How To Choose A Content Writing Agency: 10 Questions To Ask Before Signing

Updated March 01, 2026

Your CEO just forwarded a ChatGPT screenshot. Three competitors are cited as the top picks for your exact use case. Your brand isn't mentioned once. You rank on page 1 of Google for dozens of keywords and invest heavily in content each month. So why are you invisible where buyers actually research?

More B2B SaaS marketing leaders face this situation every quarter, and it makes the traditional agency evaluation process dangerously outdated. Picking an agency based on writing samples, tone matching, and cost per word made sense in 2020. Today, that approach misses the one thing that actually drives pipeline: whether your content gets cited by AI systems when buyers ask for vendor recommendations.

We've built a 10-question framework to help you vet any content writing agency against the standard that matters now, plus the red flags, comparison data, and board-ready ROI metrics you need to make the decision stick internally.

Why traditional content metrics fail in the age of AI search

We don't see traffic as the goal anymore, at least not on its own. For most of the last decade, content agency success meant page-1 rankings, domain authority, and organic session growth. Those metrics still have value, but they describe a world where Google's 10 blue links were the only meaningful discovery surface.

That world has changed, and we've watched it happen. Two-thirds of B2B buyers now say they rely on AI chatbots as much as or more than Google when evaluating vendors, according to research published by Digital Commerce 360. More importantly, a 30-day study found that AI-referred visitors converted at 23x the rate of traditional organic visitors, accounting for just 0.5% of traffic but 12.1% of all signups.

That gap, between who shows up in AI answers and who captures high-converting buyers, is where the new content ROI story is being written.

Answer Engine Optimization (AEO) is the practice of structuring content so it gets cited as a direct answer by AI models like ChatGPT, Perplexity, and Google AI Overviews. As Marcel Digital explains, SEO earns clicks while AEO earns mentions. Those are different technical challenges, and we see most agencies still optimizing for the wrong one. For a full breakdown, our guide to what AEO is and how it works covers the mechanics in depth.

Here's what this means for you: if you evaluate an agency on word count, brand voice, or keyword rankings alone, you may end up with polished content that no AI platform retrieves. That means missing the growing share of buyers who use AI to build their vendor shortlists before they ever visit your site.

The new evaluation framework: 10 questions to ask prospective agencies

We've built this as the core of your vetting process. Run every agency through these questions before you discuss pricing or timelines.

1. "What is your specific methodology for earning AI citations?"

Use this as your filter question. A credible agency should explain, in plain terms, how they structure content so LLMs retrieve it. Look for references to entity optimization, structured data (specifically JSON-LD schema), third-party validation signals, and direct-answer formatting. If they only say "we write high-quality content," they're not doing AEO.

We built the CITABLE framework specifically for LLM retrieval, with seven components:

• C - Clear entity & structure: Every piece opens with a 2-3 sentence direct answer (BLUF) that gives AI a clean passage to extract.
• I - Intent architecture: Content answers the main query plus adjacent questions buyers ask at the same stage.
• T - Third-party validation: Reviews, community mentions, UGC, and news citations build the external trust signals LLMs weight heavily.
• A - Answer grounding: Every factual claim is tied to a verifiable source. Vague, unverifiable content gets deprioritized by AI platforms.
• B - Block-structured for RAG: Sections run 200-400 words with tables, FAQs, and ordered lists to support Retrieval-Augmented Generation.
• L - Latest & consistent: Explicit timestamps and unified facts across all owned and third-party properties signal recency to AI systems.
• E - Entity graph & schema: Explicit relationships between your brand, product, use cases, and buyer segments are coded into the copy and schema markup.

If an agency can't walk you through something this specific, they're not doing AEO.

2. "How do you track share of voice in AI answers?"

Share of voice in AI search means the percentage of relevant buyer-intent queries where your brand appears compared to competitors. As our AI citation tracking comparison explains, tracking this requires systematic prompt testing across platforms, not a rankings dashboard.

Ask to see a sample report. It should show citation rates by query, platform breakdowns across ChatGPT, Perplexity, Claude, and Google AI Overviews, and a competitor comparison. If they say they "monitor Google Analytics for AI referral traffic," walk away. That's not AI visibility tracking.

3. "Can you show me an attribution model for AI-referred pipeline?"

Citations mean nothing to your CFO unless they connect to revenue. The right agency implements UTM tagging from day one, integrates with Salesforce or HubSpot, and tracks AI-referred leads through to closed-won. Ask specifically how they handle attribution for zero-click research, where a buyer reads an AI answer and navigates directly to your site later, bypassing referral tracking.

We use Predictive Performance Modeling to project expected pipeline impact before the engagement starts, so you can model ROI for budget approval rather than waiting months for data.

4. "Do you use subject matter experts or generalist writers?"

LLMs look for technical depth, industry-specific terminology used correctly, and original insight. According to Contently's analysis of leading LLM agencies, high-quality content with "information gain" (saying something others haven't said) matters more for AI retrieval than for traditional SEO. Ask to see a sample post, and ask which parts reflect the writer's original expertise versus rephrased existing sources.

5. "What is your fact-checking process?"

The CITABLE framework requires answer grounding, meaning every factual claim has a verifiable source. AI platforms that detect inaccuracies learn to deprioritize those sources over time, as Surfer SEO's research on LLM citations documents. Ask about the fact-checking step in their production workflow and how they handle corrections when data changes.

6. "How fast can we expect to see initial results?"

Traditional SEO operates on 3-6 month timelines, but AEO can move faster because AI platforms crawl and update content continuously. We target initial citations for long-tail buyer queries within 1-2 weeks, with meaningful share of voice gains measurable by 90 days. If an agency can't give you a timeline backed by a specific methodology, they're guessing.

7. "What is your contract structure?"

AI search platforms are evolving fast enough that 12-month lock-ins carry real risk. Agencies confident in their results offer month-to-month terms. Those that aren't confident need annual contracts to protect revenue. Treat any pressure to sign a long-term retainer before demonstrating initial citations as a red flag.

We operate on month-to-month terms, with transparent pricing available before you commit.

8. "How do you handle zero-click searches?"

Most B2B AI research never produces a click. A buyer asks ChatGPT "what's the best workflow automation tool for mid-market SaaS?" and gets a structured answer with your competitor's name. No click happens, but the buyer is now biased. We understand the goal of AEO is influence at the moment of decision, not website traffic. If an agency can only explain their value through traffic metrics, they don't understand the channel. Our guide to AI citation patterns breaks down how each platform makes citation decisions.

9. "Do you optimize for platforms beyond ChatGPT?"

Platform-specific optimization is essential because ChatGPT, Perplexity, Claude, and Google AI Overviews each use different signals to select citations. Optimizing for "AI search" as a single channel is like running one creative across LinkedIn and TikTok and expecting equal results. Ask for platform-specific examples. You can find Claude-specific optimization guidance in our Claude AI optimization guide and Google AI Overviews coverage in our explainer on how it works.

10. "Can you share a case study with pipeline numbers, not traffic?"

Traffic reports are easy to produce and hard to tie to revenue. Ask for a before/after showing citation rate at baseline, share of voice gains, AI-referred MQL volume, and pipeline contribution in dollars. One of our clients grew from 550 AI-referred trials to over 2,300 in four weeks, a result measurable in their CRM, not just their analytics dashboard.

Red flags: how to spot agencies bluffing about AI optimization

Even well-meaning agencies can be a year or two behind on what AI-driven buyer research actually requires. Watch for these patterns:

• The "AI-powered" gimmick: Using ChatGPT or Jasper to write content faster is not AEO. As Contently notes, the agencies delivering real results understand that LLMs don't rank pages the way Google does. Content generation is only one part of the equation.
• The SEO rebrand: If the strategy still centers on backlinks, meta descriptions, and Core Web Vitals with "AI" added to the pitch, it's traditional SEO with different packaging. Brand recognition and topical authority are the strongest predictors of AI citations, not backlink count.
• The black box: Any agency unwilling to explain their methodology in plain terms is either guessing or protecting a process that won't hold up to scrutiny. Legitimate AEO partners explain exactly what they do and why it works.
• Pricing avoidance: Agencies that refuse to give a pricing range before a discovery call are often generalists pricing on the fly after they gauge your budget. We publish our pricing on our website.

You can also run a competitive technical SEO audit to benchmark your current infrastructure before engaging any agency, which makes the vetting conversation far more specific.

Specialist vs. generalist: which model drives pipeline faster?

The debate between specialist and generalist agencies is clearer now than it's ever been. AEO builds on your existing SEO foundation but extends it with platform-specific citation strategies, entity optimization, and daily publishing cadences that generalist teams can't sustain while also managing paid ads, social, and web design.

For a direct agency methodology comparison, our Animalz vs. Directive breakdown shows how different models perform across the metrics that matter for B2B SaaS.

Making the business case: ROI and attribution for the board

Your CFO wants payback period and pipeline math, not citation rates. Here's how to translate AEO metrics into board-ready language.

Before/After: AEO impact on key metrics

The key metrics to track for your CFO conversation are:

• Citation rate: How frequently your brand appears in AI responses for your top buyer-intent queries
• Share of voice: Your citation percentage vs. your top three competitors (start with a baseline audit)
• Cost per AI-referred MQL: Total agency cost divided by AI-sourced qualified leads
• AI traffic conversion rate: AI-referred visitor conversion compared to traditional organic

Typeface's CMO AI ROI guide recommends building three scenarios (conservative, likely, optimistic) so you can defend the investment if early results come in below projection. For 15 specific AEO tactics to present alongside this ROI case, our AEO best practices guide covers the full playbook.

The right partner is not the one with the best writing samples. It's the one who can show you exactly where you're invisible in AI answers, explain why, and close that gap in a timeline your board can hold them to.

Book an AI visibility audit with Discovered Labs to see your current citation rate vs. competitors across your top buyer-intent queries. It's the clearest way to enter any agency conversation knowing exactly what you need.

Frequently asked questions

How long does it take to get cited by ChatGPT?

With the right entity structure, block formatting, and third-party validation in place, initial citations for long-tail buyer queries can appear within 1-2 weeks. Meaningful share of voice gains are typically measurable at 90 days.

Can't we just use ChatGPT to write content and save the cost?

Using AI to write content is straightforward, but getting AI to retrieve and cite your content requires entity structure, schema markup, third-party validation signals, intent mapping, and platform-specific optimization. As Contently's analysis of leading LLM agencies shows, content generation is only one component of a working AEO strategy.

What if my current SEO agency says they can do AEO too?

Run them through the 10 questions above. If they can't show you an AI visibility report with competitor share of voice or explain their citation methodology specifically, they're describing traditional SEO with an updated label.

How do I track AI-referred leads in Salesforce?

Implement UTM tagging for AI referral sources from day one, combined with "how did you hear about us?" fields at conversion to catch zero-click research patterns. Perplexity and some Claude integrations pass referral data through standard HTTP headers, while ChatGPT-sourced visits often appear as direct traffic.

Key terms glossary

AEO (Answer Engine Optimization): The practice of structuring content to be cited as a direct answer by AI models, including ChatGPT, Perplexity, Claude, and Google AI Overviews. Distinct from SEO in that the goal is citation, not ranking.

Citation rate: The percentage of times your brand is mentioned in AI-generated answers for a defined set of buyer-intent queries. Track via weekly prompt testing across platforms.

Share of voice: Your brand's citation percentage compared to competitors across a defined query set. Measured by running the same buyer-intent prompts across platforms and logging which brands appear.

GEO (Generative Engine Optimization): The broader strategic category covering all initiatives designed to improve brand visibility in AI-generated answers, including content structure, authority building, third-party mentions, and entity optimization.

AI-referred MQL: A marketing-qualified lead whose research path included an AI platform, tracked via UTM parameters, referral source data, or self-reported attribution. Converts at significantly higher rates than traditional organic MQLs, according to conversion data from Ahrefs.