For much of the past decade, customer advocacy programs followed a familiar model. Marketing wrote a case study, legal took three months to approve it, and it went up on the website with a logo next to it. When a deal got serious, sales asked for a reference call, the advocacy manager rang the same handful of customers who always said yes, and one of them spent half an hour on the phone with the prospect. If you were good at the job, you kept that list warm, ran an awards program and got a few customers on stage once a year.
Buyers increasingly encounter customer evidence through AI-generated summaries before visiting your website or speaking with sales. They ask ChatGPT or Perplexity what your customers think of you, and the AI draws on reviews, forum threads, LinkedIn posts and other sources to answer. If your evidence is limited to a generic case study and an occasional late-stage reference call, buyers — and the AI systems assisting them — have too little specific, relevant proof to work with.
The reference call still happens. It increasingly comes late in the decision process, and it asks time of the same customers again and again. What influences a deal in 2027 is credible, specific customer evidence that buyers can discover and verify before they speak with sales. That’s a different job from the one most advocacy managers were hired for, and many of them are doing it alone.
Table of Contents
1. Buyers ask ChatGPT what your customers think of you
TrustRadius asked 1,862 technology buyers in January 2026 what they use when they buy. 74% said reviews. Only 13% said analyst reports, which is down 63% since 2022. Your own marketing material came last on the list. 63% had used AI somewhere in the purchase, and 94% of those check what the AI tells them at least some of the time, mostly against reviews and against people they know.
And the AI itself gets its answers from somewhere else. Muck Rack went through 25 million links that ChatGPT, Claude and Gemini cited in their answers. 84% came from earned media, including journalism, reviews, forums and independent posts; journalism alone accounted for 27%. Paid and advertorial content accounted for 0.3%. Your website can still appear, but your case study page competes with a much wider range of sources.
Buyers have long given independent customer evidence more weight than vendor claims. AI now shapes which sources many of them encounter first. In 2027 the advocacy job is making sure that, when buyers and their AI assistants go looking, they find something true, recent and specific, with a customer willing to stand behind it.
2. The problem with reference calls
UserEvidence asked 811 buyers, sellers and marketers about customer proof for its 2025 Evidence Gap report. Buyers put the one-to-one customer reference second to last for trust, at 42%. A testimonial with a customer’s name on it was trusted by 64%, and an anonymous statement verified by a third party by 60%. The comparison challenges the assumption that a resource-intensive reference call is necessarily more persuasive than a short, independently verified customer statement. That does not make a live reference unhelpful in every deal. It means the format alone is no guarantee of trust. And 53% of sellers said a deal had slowed down or fallen through in the past year because they lacked proof to show.
Repeated reference requests can also become an unreasonable burden on the customer. In the programs we run, a reference customer will take two, maybe four calls a quarter before they start finding reasons to say no. The same handful get asked every time because they’re the ones who said yes last time. The people who like you most get worn out first, and whatever they said on the call sits in one prospect’s notes and nowhere else.
Keep the reference call for the deals that genuinely need it, often late, when a buyer wants to test an important assumption with someone like them. Earlier in the journey, evidence should be discoverable and reusable. A customer answering a question in a public community or on an independent review site can inform many more buyers than one private conversation — provided the answer can be found and its context checked.
3. Buyers make a short list before they call you
6sense surveyed close to 4,000 B2B buyers for its 2025 Buyer Experience Report. 94% of buying groups had their shortlist in order of preference before they spoke to a single seller, and the vendor they put first won 77% of the time. By the time sales asks you for a reference, the buyer may already have a preferred vendor based on information found weeks earlier. A reference call can confirm that preference — or call it into question.
Where did they find those things? SE Ranking looked at 30,000 searches with buying intent and found a review platform cited in 34.5% of Google’s AI Overviews. At the same time, organic traffic to G2 fell 84.5% between January 2024 and December 2025, and TrustRadius lost 92%. That measures search traffic, not total use of review platforms: their content still matters directly and through AI-generated answers.
G2’s own 2026 report has review sites and AI chatbots neck and neck as the biggest influence on the shortlist, 38% against 37%, and in a second G2 study 45% of buyers said a review-site citation is the most convincing thing an AI answer can include. G2’s 2026 acquisition of Capterra, Software Advice and GetApp also consolidated several major software discovery platforms under one owner.
So the advocacy work has to happen before the deal exists. Reviews, public answers and relevant customer stories need to address the questions buyers ask most — before a buyer you’ve never heard of asks an AI assistant to build a shortlist.
4. Where advocates come from: things customers already said
Most advocacy programs find advocates the hard way. The advocacy manager asks the account managers who might be willing, gets a few names, chases them, and one or two say yes. It takes weeks, it depends on who the manager happens to know, and it misses every happy customer nobody thought to mention.
Customers say good things all the time. On the call where the migration finished. In the last line of a support ticket they rated five out of five. In the comment box under a survey score. In a community thread where they helped someone else. Almost none of it goes anywhere, because the person who heard it was busy and the advocacy manager wasn’t in the room.
In 2027, finding advocacy opportunities need not depend on someone remembering what a customer said in a meeting. With approved access, an AI agent reads call transcripts, closed tickets, survey comments and community posts, and flags statements that are specific and useful: a number, a before and after, a problem that went away. That sentence becomes a nomination, not an automatic publication.
The agent drafts a message from the person the customer already talks to: “I was going back through my notes from our Tuesday call. You said the migration took two weeks instead of six. Would you mind if we quoted you on that on our website?”
That person checks the context, fixes the draft and sends it. Our rule for AI across all five programs is the same: it reads and drafts, and a person decides what reaches the customer.
Asking to publish something a customer already said is smaller than asking them to be a reference. An approved statement can appear on your website or in a public community. Reviews are different: invite customers to submit honest assessments in their own words, not just those likely to be positive. Check the original context, confidentiality and permission before publishing. A score on its own isn’t a nomination. A 9 tells you the customer is happy, not what they can speak to. The sentence next to the score is what the agent is listening for.
This works best when one person is accountable for the relationship and coordinates the request. That person knows who ran onboarding, what came up in the last review meeting and when the renewal is due. A cold email from someone in marketing the customer has never met is a bigger ask, and it raises the risk that the same customer gets approached twice.
5. Your advocates have less time than they used to
Microsoft’s 2025 Work Trend Index surveyed 31,000 people in 31 countries. 80% said they don’t have enough time or energy to do their own job. 53% of leaders said productivity has to go up. Microsoft also found interruptions approximately every two minutes during core working hours, with its most heavily interrupted users receiving hundreds of notifications a day. Those are your advocates. The ops lead who did the migration is now running AI agents on top of her own work and covering for a colleague who left in March.
Look at what an advocacy program asks that person for. A case study may involve an interview, rounds of review and legal approval on both sides. A video testimonial can require production time and a release form. A conference talk may mean travel and slide preparation. A reference call takes time on the prospect’s schedule, sometimes repeatedly. Each can produce valuable evidence, but each is also a substantial ask of someone whose primary job lies elsewhere.
For 2027, replace large, infrequent requests with smaller contributions at appropriate moments in the customer relationship. Approve a quote we pulled from your call. Answer one question from another customer in the community. Share an independent review in your own words. Agree to a sentence on our website with your name on it. These contributions can build a broader, more reusable body of evidence without demanding a full production every time. The bigger stories still happen, when the customer has a relevant outcome to share and is willing to devote the time.
6. The advocacy job belongs to whoever owns the customer
In the Customer Marketing Alliance’s 2026 State of Customer Marketing, 62% of companies have an advocacy program, advocates typically represent only a small share of the customer base, and 43% of customer marketing teams are one person, up about six points in a year. That’s a small team with a job description that may still list case studies, references, awards and events.
Small customer marketing teams cannot sustainably add new responsibilities while preserving every labor-intensive practice of the traditional model. The work that matters in 2027 is making specific customer evidence available: a review, a public answer, a quote with a name on it, a story with numbers. The relationship owner should be accountable for identifying opportunities and coordinating asks, even where specialists produce or publish the material. Last week’s guide said much the same about community. One owner can see which customer did the migration, which one might answer a question in public and which one has already taken a reference call — because the information sits in one shared view.
With appropriate access and controls, AI examines calls, tickets, surveys and community discussions for evidence worth developing. It drafts the ask, records permission, tags the result and helps check what AI assistants say about the company. The relationship owner decides whom to approach; a person verifies every statement before publication. The result should be a searchable evidence library, organized by industry, use case, measurable outcome, buyer role, recency and publication rights. Sales and marketing can then find usable proof without asking the same customer again.
7. Four decisions to make before the budget closes
1. What counts as proof. A true, specific customer statement that answers a buyer’s question and can be verified. Named quotations, independently submitted reviews, public answers and third-party-verified anonymized evidence all have a place. Generic claims and logo walls are not enough.
2. Where it lives. Make approved evidence available where buyers look: independent review sites, public communities, relevant customer posts and your own website. Maintain an internal evidence library for material with restricted permissions. Once a month, ask ChatGPT, Claude and Perplexity the ten questions buyers ask about you, inspect the sources they cite and note the gaps.
3. How customers get in, and how much you ask. Use approved AI workflows to flag specific customer statements from calls, closed tickets, surveys and community posts. The relationship owner checks the context and asks permission to reuse each statement. Keep requests small and well timed. Invite independent reviews through a fair, non-selective process; reserve substantial case studies for outcomes that warrant them.
4. Who owns it. One accountable relationship owner per customer, across Success, Community, Advocacy, Loyalty and Expansion. Specialists can support the work, but the owner coordinates requests and keeps the full customer context visible. For many organizations that means redesigned responsibilities and change management, not another task added to an advocacy manager’s list. Measure the relevance, availability, discoverability and use of customer evidence — not only advocate counts or assets produced.
Write the four down on one page and get it signed by whoever owns the 2027 budget. If it isn’t in the budget by December, you usually don’t get to do it.
8. Why one owner matters
In many companies the advocacy manager keeps one customer list, the community manager another, and the CSM a third. If those systems do not share context, a customer can hear from several teams in the same quarter, each asking for something. The issue is not the number of specialists. It is the absence of coordinated ownership.
Decision 4 makes the other three sustainable, and it’s the one many companies skip. All five goals sit within one function, with one accountable owner per customer relationship, AI assisting with analysis and drafts, and people deciding what reaches customers. That’s how the SCALE™ operating model works. Advocacy and community no longer operate as separate customer relationships: their specialists work through a coordinated post-sale function. Getting there may require a reorganization, a sponsor at the top, a plan for people whose roles change and several quarters of implementation. The Success and Community guides describe the same move from their side.
Ask yourself four things:
1. Does the person who asks a customer for a reference know when that customer’s renewal is?
2. Does the person who asks for a reference know what else that customer was asked for this month?
3. Can anyone on your team say which proof was used in which deal last quarter?
4. If a customer said something good on a call last week, did anyone ask to publish it?
Two or more negative answers suggest fragmented relationship ownership — a structural problem an advocacy program cannot resolve on its own.
9. In one line
Make proof before the deal, where buyers can find it, from things customers already said, one well-timed ask at a time.
This week, read last month’s call notes and survey comments and count the good things customers said that nobody asked to publish. Then ask an AI assistant what customers say about your company, examine the sources it cites and identify what’s missing. And think about the last reference call your team set up. Did the buyer ask for it, or did sales?
If you’re setting your 2027 customer advocacy strategy, we can help. Advocacy Maven has worked on more than 300 customer programs across all five SCALE™ programs: Success, Community, Advocacy, Loyalty, and Expansion. See how we work, or read what our clients say about the programs we have built with them.
Related guides
2027 Customer Success Strategy: Own the Whole Relationship
2027 Customer Community Strategy: Life After How-To
2027 Loyalty Strategy: coming October 26
2027 Expansion Strategy: coming November 2
2027 AI Discoverability Strategy: coming November 9
Summary: 2027 Customer Advocacy Strategy: Get It on the Record
A successful 2027 customer advocacy strategy moves beyond traditional case studies and repeated reference calls to create verified, discoverable customer proof that influences buyers before they contact sales. By capturing specific outcomes customers have already shared, using AI to identify advocacy opportunities, and making approved evidence accessible through reviews, communities, and other trusted sources, companies can strengthen buyer confidence without overburdening advocates. The 2027 customer advocacy strategy also requires coordinated relationship ownership, smaller and better-timed requests, and clear measures of evidence quality, visibility, and commercial use. Under the SCALE™ operating model, advocacy becomes part of a unified post-sale function, turning authentic customer experiences into reusable proof that supports trust, retention, and growth.
FAQ 1: Is the customer reference call dead in 2027?
No. It remains useful when a buyer needs to discuss a specific implementation, risk or outcome directly with a peer. UserEvidence’s 2025 survey found 42% trust in one-to-one references, compared with 64% for named testimonials. And 6sense found that 94% of buying groups ranked their shortlist before talking to a seller. The reference call often validates a preference already formed rather than creating it.
Keep it for situations where a conversation adds something written evidence cannot. Earlier in the journey, make customer proof discoverable and reusable through reviews, public answers and relevant stories.
FAQ 2: What should a 2027 customer advocacy strategy include?
Four decisions: what counts as credible proof, where buyers and sales can find it, how customers are asked to participate, and who is accountable for the relationship. Begin with verified statements that answer buyer questions; publish them according to customer permissions; review what AI assistants cite; and measure how often evidence is available and used.
For the first quarter, review recent call notes, closed tickets and survey comments. Identify specific outcomes worth documenting, verify the context, and invite customers to approve appropriate uses. Keep independent review outreach separate and representative of the customer base.
FAQ 3: How do we find customer advocates without chasing them?
Start with what customers have already said on calls, in support tickets, in survey comments and in public community discussions. With approved access, an AI agent flags statements that contain a specific outcome and keeps the source and date attached. The relationship owner checks the context and asks whether the customer is willing to make the statement available.
The ask is small because it begins with the customer’s own words and comes from someone who knows the relationship. An approved quote might appear on your website or in the community; other evidence can be retained for restricted sales use. Independent reviews must be written by customers themselves, following the review platform’s rules. You can still develop substantial case studies when the outcome deserves that treatment.
FAQ 4: Should we still do case studies and video testimonials in 2027?
Yes, but not necessarily as the first ask. Case studies remain valuable when they provide relevant context and measurable outcomes; videos can make a customer’s experience more tangible. Both may require interviews, production and review time. Microsoft’s 2025 Work Trend Index found that 80% of respondents lacked sufficient time or energy for their own work. Start by asking for a small, useful contribution: permission to quote a result, an answer to a community question, or an independent review in the customer’s own words.
Develop a full case study when the customer’s experience merits a deeper account and the customer is willing to participate. A strong, specific story is more useful than several generic ones, whether it starts with a brief quotation or a formal interview.
FAQ 5: How does AI change customer advocacy?
AI changes how customer evidence is discovered and summarized. Muck Rack found that 84% of citations across ChatGPT, Claude and Gemini came from earned media, a broad category that includes journalism and other independent sources. Separately, G2 found that 45% of buyers considered a review-site citation especially convincing in an AI answer. Advocacy therefore needs to produce verifiable evidence and make it available where buyers are likely to encounter it — not assume every published statement will appear in an AI response.
AI also assists with the work. With approved access, it scans transcripts, tickets, surveys and community discussions for relevant customer outcomes, drafts permission requests, records responses and helps assess AI visibility. A person verifies the original statement, decides whom to approach and approves what is published. The AI discoverability guide in this series, out November 9, covers the testing method in full.
FAQ 6: How does advocacy fit into the SCALE™ operating model?
Advocacy is one of five coordinated programs: Success, Community, Advocacy, Loyalty and Expansion. SCALE™ unifies their accountability under one owner per customer relationship, while specialists can contribute across programs. AI identifies potential proof in calls, tickets, surveys and community activity and drafts the request; the owner checks the context, coordinates outreach and secures permission. Each approved contribution is tagged for appropriate reuse, and reporting covers evidence quality, discoverability and commercial use.
The first three decisions can begin this quarter without changing the organization chart. The fourth establishes lasting accountability and may call for a broader redesign of the post-sale function. The Success guide describes the same change from the retention side.
Co-founder and Chief Customer Officer at Advocacy Maven, the customer programs agency that helps companies turn every customer program into a trusted growth engine, Diana brings nearly 20 years of experience as a global strategist, local implementer, and agency operator, advising companies on how to design, manage, and scale their Success, Community, Advocacy, Loyalty, and Expansion programs.
Since 2013, Advocacy Maven has supported more than 300 customer programs globally, created 5,000+ customer assets in more than 30 languages, and mobilized over a quarter million advocates. The agency is consistently recognized as a leader by G2 across Marketing Strategy, Brand Advocacy, and Content Writing services.
