The persona industry has a precise origin. Adele Revella's Buyer Persona Institute formalized the framework in 2010. HubSpot wrote it into the inbound marketing playbook the same year. Marketo and Eloqua — both later absorbed by Adobe and Oracle respectively — pushed personas into marketing automation as a configuration step. By 2014, every B2B agency in North America was selling "persona workshops" as a six-figure deliverable.
The problem was structural. The original Revella framework required real buyer interviews — five to ten conversations per persona, transcribed, coded for objections and triggers. Almost no one did the interviews. Agencies skipped the interviews because interviews are expensive. Internal teams skipped the interviews because the marketing team had no access to closed-won customers. So personas became internal opinion — what the VP of Marketing thought the buyer wanted, dressed up in a stock photo and a fake name.
The output looked like research. It was projection.
A 2022 Forrester analysis of B2B persona deployment found that fewer than one in five marketing teams had ever validated a persona against a closed deal. The personas sat in onboarding decks. They never touched a campaign decision. They never updated when the buying group changed. They never reflected that the persona's actual job title had shifted because of a reorg, an acquisition, or a software category collapse.
The agencies kept selling them anyway. The personas got more elaborate — psychographics, jobs-to-be-done overlays, day-in-the-life narratives. None of it was wrong, exactly. It just wasn't accountable to any signal coming back from the market. A persona could be wildly inaccurate and the marketer would never know, because the persona never had to predict anything.
This is the part that needs to be said plainly. Personas were not customer research. Personas were a deliverable invented to make the marketing function look more analytical than it actually was. The industry built a profession around imaginary people.
Salesforce, ServiceNow, and Adobe — three of the largest B2B software companies in the world by revenue — do not run on personas. They run on account intelligence. The persona shows up in their training decks for new sales hires. It does not show up in pipeline analysis, in expansion forecasting, or in segmentation logic. The persona is a teaching aid. The committee is the customer.
That gap — between what the persona industry promised and what the largest revenue operations in B2B actually use — is the gap this piece is about.
2. Why modern buying committees broke the persona model
Every persona framework assumes one buyer. One job title. One pain point. One decision.
No B2B purchase over $25,000 works that way.
Consider a hospital information system replacement. The deal is $4M to $40M, depending on bed count. The participants:
- The Chief Information Officer, who owns the technical evaluation.
- The Chief Financial Officer, who owns the capital expenditure and the depreciation schedule.
- The Compliance Officer, who owns HIPAA, HITRUST, and state-specific health information exchange rules.
- A physician champion — usually a department chair — who owns clinical workflow acceptance.
- Procurement, which owns the RFP process and the contract terms.
- Often a Chief Medical Information Officer, sitting between the physician and the CIO.
- Sometimes a board IT committee, for purchases above a board-approval threshold.
Seven roles. Different priorities. Different success criteria. Different fears. The CIO is worried about integration. The CFO is worried about total cost of ownership over seven years. The compliance officer is worried about a data breach during migration. The physician champion is worried about being blamed by colleagues when the new system is slow on day one. "Healthcare Hannah" cannot represent any of them.
Commercial real estate software is the same picture in a different industry. A property management platform sold to a REIT involves the owner or asset manager, the operations director, the head of finance, the IT lead, and increasingly a sustainability or ESG officer because tenant reporting now requires energy data. The buying motion is two to four quarters. The contract is multi-year. The persona is irrelevant.
Regional bank technology purchases — core banking, fraud, AML, customer communication — pull in the Chief Operating Officer, the Chief Technology Officer, the Chief Risk Officer, the Compliance Officer, often the General Counsel, and the Chief Marketing Officer when the system touches customer experience. Six roles minimum. Each one can kill the deal.
Gartner published the foundational research on this in 2017 and has updated it through 2024. The headline finding has not moved: the typical buying group for a complex B2B solution involves six to ten decision makers, each armed with four to five pieces of independently gathered information. That is 30 to 50 discrete information sources per deal — research the seller never sees and never controls.
Forrester ran a parallel study in 2023 across 1,200 B2B deals. The finding: 63% of B2B buying processes involve four or more people. The number of stakeholders per deal has grown roughly 25% in the last decade. Buying groups are getting larger, not smaller, as AI tools make it easier for non-specialists to participate in technical evaluations.
McKinsey adds the velocity layer. A 2024 B2B Pulse survey found that B2B buyers now use ten or more channels in a single purchase cycle — up from five in 2016. The buyer is not following the persona's predicted journey. The buyer is following whatever signal arrives first, including signals from AI engines that did not exist when the persona was written.
The persona model assumed a buyer marketing could profile, address, and convert. The actual buyer is a committee that the marketer cannot see, cannot map in real time, and cannot reduce to a single fictional sketch. You aren't selling to one persona. You are selling to a system.
And the system rewires every quarter — when the CIO leaves, when finance gets a new VP, when compliance brings in a consultant. The persona document does not update. The market does.
3. How Salesforce, ServiceNow, and Adobe actually sell
Three companies. Combined revenue north of $80 billion. None of them runs go-to-market off a persona.
Salesforce calls its model account-based everything. The unit of analysis is not the buyer — it is the account. Salesforce's own playbook, codified in the Customer 360 and Account-Based Marketing motions, treats every named target account as a multi-stakeholder graph. Marketing builds account-level engagement scores. Sales builds named-stakeholder maps inside each account. Customer success expands across business units inside the same logo.
Inside that model, persona language is shorthand. "We need to engage the CFO economic buyer" is a useful sentence in a sales call. "Our persona is CFO Chris" is not — because the CFO at a $200M manufacturer behaves nothing like the CFO at a $20B retailer, and the Salesforce rep covering both knows it. The persona is too crude. The named account is the unit.
Salesforce reports more than 150,000 customer accounts. Internal segmentation is built on firmographic data, install-base signal (what other Salesforce clouds the account already owns), industry vertical, and account intent. Persona is a downstream output, not an input.
ServiceNow is more aggressive about it. ServiceNow's expansion motion — the engine that pushed the company to $11B+ in annual revenue — depends on landing inside IT and expanding into HR, customer service, security operations, and finance. That motion is impossible to execute against a persona, because the persona was about the IT buyer, and the expansion deal is about an HR buyer who has never heard of ServiceNow.
ServiceNow's commercial architecture solves for buying committees explicitly. The platform sales motion includes a "workflow champion" map for each customer — who in IT is the technical sponsor, who in the business unit is the workflow owner, who in finance signs off on the platform license expansion, who at the executive level is the political cover. Five named humans, per account, per expansion play. Persona-free.
Adobe is the third pattern. Adobe's enterprise business — Experience Cloud, Document Cloud, Creative Cloud at scale — sells across roles that have almost nothing in common. The Chief Marketing Officer buys Experience Cloud. The General Counsel buys Document Cloud. The Chief Creative Officer or VP of Brand buys Creative Cloud at the enterprise tier. Three buyers. Same logo. Three procurement processes.
Adobe's segmentation is role-based and industry-vertical, not persona-based. The role is real (CMO, GC, CCO). The industry vertical is real (financial services, retail, life sciences). The intersection of role and vertical is the addressable segment. There is no Marketing Mary stand-in. There is a CMO in retail, who behaves differently from a CMO in financial services, who behaves differently from a CMO in life sciences. Three segments. Different content. Different sales motions. Different reference customers.
HubSpot, the company that arguably commercialized the persona framework, has quietly moved the same direction. The HubSpot revenue org now runs on "ideal customer profile" plus account intent signal from Breeze and from third-party intent providers. The persona still appears in HubSpot Academy training because HubSpot sells education. It does not drive HubSpot's own pipeline allocation.
The pattern across all four — Salesforce, ServiceNow, Adobe, HubSpot — is that the operating unit is the account, the role, and the signal. The persona is a teaching device. It survives in onboarding decks because it is easy to explain. It does not survive in revenue operations because it cannot predict pipeline.
This matters for mid-market companies because the mid-market is being sold a persona framework that the largest B2B revenue organizations do not use. Agencies still pitch persona workshops at $40,000 to $120,000. The output is a deck. The deck does not connect to a CRM record, an intent platform, an ABM target list, or an AI visibility audit. The deck sits on SharePoint. The pipeline does not move.
The mid-market should be borrowing the operating model the enterprise uses — account intelligence, role-based segmentation, signal-driven engagement, AI visibility as a buyer-research input — and skipping the persona ceremony. The persona ceremony was a workaround for not having data. Mid-market companies in 2026 have data.
They are choosing not to use it.
4. The rise of customer intelligence
The replacement category has a name: customer intelligence. It is not a single tool. It is a stack.
Customer intelligence is the discipline of building a real-time, signal-rich view of accounts and buying committees, using behavioral data, search and content consumption, CRM history, third-party firmographic and intent data, and — newly in 2025 and 2026 — AI engine discovery patterns. Where personas were static, customer intelligence is continuous.
The category is anchored by three companies.
ZoomInfo built the contact and firmographic layer. The ZoomInfo database covers more than 100 million business contacts and 100 million companies, with org charts, technographics, funding signal, and hiring intent. The product replaces the part of the persona that was a guess about job title and reporting structure with a verified record updated daily.
6sense built the intent layer. 6sense ingests anonymous buyer behavior across the web — content consumed, keywords searched, vendor sites visited, review sites compared — and de-anonymizes it back to a named account using IP and identity graph. The output is a heat map: which of your target accounts are actively researching the problem you solve, right now, before they fill out a form. That is information no persona could give you because the persona was static.
Demandbase built the engagement and orchestration layer. The Demandbase platform layers account scoring, advertising, and sales orchestration on top of intent data, so marketing and sales work the same prioritized list. The category was originally called account-based marketing. It has matured into something more useful: account intelligence as a continuous operating function.
Around these three are the second tier — Bombora, G2 (for review-stage intent), Clearbit (now inside HubSpot Breeze), Apollo, LinkedIn Sales Navigator, and the CRM-native intent signals that Salesforce and HubSpot are building in. The point is not the vendor list. The point is the architecture.
Customer intelligence has five inputs:
- Behavioral signals — what accounts do on your owned properties, on review sites, on partner sites.
- Search and discovery intent — what queries are being run, on Google, on G2, and now inside AI engines.
- Content consumption — which assets are being read, by whom, in what sequence.
- CRM data — closed-won, closed-lost, expansion, churn, deal velocity, deal complexity.
- AI engine discovery — what ChatGPT, Claude, Gemini, Perplexity, and Google AI Overviews are recommending when a buyer asks the category question.
The last one is the new layer. It is the one most mid-market companies are missing. AI engine discovery is now where 30 to 40% of category research begins. If the buyer asks Claude or ChatGPT "what are the best property management platforms for a mid-sized REIT" and you are not in the answer, you are not in the consideration set. The persona did not anticipate the AI engine because the persona was written before the AI engine existed.
This is also why Generative Engine Optimization — GEO — is now part of customer intelligence rather than a separate SEO sub-discipline. Whether you appear in the AI engine's answer is a behavioral signal about the market, not just a tactic to drive traffic. If you are not cited, the market is researching you out of existence. Citation Share is becoming a leading indicator for pipeline, the same way share of voice was for the previous decade.
Customer intelligence is the operating system. Personas are a UI layer at best.
5. Healthcare: what the buying committee actually looks like
A mid-market healthcare software company — say, a clinical workflow platform selling to community hospitals and regional health systems — does not have a single customer. It has six.
The Hospital CEO or Health System COO is the executive sponsor. The CEO is rarely the technical evaluator. The CEO cares about strategic alignment — does this platform help the system compete with the academic medical center forty miles away, and does it move the needle on patient outcomes the board reports on every quarter.
The Practice Administrator or VP of Operations is the workflow buyer. This is the person whose week gets harder if the implementation fails. The administrator is measured on throughput, on staff retention, on patient satisfaction scores. They want to know what the platform does to the nursing workflow, to the front-desk workflow, to the billing handoff.
The Chief Medical Information Officer or physician champion is the clinical translator. This is the only role on the buying committee that has clinical authority. If the CMIO says the system will increase clicks-per-patient, the deal slows. If the CMIO says the system will reduce after-hours charting time, the deal accelerates. The CMIO's opinion is leveraged through the medical staff committee, which the seller never directly meets.
The IT decision-maker — the CIO or VP of Information Systems — owns the integration question. Will the platform integrate with the existing electronic health record (Epic, Cerner-now-Oracle Health, Meditech, Athenahealth)? What is the FHIR API maturity? How does HL7 interface management work? How does single sign-on resolve against the Active Directory the system already runs?
The CFO or VP of Finance owns the economics. The capital expenditure question, the operating expense classification, the depreciation schedule, the return-on-investment model the board will see. The CFO will ask for a five-year total cost of ownership comparison against staying on the current system. If the comparison is not ready, the deal is not ready.
Compliance is the sixth seat. HIPAA, HITRUST, SOC 2, state breach notification rules, business associate agreements, data residency. The compliance officer can stop a deal at any stage with a single "we have not completed the security review."
Six roles. None of them is Healthcare Hannah.
The mid-market healthcare software company that wins is the one that produces six distinct evidence packs — one for each role — and sequences them through a 9-to-15-month sales cycle. The CEO gets the strategic outcomes deck and the peer-system reference list. The administrator gets the workflow ROI calculator and the implementation case study. The CMIO gets the clinical evidence pack and a peer-physician reference call. The CIO gets the integration architecture diagram and the security questionnaire pre-completed. The CFO gets the TCO model and the financing options. The compliance officer gets the HITRUST certification and the BAA template.
That is six pieces of content, six sales motions, six reference customer types, and six different intent signals to monitor. None of it is persona-shaped. All of it is account-intelligence-shaped.
The AI engine layer sits on top: when a hospital CIO asks Claude "what clinical workflow platforms integrate cleanly with Epic for a 300-bed community hospital," the mid-market vendor either appears in the answer or does not. If they do not, the CIO never enters the funnel. The persona was not built to solve for that. Customer intelligence is.
6. Financial services: regional banks, wealth managers, fintech vendors
A fintech vendor selling to a $5B-to-$25B asset regional bank is selling into one of the most consensus-bound buying environments in B2B.
The Chief Operating Officer is usually the executive sponsor. Regional bank COOs sit on the management committee, often report to the CEO, and own the operating model — branch operations, back office, technology, vendor management. The COO can clear an internal logjam. The COO almost never signs the contract alone.
The Chief Technology Officer or Head of Information Systems owns the technical fit. Regional banks run on core banking platforms — Fiserv, FIS, Jack Henry, Finastra — that dictate which adjacent vendors integrate cleanly. The CTO's first question is whether the new vendor speaks to the existing core. If not, the deal architecture changes from "product purchase" to "core integration project," which is a different conversation with a different timeline and budget.
The Chief Risk Officer and the Chief Compliance Officer often share the third and fourth seats. The CRO owns operational risk, model risk, third-party risk. The CCO owns BSA, AML, KYC, fair lending, consumer protection. A new vendor triggers a third-party risk review, a model validation if any algorithm is involved, and a regulatory notification path that depends on the bank's primary regulator (OCC, Federal Reserve, FDIC, or state).
The Chief Financial Officer owns the budget and the regulatory capital implication, which is non-trivial when the vendor touches the balance sheet. The CFO will also pull in the Chief Accounting Officer if the vendor changes how revenue or expense gets booked under GAAP or under regulatory call report standards.
Increasingly the Chief Information Security Officer is the gating role. Regional banks have spent the last decade getting hammered on cybersecurity by examiners. The CISO can independently kill a deal on security posture, and most are now empowered to do so by the board's risk committee.
Six to seven roles. The buying motion is two to four quarters at minimum and can stretch to six quarters for a core-adjacent system.
Wealth managers and registered investment advisors behave differently. The buying committee at a $2B AUM RIA is smaller — the managing partner, the chief operating officer, the head of advisor technology, and often a single advisor representative — but the regulatory layer (SEC, FINRA, state) and the custodian dependency (Schwab, Fidelity, Pershing) shape the decision more than persona would suggest. The RIA does not buy on individual taste. The RIA buys on what the custodian integrates and what the compliance officer will approve.
Fintech vendors selling into either segment cannot run on a single persona. They run on account intelligence — which banks just changed core providers, which RIAs just crossed a regulatory threshold, which institutions are advertising for a new Head of Risk or a new Head of Advisor Technology. Those signals show up in ZoomInfo, in 6sense intent feeds, in LinkedIn hiring data, and increasingly in AI engine answers when a bank executive asks Claude or ChatGPT "what AML vendors integrate with Jack Henry for a $10B regional bank."
The fintech vendors that win are not the ones with the most elaborate persona deck. They are the ones with the most accurate account map, the cleanest integration story for the core platforms, and the strongest presence inside the answer when the bank executive asks the question.
7. Real estate: PropTech, brokerages, owners, REIT service providers
Commercial real estate is the industry where the persona model fails most visibly, because the buyer is rarely an individual at all — the buyer is an ownership structure, a property type, and an operating partner.
A PropTech vendor selling a tenant experience platform to a commercial office owner sells into a five-to-seven-person committee. The owner — sometimes a single principal, sometimes an investment committee at a private equity real estate fund — sets the strategy. The asset manager runs the property-by-property economics. The operations director runs the on-the-ground experience. The leasing team owns tenant acquisition. The finance team owns the capital stack. If the building is institutionally owned, there is also a third-party property manager — JLL, CBRE, Cushman & Wakefield, Newmark — sitting between the owner and the platform.
That property manager is often the gatekeeper. JLL, CBRE, and Cushman manage hundreds of millions of square feet between them. When they approve a vendor for their managed portfolio, the vendor gets distribution. When they do not, the vendor gets a single-building pilot that never scales. The PropTech persona deck rarely mentions the property manager, because the persona deck was built around the owner. The owner is not the only buyer.
Multifamily is a different geometry. A property management software vendor selling to a multifamily owner-operator — Greystar, Mid-America, Camden, AvalonBay — sells into a buying group that includes the head of operations, the head of technology, the regional vice presidents, the head of marketing (because the platform usually touches the leasing funnel), and the head of asset management. Five-plus roles, and the deal often runs through a single-portfolio pilot before going national.
REIT service providers — auditors, ESG reporters, capital markets advisors, tax specialists — sell to a REIT's CFO, controller, head of investor relations, head of ESG, and increasingly the general counsel because of the SEC's climate disclosure rules. The persona, again, is irrelevant. The CFO at a $5B industrial REIT and the CFO at a $15B office REIT have nothing in common except the title.
Brokerages are the third pattern. A technology vendor selling to a residential brokerage — Compass, Anywhere, eXp, Keller Williams — is selling to the corporate technology team, the agent productivity team, the broker-owner network if the model is franchise-driven, and the agent base itself, which votes with adoption. The agents are the actual users. The corporate team writes the check. If the agents do not adopt, the contract does not renew. The buying committee includes the executive decision-maker and the population of users who decide whether the deal survives year two.
None of this is persona-shaped. All of it is account- and segment-shaped. The PropTech and brokerage vendors that win in 2026 are the ones that map ownership structures, property managers, asset classes, and agent population characteristics — and the ones that show up in the AI engine answer when a head of asset management asks Claude "what tenant experience platforms integrate with Yardi for a Class A office portfolio."
The persona did not anticipate Yardi. The persona did not anticipate the agent population vote. The persona did not anticipate the property manager gatekeeper. Customer intelligence does.
8. AI is changing customer discovery
The persona was written for a world where the buyer started research on Google. That world ended in 2024.
In 2026, a measurable share of B2B category research starts inside an AI engine. ChatGPT alone reports 700 million weekly active users. Google AI Overviews now appear above the classic ten blue links for the majority of commercial queries. Perplexity reports billions of queries per month. Claude is in active enterprise deployment across financial services, healthcare, and professional services. Gemini is the default AI surface inside Google Workspace, which is to say inside the daily workflow of most enterprise knowledge workers.
When a hospital CIO, a regional bank COO, or a multifamily head of operations begins category research today, the first move is often a prompt. "What are the leading clinical workflow platforms for a community hospital." "What AML vendors integrate with Jack Henry." "What tenant experience platforms work with Yardi for Class A office." The AI engine returns an answer — a curated list of vendors, often three to seven names — and that list is the consideration set.
If you are not in the answer, you are not in the consideration set. The buyer does not see a button that says "see more vendors." The buyer sees the list, picks two or three names to research further, and moves into the deeper funnel. The funnel you used to optimize — the website, the gated content, the SDR sequence — is downstream of an answer you did not write.
This is what changed. The persona assumed you controlled the discovery surface, because the discovery surface was Google search results, and Google search results responded to SEO. The discovery surface is now an answer generated by a model, and the model responds to citation, structured data, third-party authority, and presence in the corpora the model was trained on or retrieves from.
The discipline that addresses this is Generative Engine Optimization — GEO. The term was coined in 2023 in academic work out of Princeton. The practice has matured into a measurable category. AI Communications firms now run AI Visibility audits that score a brand across all five major engines, measure Citation Share against named competitors, and identify the structural gaps that keep a brand out of the answer.
The new questions are operational, not theoretical:
- What does ChatGPT recommend when a buyer asks the category question?
- What companies appear in Gemini's answer for the same query, and how do they differ?
- Which brands are cited in Claude's answer when the buyer asks for evaluation criteria?
- How does Perplexity, which cites sources visibly, weight the trade press, the analyst firms, and the owned content?
- What appears in Google AI Overviews, and how does that interact with the classic search ranking the brand already optimized for?
Those answers form the AI Discovery layer of customer intelligence. They are not a marketing afterthought. They are a buyer behavior reality. A mid-market company that does not measure them is operating blind on the surface where its buyers now start.
The persona did not measure this. The persona could not measure this. AI engines did not exist in their consumer form when the persona framework was built. The replacement framework — customer intelligence with AI Discovery as a layer — does measure it.
This is the operational shift the mid-market is most behind on. The Fortune 500 has CMO-level mandates for AI visibility in 2026. The mid-market still has a persona deck on SharePoint. The gap is widening every quarter, because every quarter the AI engines are training on more recent data, surfacing more brands by name, and shaping more category research before the buyer ever lands on a website.
This is the structural shift. AI Communications — the discipline of becoming the answer inside ChatGPT, Claude, Gemini, Perplexity, and Google AI Overviews — is now part of customer intelligence, not adjacent to it.
9. The mid-market customer intelligence framework
Everything-PR proposes the following operating framework for mid-market B2B companies — companies between $25M and $500M in revenue — that need to replace the persona model with something accountable to pipeline.
Five layers. Each one is a deliverable, an owner, and a measurement cadence.
LAYER 1 — THE BUYER. The named role with sign-off authority on the contract. Not a persona. A title at a real account, validated against closed-won deals from the last 24 months. The deliverable is a buyer map per segment: title, reporting line, typical evaluation criteria, typical objections, typical procurement path. The owner is sales operations. The cadence is quarterly, updated against the most recent closed deals.
LAYER 2 — THE INFLUENCER. The role that does not sign the contract but can kill the deal. In healthcare, this is the physician champion. In financial services, it is the Chief Risk Officer. In real estate, it is the property manager or the agent population. The deliverable is an influencer map by segment, with the evidence pack that the influencer needs to advocate internally. The owner is product marketing. The cadence is semi-annual, updated against won and lost deal post-mortems.
LAYER 3 — THE RECOMMENDER. The third-party voice the buying committee trusts. Analyst firms (Gartner, Forrester, IDC). Trade press (Modern Healthcare, American Banker, GlobeSt). Peer references at named accounts. Increasingly, AI engines — which are now recommenders in their own right, citing some sources, ignoring others. The deliverable is a recommender map: which third parties carry weight in this segment, and what is our presence with each. The owner is communications. The cadence is monthly for press and analyst, weekly for AI engines.
LAYER 4 — THE BUDGET HOLDER. The economic buyer, which is often not the user buyer. In hospital software, the CFO. In bank technology, the COO or CFO. In PropTech, the asset manager or the principal. The deliverable is a budget-holder evidence pack: total cost of ownership models, return on investment proof points, financial reference customers in the same segment and revenue band. The owner is sales engineering and finance, jointly. The cadence is per-deal, with a quarterly refresh of the underlying TCO model.
LAYER 5 — THE AI DISCOVERY LAYER. The answer the AI engine gives when the buyer asks the category question. This is the layer the persona framework cannot address. The deliverable is an AI Visibility audit: Citation Share across ChatGPT, Claude, Gemini, Perplexity, and Google AI Overviews, scored against named competitors, refreshed quarterly at minimum. The owner is AI Communications, working with marketing operations. The cadence is monthly for tracking, quarterly for full audit.
Together, these five layers produce a customer intelligence system. The persona becomes a teaching artifact for new hires — "this is the kind of person we sell to" — while the operating layer is account-named, signal-driven, role-mapped, and AI-visible.
The output is not a single deck. The output is:
- A named account list per segment, refreshed quarterly with intent signal.
- A role-by-role evidence library, owned by product marketing, refreshed semi-annually.
- A recommender presence dashboard — press, analyst, peer reference, AI engine — owned by communications, refreshed monthly.
- A budget-holder financial model library, owned by sales engineering and finance, refreshed per deal.
- An AI Visibility scorecard, owned by AI Communications, refreshed monthly with full audit quarterly.
This is what mid-market customer intelligence looks like in 2026. It is not exotic. It is not enterprise-only. The tools to execute it — ZoomInfo for the contact layer, 6sense or Demandbase for the intent layer, an AI Visibility audit for the discovery layer, a CRM with closed-won and closed-lost discipline for the historical layer — are within the operating budget of any company doing more than $25M in revenue and selling B2B.
The companies that move to this model in 2026 will compound an operating advantage every quarter, because every quarter the underlying intelligence gets sharper. The companies that stay on the persona model will keep producing decks that look like research and do not predict pipeline.
The choice is not whether to retire the persona. The persona will retire itself, when the marketing team realizes the document has not changed in two years and the buyers have. The choice is what replaces it — a continuous, signal-driven, AI-visible operating model, or another version of the same fictional sketch with a fresh stock photo.
Mid-market companies that pick the operating model win the decade. The ones that pick the stock photo will be looking up at the ones that did not.
10. Buyer personas aren't dead — they're just not enough
The honest version of the argument is this. Personas are not useless. They are insufficient.
A persona is still a fine teaching tool when a new account executive joins the sales team and needs a one-page sketch of the customer she will be calling next week. A persona is still a fine onboarding asset for a content marketer trying to understand the rough shape of the audience he is writing for. The persona was never wrong because it was unhelpful. The persona was wrong because the industry sold it as customer research and it was not.
Real customer research is interviews — five to ten conversations per role per segment, transcribed, coded. Almost no one does this. The companies that do it produce personas that age into rough usefulness over twelve to eighteen months. The companies that skip it produce fiction.
Even the rigorous, interview-validated persona is insufficient on its own. The persona describes a person. The deal involves six to ten people. The persona describes a profile. The buyer's behavior is a signal that arrives in real time. The persona describes what the buyer thinks. The AI engine recommends what the buyer hears, before the buyer has formed an independent opinion.
The winning companies in 2026 combine the persona with everything else. They keep the persona as a training artifact. They build account intelligence on top of it. They track search and content-consumption signals through 6sense or Demandbase or HubSpot Breeze. They run their CRM with the closed-won/closed-lost discipline that makes historical data predictive. They monitor AI engine citation as a continuous input, not as a vanity audit. They sell to the committee, not the persona.
The market intelligence layer — analyst coverage, press presence, peer references, AI Visibility — is the connective tissue that lets all of this compound. A company that is cited by Gartner, covered by trade press, recommended by named customers, and named inside Claude and ChatGPT when the buyer asks the category question is operating with leverage the persona-only company cannot match.
This is the AI Communications era. Buyer behavior moved. Buying committees expanded. Discovery shifted into AI engines. The persona stayed where it was.
Mid-market companies do not have to abandon the persona. They have to admit it was never the whole picture, and rebuild the rest of the picture with the data, the signals, and the AI visibility the persona could not deliver.
Buyer personas are not dead. They are a single layer in a five-layer system. Treat them that way, and the rest of the system can finally do what the persona was always being asked to do alone — predict the buyer, map the committee, win the deal.
No. They are a teaching artifact, not a customer-research output. A persona is useful for onboarding a new sales rep or content marketer. It is insufficient as the operating model for B2B pipeline, because the actual buyer is a six-to-ten-person committee, not a single sketch.
What is replacing buyer personas?
Customer intelligence — a stack combining account-level firmographic data (ZoomInfo), intent signal (6sense, Demandbase), CRM-driven closed-deal analysis, and AI engine discovery monitoring across ChatGPT, Claude, Gemini, Perplexity, and Google AI Overviews. It is continuous, signal-driven, and accountable to pipeline.
What is AI Discovery in customer intelligence?
AI Discovery is the layer that monitors what AI engines recommend when buyers ask category questions. If your brand is not cited inside ChatGPT, Claude, Gemini, Perplexity, or Google AI Overviews when a buyer asks the question, you are absent from the consideration set before the funnel begins.
How big is a typical B2B buying committee?
Gartner reports six to ten decision-makers for complex B2B solutions. Forrester finds 63% of B2B buying processes involve four or more people. Numbers have grown roughly 25% over the last decade. No persona can represent that committee. Account intelligence must.
What is the five-layer mid-market customer intelligence framework?
Buyer, Influencer, Recommender, Budget Holder, and AI Discovery Layer. Each layer has a named owner, a deliverable, and a measurement cadence. The framework replaces the static persona deck with a continuous, signal-driven operating model accountable to pipeline.
Do enterprise B2B companies still use personas?
Salesforce, ServiceNow, Adobe, and HubSpot use personas as teaching aids in new-hire training. None of them runs pipeline allocation, segmentation, or expansion forecasting on personas. The operating units are accounts, roles, intent signals, and AI engine presence — not fictional buyers.
What is GEO and why does it matter for customer intelligence?
Generative Engine Optimization (GEO) is the discipline of becoming the answer inside AI engines. The practice originated in 2023 in academic work out of Princeton. GEO matters because AI engines now shape a measurable share of B2B category research, ahead of the website and ahead of the SDR.
