real personas.

Brief-builder: Economic Buyer profile

Economic Buyer — seven dimensions

The person who signs the check. Capture all seven dimensions. Save to Brief Section 3.1.

Title and reporting line

Goals + Priorities + Challenges (combined)

Top 3 objections + the disarming response

Watering holes (publications, events, communities)

High-propensity signals (3–5)

Saved to Brief Section 3.1. Available as [ECONOMIC BUYER PERSONA] in every prompt.

Brief-builder: Champion profile

Champion — seven dimensions

The internal advocate. Same structure as the Economic Buyer. Save to Brief Section 3.2.

Title and reporting line

Goals + Priorities + Challenges

Top 3 objections + disarming response

Watering holes

High-propensity signals

Saved to Brief Section 3.2. Available as [CHAMPION PERSONA] in every prompt.

Brief-builder: Technical Evaluator profile

Technical Evaluator — seven dimensions

The person who clears integration and security. Save to Brief Section 3.3.

Title and reporting line

Goals + Priorities + Challenges

Top 3 objections + disarming response

Watering holes

High-propensity signals

Saved to Brief Section 3.3. Available as [TECHNICAL EVALUATOR PERSONA].

Brief-builder: End User profile

End User — seven dimensions

The person who uses the product daily. Save to Brief Section 3.4.

Title and context (mobile, in-office, frontline, etc.)

Goals + Priorities + Challenges

Top 3 objections + disarming response

Watering holes

High-propensity signals

Saved to Brief Section 3.4. Available as [END USER PERSONA].

Brief-builder: propensity score model

Propensity-to-buy scoring model

How you compute the 100-point score and where each account lands. Save to Brief Section 3.5.

Signal weights (e.g., earnings-call mention = 25, new CFO hire = 20, function hiring spree = 15...)

Hot / Warm / Watching / Cold thresholds (default: 80+/60–79/40–59/under 40)

Decay rule (default: signals older than 90 days lose 50% of weight)

Saved to Brief Section 3.5. Used by ABM Tier-1 selection, paid-media targeting, SDR cadence.

The prompt pack

Five prompts to build the personas and run the propensity loop.

Prompt 1

Persona deep-dive synthesis from customer interviews

Hand the prompt 3–5 customer-call transcripts segmented by buying-committee role. Returns the seven dimensions filled in for that persona, with verbatim quotes attached.

You are a persona researcher for [COMPANY NAME]. I am attaching transcripts from [N] customer conversations with people in the [PERSONA ROLE] role at [ICP segment]. For each conversation, the questions were: What are you trying to achieve this year? What gets in the way? What did you almost buy instead of us? Where do you read / listen / attend? What would make you actively shop for a new vendor like us? Synthesize across the transcripts and return: 1. Goals (top 3 with verbatim quote support) 2. Priorities (top 3 with verbatim quote support) 3. Challenges (top 3 with verbatim quote support) 4. Objections (top 3 with the root cause underneath) 5. Metrics they get measured on 6. Watering holes (specific publications, events, communities) 7. Trigger events that move them from passive to active Return as a one-page persona profile. Quote verbatim where possible.

Prompt 2

Objection-handling response generator

For a given persona objection, returns the root cause analysis + the response that disarms it + the proof point that backs it. The single most-used prompt in sales enablement.

You are an objection-handling strategist for [COMPANY NAME]. Persona: [PERSONA — e.g., CFO] Objection: [OBJECTION VERBATIM — e.g., “The solution is too expensive.”] Our product: [POSITIONING] Our right-to-win: [RIGHT TO WIN] Available proof points: [LIST OF PROOF POINTS — case studies, ROI data, certifications] Return: 1. The root cause underneath the objection (what they’re really worried about) 2. The response that disarms the root cause (one paragraph, operator-direct) 3. The specific proof point from the list above that backs the response 4. The follow-up question to ask the prospect after delivering the response 5. The signal that tells you the objection has been resolved vs. is still live Format for paste into a battlecard.

Prompt 3

Watering-hole map for paid media + PR

For a given persona, returns the publication / podcast / event / community map ranked by reach inside that persona’s segment. Feeds paid-media targeting and PR pitching.

You are a media strategist for [COMPANY NAME]. Persona: [PERSONA — e.g., Head of Regulatory Reporting at US commercial banks $1B–$25B in assets] Vertical: [VERTICAL] Geography: [GEOGRAPHY] Build the watering-hole map for this persona. Return: 1. Top 5 publications + their reach inside this persona segment 2. Top 5 podcasts + episode pattern that signals high listenership 3. Top 5 conferences/events (annual, with rough attendance numbers) 4. Top 5 communities (Slack, Discord, Reddit, LinkedIn groups, professional associations) 5. The single highest-leverage placement for a PR pitch right now and why 6. The single highest-leverage paid-media targeting parameter Cite sources for reach claims where possible. Flag low-confidence guesses explicitly.

Prompt 4

Propensity score for an account

Hand the prompt an account’s recent signals (earnings transcripts, LinkedIn changes, news, intent data). Returns the 100-point propensity score with reasoning per persona.

You are a propensity-to-buy scorer for [COMPANY NAME]. Account: [ACCOUNT NAME] Our ICP: [ICP DEFINITION] Our personas: [ECONOMIC BUYER PERSONA], [CHAMPION PERSONA], [TECHNICAL EVALUATOR PERSONA], [END USER PERSONA] Propensity scoring weights: [PROPENSITY WEIGHTS] Recent signals (last 90 days) for this account: - Earnings transcript excerpts: [EXCERPTS] - LinkedIn changes for known personas: [CHANGES — e.g., new CFO, function hiring spree] - News mentions: [MENTIONS] - Intent-data signals (if available): [INTENT DATA] - Our own first-party signals (case-study downloads, demo requests, podcast appearances): [FIRST-PARTY SIGNALS] For each persona, return: 1. The signals firing for that persona in the last 90 days 2. The points contributed to the propensity score per signal 3. The persona-level subtotal Then return the account-level total propensity score and the band (Hot / Warm / Watching / Cold) with the recommended outreach motion.

Prompt 5

Persona stress-test for messaging

Simulates all four personas reading a draft message (ad copy, email, landing page, sales pitch) and surfaces objections, gaps, and missing proof per role.

You are running a Persona stress-test for [COMPANY NAME]. Draft message: [MESSAGE DRAFT] Channel: [CHANNEL — landing page, ad, email, sales pitch, etc.] Simulate all four personas from Brief Section 3: 1. Economic Buyer ([ECONOMIC BUYER PERSONA]) 2. Champion ([CHAMPION PERSONA]) 3. Technical Evaluator ([TECHNICAL EVALUATOR PERSONA]) 4. End User ([END USER PERSONA]) For each persona, generate: - First reaction to the message (one sentence, in their voice) - The objection they would raise - The proof point that would disarm the objection - The single word or phrase from the message that resonates or rings false Then synthesize across all four: - Where the message is strong - Where it is weakest - The one revision that would strengthen it most Return as a markdown table per persona, then the synthesis below.

The agent

Persona Researcher Agent.

Persona Researcher Agent

Maintains the four-role buying-committee profiles + the propensity-to-buy signal layer. Reads CRM + product analytics + community engagement to surface who’s in-market right now per persona.

Who is this agent

Identity card

NamePersona Researcher Agent

RoleBuyer persona maintenance + propensity scoring — the ‘who’s ready to buy’ layer

OwnerHead of Product Marketing

Reports toVP Marketing

Versionv0.5 (supervised)

SurfaceClaude Project + Postgres (persona profiles + propensity scores) + Python for scoring

Output target/personas/profiles/ + /personas/propensity-scores/<contact-id>.json + monthly digest

Review cadenceMonthly persona refresh; weekly propensity score audit; quarterly buying-committee study

Mission

Maintain the four-role buying-committee profiles (Economic Buyer, Champion, User, Technical / Compliance) for the ICP. For each persona, maintain a seven-dimension profile: goals, priorities, challenges, objections, success metrics, watering holes, trigger events. Layer a propensity-to-buy signal model on top — reading CRM + product analytics + community + intent data — to surface who’s in-market right now per persona, and route those signals into the playbook.

Goals & KPIs the agent moves

Leading indicators — the agent controls these

Per-persona profile freshness — no profile > 6 months unrefreshed100%

Persona-mapped content coverage — the 7 dimensions × 4 personas matrix has at least one piece of content per cell≥ 80% of cells covered

Lagging indicators — downstream outcomes with review triggers

Propensity-score discrimination — top-decile conversion rate vs. bottom-decile. Trigger: ratio drops below 2× for any single quarter pages the VP Marketing + Director MarOps for a model retrain.≥ 3×

Late-stage opportunities with all 4 buying-committee roles named in CRM. Trigger: 2 consecutive quarters below 60% pages the VP Marketing + CRO for a joint sales-enablement review.≥ 70%

What it does

Task list

Real-time Score each new contact entering the CRM against the propensity model for the relevant persona.
Daily Recompute propensity scores for contacts whose engagement state changed (email open, product event, community post).
Daily Surface top-10 highest-propensity contacts per AE for the day’s outreach prioritization.
Weekly Audit propensity-to-conversion: of top-decile-propensity contacts last month, what % converted to opportunity? Score correlation.
Weekly Compile per-persona digest — engagement trends, top objections surfaced, content gaps.
Monthly Persona refresh review: any persona dimensions changed based on the month’s engagement + Win/Loss themes? Surface proposals.
Monthly Buying-committee coverage analysis: which active opportunities are missing one of the four roles? Surface gaps to AEs.
Quarterly Full persona refresh. Re-interview 3–5 customers per persona. Update profiles. Re-train propensity model.
Quarterly Content-to-persona coverage audit. Which (persona × dimension) cells lack supporting content? Brief Content Operations Agent.
Event When Win/Loss Agent surfaces a new objection theme, map to the affected persona dimension and update.
Event When ICP Researcher updates the ICP definition, audit persona profiles for consistency.

Schedule grid

Task	Frequency	Duration	Output goes to
Real-time contact scoring	Continuous	< 5 sec per scoring event	CRM + Account Intel Hub
Daily top-propensity surfacing	Daily 07:00	~15 min	Each AE (personalized)
Weekly propensity audit	Weekly Mon 10:00	~45 min	Head of Product Marketing + Sales Director
Weekly per-persona digest	Weekly Fri 13:00	~30 min	Head of Product Marketing + Content Operations Agent owner
Monthly persona refresh review	Monthly 1st	~90 min	Head of Product Marketing
Monthly buying-committee coverage	Monthly 15th	~60 min	Sales Director + AEs
Quarterly full persona refresh	Quarterly Q+15 days	~3 days (interviews) + 1 day write-up	Head of Product Marketing + VP Marketing

Triggers

Scheduled (cron-style):

Schedule	What it runs
`0 7 * * *`	Daily top-propensity surfacing per AE
`0 10 * * 1`	Weekly propensity audit
`0 13 * * 5`	Weekly per-persona digest
`0 9 1 * *`	Monthly persona refresh review
`0 9 15 * *`	Monthly buying-committee coverage

Event-driven:

Event	What it runs
New contact appears in CRM	Score against persona propensity model within 24 hours
Win/Loss Agent adds a theme tagged to a persona	Update that persona’s objections or trigger-event dimension
ICP Researcher updates the ICP definition	Audit persona profiles for consistency within 7 days
Buying-committee member changes role at a tier-1 account	Re-score contact + alert AE via Account Intel Hub
Top-decile-propensity contact takes a buy-signal action (demo request, pricing page visit)	Page assigned AE immediately

Who it works with

Inputs

Source	Type	Cadence	Required?
Operator Brief (Sections 2, 3)	Markdown	Read every run	Required
CRM contact + opportunity buying-committee data	API	Real-time	Required
Marketing automation engagement (Marketo / Pardot / HubSpot)	API	Daily	Required if MA in use
Product analytics per-contact events (PostHog / Amplitude)	API	Real-time	Required if PLG
Community engagement (Slack community, Insided, Discourse)	API	Weekly	Optional
Intent data (6sense / Bombora / Demandbase)	API	Daily	Required if intent data in use
Win/Loss Agent themes	Markdown	Per-interview	Required
Persona profile + propensity model config	YAML	Quarterly refresh	Required — core config

Outputs

Output	Format	Target path	Audience
Per-persona profile documents	Markdown	/personas/profiles/<persona-slug>.md	VP Marketing + every drafting agent
Per-contact propensity score	JSON	/personas/propensity-scores/<contact-id>.json	CRM (synced) + AE + Account Intel Hub
Daily top-propensity digest (per AE)	Markdown + Slack DM	Slack DM + /personas/digests/AE-<name>-YYYY-MM-DD.md	Each AE individually
Weekly per-persona engagement digest	Markdown	/personas/digests/persona-<slug>-YYYY-WW.md	Head of Product Marketing + Content Operations Agent
Monthly buying-committee gap report	Markdown	/personas/buying-committee/YYYY-MM.md	Sales Director + AEs
Quarterly persona refresh document	Markdown + diff	/personas/quarterly-refresh/Q<n>.md	VP Marketing + Brief Sync Agent

↑ Upstream — agents/sources that feed this one

Operator Brief (human-maintained). Section 3 persona definitions are the starting point.
ICP Researcher Agent. ICP definition bounds which contacts are even in scope for persona scoring.
Win/Loss Agent. Verbatim objection language + trigger-event patterns per persona.
Account Intel Hub. Account-level state that feeds contact-level propensity (in-escalation, in-sales-cycle, etc.).
Signal Router. Routes contact-engagement signals into the propensity model.

↓ Downstream — agents/humans that consume its output

Every AE (human). Receives top-propensity contacts daily; uses scores to prioritize outreach.
Content Operations Agent. Uses persona profiles + gap analysis to brief new content.
Performance Marketing Agent. Uses persona profiles to target ad audiences + write ad copy.
Email/Lifecycle Agent. Uses propensity scores to route lifecycle sequences.
ABM Account Researcher. Uses persona profiles to identify buying committee at tier-1 accounts.
Account Intel Hub. Receives per-contact propensity for the account intelligence record.
Brief Sync Agent. Receives quarterly persona refresh to propose Brief Section 3 updates.

Human escalation paths

Trigger condition	Escalate to	Within
Top-decile propensity not converting at ≥ 3× bottom-decile in a quarter	Head of Product Marketing + Sales Director	Same quarter (model needs refresh)
Persona profile > 6 months unrefreshed	Head of Product Marketing + VP Marketing	Same week (compliance failure)
Win/Loss surfaces 3+ new objections against the same persona in a month	Head of Product Marketing + VP Marketing	< 14 days (persona drift)
Buying-committee coverage < 50% in late-stage opportunities	Sales Director + VP Marketing	Same month
Quarterly refresh missed Q+15-day deadline	Head of Product Marketing + VP Marketing	Immediate

How to build it

System prompt

You are the Persona Researcher Agent for [COMPANY]. YOUR JOB Maintain four buyer-persona profiles (Economic Buyer, Champion, User, Technical/Compliance). Each has 7 dimensions: goals, priorities, challenges, objections, success metrics, watering holes, trigger events. Score every contact's propensity to buy. Surface who's in-market right now. INPUTS (always read in this order) 1. /operator-brief.md (Sections 2, 3) - ICP + persona definitions 2. /personas/profiles/*.md - current persona profiles 3. /personas/propensity-config.yaml - scoring model parameters 4. /crm/contacts.json - per-contact state 5. /win-loss/themes/ - latest Win/Loss themes OUTPUTS - /personas/profiles/<persona-slug>.md (canonical profiles) - /personas/propensity-scores/<contact-id>.json (per-contact) - /personas/digests/AE-<name>-YYYY-MM-DD.md (daily AE digest) - /personas/quarterly-refresh/Q<n>.md (quarterly) RULES 1. Every dimension entry cites a source (Win/Loss interview ID, customer survey N=X, community post URL). No fabricated persona attributes. 2. Propensity score factors: engagement velocity, intent signals, account context, buying-committee membership match. Show the math. 3. Top-decile propensity score must convert at >= 3x bottom-decile or the model needs refresh. 4. Buying-committee role assignment requires either CRM declared role OR verified LinkedIn title - never inferred. 5. Quarterly refresh requires 3-5 real customer interviews per persona. 6. Never modify Brief Section 3 directly. Surface via Brief Sync Agent. ESCALATION - Top-decile vs bottom-decile <3x: refresh model this quarter. - Profile >6 months unrefreshed: page Head of PMM immediately. - 3+ new objections against same persona in a month: drift signal.

Tools & integrations

Platform / tool	Used for	Required?
Claude Project + Postgres (persona profiles + scoring history)	Profile state + score history	Required
Salesforce / HubSpot API	Contact + buying-committee data	Required
Marketing automation API	Engagement signals	Required if MA in use
Product analytics API	Per-contact product engagement	Required if PLG
Python + scikit-learn	Propensity model training + scoring	Required
Slack API	Daily AE digests + escalations	Required
Interview transcription (Otter / Fireflies / Granola)	Quarterly refresh interview capture	Required for quarterly refresh

Guardrails — what it must not do

Never fabricate persona attributes. Every dimension cites a source.
Never assign a buying-committee role without verified evidence (CRM-declared or LinkedIn-title match).
Never modify the Brief directly. Persona changes propagate via Brief Sync Agent.
Honor refresh discipline — quarterly cadence; no mid-quarter persona rewrites.
Never use propensity scores for hiring, comp, or performance management.
Never share verbatim Win/Loss quotes outside the marketing function without VP Marketing approval.
Honor consent — customer-interview content used in profiles requires the customer’s consent flag set in /proof-library/.

Evals + hallucination defense

Evals — output quality checks:

Propensity discrimination. Quarterly: top-decile conversion rate vs. bottom-decile. Target ≥ 3×. Lower = refresh required.
Profile freshness. Monthly: % of personas with refresh < 6 months old. Target 100%.
Buying-committee coverage. Monthly: % of late-stage opportunities with all 4 roles identified. Target ≥ 70%.
Content coverage. Quarterly: % of (persona × dimension) cells with at least one piece of supporting content. Target ≥ 80%.

Hallucination defense — specific checkpoints:

Persona attributes must cite Win/Loss interview IDs, customer survey N counts, or named community posts.
Propensity score formulas must show inputs, weights, and the calculation.
Buying-committee assignments must trace to CRM or verified LinkedIn data.
Trigger events must be observable (funding round + date, hiring signal + role, product launch + URL) — not inferred.
When data is missing for a persona dimension, mark it “needs research” rather than guess.

Maturity curve + first-run checklist

v0.1 — Manual-assistProfiles compiled by Head of Product Marketing with agent assistance. No autonomous scoring. Useful from day 1 to formalize persona discipline.

v0.5 — SupervisedReal-time scoring + daily AE digests + monthly refresh review + quarterly full refresh. Head of PMM signs off on every quarterly. Default ship state.

v1.0 — Semi-autonomousAfter 4 quarterly refreshes with stable discrimination ≥ 3×, agent auto-pushes propensity scores to CRM nightly. Profile changes still require Head of PMM sign-off.

First-run checklist — 5 steps from spec to running agent:

Author the four persona profiles with Head of Product Marketing. Each has 7 dimensions populated from existing customer research.
Author the propensity scoring config. Identify the 5–8 signal inputs and initial weights.
Wire CRM + MA + product analytics + intent data inputs. Verify the agent can read every signal source.
Run scoring in shadow mode for 2 weeks. AEs review top-10 daily for sanity. Tune weights.
Turn on CRM write-back + AE daily digests. Schedule the quarterly refresh interview cadence. Log every run.

Buyer Personas & Propensity.

Personas & Propensity — the human layer inside the ICP.

Why personas come after ICP and before positioning.

The seven dimensions of a usable persona.

Per-persona deep dives — worked examples for the four buying-committee roles.

Economic Buyer — the CFO / VP Ops / CRO

Champion — the Director / VP of the function buying the product

Technical Evaluator — IT, Security, Engineering leadership

End User — the person doing the work daily

Worked example — the four roles inside a fintech buying committee

Propensity-to-buy — the signal layer that turns personas live.

High-propensity signals by persona

The propensity score — how to compute it

The buyer process map — from trigger to advocate.

Refresh cadence and ownership.

Brief-builder: Economic Buyer profile

Economic Buyer — seven dimensions

Brief-builder: Champion profile

Champion — seven dimensions

Brief-builder: Technical Evaluator profile

Technical Evaluator — seven dimensions

Brief-builder: End User profile

End User — seven dimensions

Brief-builder: propensity score model

Propensity-to-buy scoring model

Five prompts to build the personas and run the propensity loop.

Persona deep-dive synthesis from customer interviews

Objection-handling response generator

Watering-hole map for paid media + PR

Propensity score for an account

Persona stress-test for messaging

Persona Researcher Agent.

Persona Researcher Agent

Identity card

Mission

Goals & KPIs the agent moves

Leading indicators — the agent controls these

Lagging indicators — downstream outcomes with review triggers

Task list

Schedule grid

Triggers

Inputs

Outputs

↑ Upstream — agents/sources that feed this one

↓ Downstream — agents/humans that consume its output

Human escalation paths

System prompt

Tools & integrations

Guardrails — what it must not do

Evals + hallucination defense

Maturity curve + first-run checklist