AfterQuery growth playbook

Add a "For Experts" item to the top navigation and footer, linked to experts.afterquery.com with tracked links. Add a short experts section to the homepage: who AfterQuery hires (by domain), how the work is paid, and how the application works, with a link to the Help Center. Add a banner on the careers page that separates employee roles from contract expert work. On the contact page, move the experts link above the form as a clear button. Replace "automatically discarded" with an auto-reply that sends people to the experts site. Measure the new traffic source in the expert funnel.

Brand search from press coverage and research posts already reaches afterquery.com. A visible door catches experts who now bounce or write in and get discarded. The door costs almost nothing and feeds the supply targets that three new hires are measured on.

[ ] "For Experts" in header and footer (tracked link to experts.afterquery.com)
[ ] Homepage section: domains hired, how pay works, how the application works
[ ] Link to the Experts Help Center
[ ] Careers banner: employee roles vs contract expert work
[ ] Contact page: experts button above the form
[ ] Auto-reply instead of "automatically discarded"
[ ] Main-site source tracked through application, assessment and first task

• More low-fit applicants

  List the domains and credentials wanted, and let the experts site's screening do the rest.

• Buyers confused by an expert link

  Keep "Get data" as the main button and put experts second.

• Your websiteHomepage (checked twice on November 9, 2026): H1 "We teach machines how experts think."; 31 links (navigation: Research & Blog, For Enterprises, Data Partnerships, Our Products, Careers, "Get data"), none to the experts site or the experts help center; the only "join" is "Join the team revolutionizing AI Research and Training"; case results: a dataset worth "11.4 GDPval points", a legal benchmark with "5,161 cases across 28 practice areas", and perfect solves rising "from 47% to 74%".Source
• Your websiteContact page: a buyer form ("What are you interested in?": "Request a Custom Dataset", "Browse our Off-the-shelf Datasets", "Enterprise AI Consulting", "End-to-end Enterprise AI Implementation"); below it: "If you're looking for our Experts site (contract jobs), please refer to experts.afterquery.com." and "All messages seeking Expert roles are automatically discarded."; no stated response time.Source
• Your websiteCareers page: "Shape how AI learns."; 41 employee roles across Engineering, Growth, Internal Ops, Operations, Research, Revenue and Marketing; no mention of the expert network or of contract work for experts.Source
• HiringGrowth Associate: "grow our expert network from 100K to hundreds of thousands and beyond"; "Build, optimize, and scale automated expert sourcing pipelines"; "Expand AfterQuery's Conference footprint"; "Expand Campus Partnership efforts with elite universities and MBA/PHD programs".Source
• Your websitedocs.afterquery.com is the "AfterQuery Experts Help Center" ("Contributor Pay Basics", "AfterQuery Experts Application Process", "Working with AfterQuery Experts"); the main site's navigation and footer do not link to it.Source
• Competitor researchRivals' own homepages: Mercor's top navigation has "Experts" with "Find work" and shows "$90/hr average contracted rate", "$4M+ daily payouts" and "5M+ experts"; micro1 links "Join as an expert", "Refer & earn" and "Interview Prep", plus a data-partnership referral program; Turing has "Join expert network" in its header and "Off-The-Shelf Datasets"; Surge AI's navigation has "Off-the-Shelf", "Expert Workforce" and a list of its public benchmarks.See the research Source

Agree one definition (for example, experts who passed verification, and the subset active in the last 90 days) and one date. Publish it on the homepage, products page and press kit, and add per-domain counts where they help sales (attorneys, clinicians, engineers). Update the number quarterly from the marketplace database.

A consistent, defined number is more believable than a big one, and an "active" figure answers the buyer's real question: can you staff my project?

• The active number looks small

  Show both verified and active, with per-domain depth.

• Your websiteSeries A announcement (April 9, 2026): "$30 million Series A at a $300 million valuation, led by Altos Ventures"; "surpassed $100 million in revenue run rate"; "nearly 100,000 verified practicing professionals"; the money goes partly to "expanding our expert network".Source
• Your websiteProducts page: 12 offerings, including Off-the-shelf Data ("Request access" goes to the contact form), Loss Analyses, Custom Evals and Professional Domains with "300,000+ verified practitioners"; no prices and no page for any individual dataset.Source
• HiringGrowth Associate: "grow our expert network from 100K to hundreds of thousands and beyond"; "Build, optimize, and scale automated expert sourcing pipelines"; "Expand AfterQuery's Conference footprint"; "Expand Campus Partnership efforts with elite universities and MBA/PHD programs".Source

Add a short terms box to /data-partnerships/refer: how the reward is calculated (for example, a share of the first payment, capped at $85k), when it is paid, who qualifies, and an example. Add the page to the sitemap. Send each referrer a status email at every stage (received, NDA, sample, terms).

Clear terms turn a headline number into a reason to act, and status updates keep referrers sending the next introduction.

• Low-quality introductions

  Pay only at signed terms and state the fit criteria.

• Your websiteData Partnerships page: "Your company's data, licensed with care."; operating, winding-down and closed companies are welcome; personal details removed or pseudonymized and a re-identification test before delivery; six steps from inquiry to payment; a "Refer and earn up to $85k" button. The /data-partnerships/refer form gives no amount, eligibility or payment timing, only that referrals are tracked so "payment is ensured if we move forward with the partnership"; that page is not in the sitemap.Source
• Your websiteSitemap (34 URLs, checked November 9, 2026): 15 blog posts and 9 knowledge pages, all knowledge pages dated April 8, 2026 with no author; no page for experts, no dataset catalogue, no industry or comparison pages, no security or trust page.Source
• HiringData Acquisition Lead: "Personally lead negotiation of AfterQuery's highest-value, most complex enterprise data-purchase agreements" and "set offer prices, deal structures and walk-away points in a market with little public pricing data", ideally deals of "seven figures or more". A Quality Control Lead will "Own quality metrics and KPIs, reporting on error rates, root causes, and improvement trends".Source

Branch the form on the interest dropdown. Off-the-shelf browsers see the dataset list (or a sample request) at once. Custom dataset and enterprise buyers get a calendar link and a stated reply time. Data partners go to the partnership start page. Log the branch in Attio so attribution shows which path closes.

Instant next steps reduce drop-off between form and first call, and routing by intent puts each lead with the right team.

• Calendar slots wasted by poor fits

  Show the calendar only after company and use-case fields.

• Your websiteContact page: a buyer form ("What are you interested in?": "Request a Custom Dataset", "Browse our Off-the-shelf Datasets", "Enterprise AI Consulting", "End-to-end Enterprise AI Implementation"); below it: "If you're looking for our Experts site (contract jobs), please refer to experts.afterquery.com." and "All messages seeking Expert roles are automatically discarded."; no stated response time.Source
• HiringRevenue hiring: a GTM Operations Lead to "Architect CRM and pipeline hygiene" and "Own analytics & attribution", including "ROI on sales events (conferences, dinners, etc.)", on Attio and Ramp; an Engagement Manager as "primary point of contact for key frontier AI labs" who will "Identify and close upsell, cross-sell, and new use-case opportunities"; an Associate to "Create and maintain sales collateral".Source
• Your websiteProducts page: 12 offerings, including Off-the-shelf Data ("Request access" goes to the contact form), Loss Analyses, Custom Evals and Professional Domains with "300,000+ verified practitioners"; no prices and no page for any individual dataset.Source

Serve the hero image in AVIF/WebP at the displayed size with priority loading. Inline or defer the render-blocking CSS and fonts. Re-test in PageSpeed and watch the field LCP over the next 28 days.

Faster first paint keeps more mobile visitors, and passing Core Web Vitals removes a ranking handicap for brand and category searches.

• Visual quality drops

  Compare at common phone widths before shipping.

• PageSpeedGoogle PageSpeed (November 9, 2026): mobile score 85, desktop 99; real Chrome users' LCP is 4.4 s at the 75th percentile, so the homepage fails Core Web Vitals on LCP (INP 136 ms and CLS 0.09 pass); the LCP element is an image, with 309 KB of image savings and 550 ms of render-blocking requests flagged.Source
• Your websiteHomepage (checked twice on November 9, 2026): H1 "We teach machines how experts think."; 31 links (navigation: Research & Blog, For Enterprises, Data Partnerships, Our Products, Careers, "Get data"), none to the experts site or the experts help center; the only "join" is "Join the team revolutionizing AI Research and Training"; case results: a dataset worth "11.4 GDPval points", a legal benchmark with "5,161 cases across 28 practice areas", and perfect solves rising "from 47% to 74%".Source

Add a standard end block to every post. It names the dataset or service used (for example, the Off-The-Shelf Office Agent Training Dataset, a custom eval, post-training), gives one line on what a buyer gets, and offers a "Request a sample" or "Scope a similar eval" button that pre-fills the contact form with the post's name. Add the same block to the four case-study cards on the For Enterprises page.

Readers who just saw a measured result have the strongest intent, and pre-filled context lets sales open the conversation with that result.

• Research feels salesy

  Keep the block short, below the conclusion and factual.

• Your websiteResearch page: 15 posts from October 2025 to November 8, 2026; public benchmarks SpreadsheetBench 2 and Legora BAR; the newest benchmark, CrossSearch, has only a blog post. The GDPval case study used AfterQuery's "Off-The-Shelf Office Agent Training Dataset" (GDPval 35.3 to 46.7, against a teacher's 49.5) and ends: "Get in touch here to access our off-the-shelf GDPval, Office Agent, and agentic post-training datasets".Source
• Your websiteProducts page: 12 offerings, including Off-the-shelf Data ("Request access" goes to the contact form), Loss Analyses, Custom Evals and Professional Domains with "300,000+ verified practitioners"; no prices and no page for any individual dataset.Source
• HiringTechnical Content Writer (Contract, posted September 21, 2026): "Write and adapt content for AfterQuery's brand and founder social media accounts, including X and LinkedIn", about 10-15 hours a week to start.Source

Build /datasets with one page per sellable dataset: capability area, size, task format, expert domains, quality checks, licence terms, and a measured result where one exists (for example, the GDPval lift). Each page offers a sample request and a scoping call. Keep prices on request, but show the licence options.

Buyers can qualify themselves, procurement gets the facts it needs, and every dataset page can rank for the capability it teaches.

• Customers' exclusivity terms

  List only datasets cleared for non-exclusive licence.

• Competitors copy the formats

  Keep samples gated and the full specs under NDA.

• #4 Split the contact form so dataset browsers, enterprise buyers and data partners each get the right next step

• Your websiteProducts page: 12 offerings, including Off-the-shelf Data ("Request access" goes to the contact form), Loss Analyses, Custom Evals and Professional Domains with "300,000+ verified practitioners"; no prices and no page for any individual dataset.Source
• Your websiteResearch page: 15 posts from October 2025 to November 8, 2026; public benchmarks SpreadsheetBench 2 and Legora BAR; the newest benchmark, CrossSearch, has only a blog post. The GDPval case study used AfterQuery's "Off-The-Shelf Office Agent Training Dataset" (GDPval 35.3 to 46.7, against a teacher's 49.5) and ends: "Get in touch here to access our off-the-shelf GDPval, Office Agent, and agentic post-training datasets".Source
• Competitor researchRivals' own homepages: Mercor's top navigation has "Experts" with "Find work" and shows "$90/hr average contracted rate", "$4M+ daily payouts" and "5M+ experts"; micro1 links "Join as an expert", "Refer & earn" and "Interview Prep", plus a data-partnership referral program; Turing has "Join expert network" in its header and "Off-The-Shelf Datasets"; Surge AI's navigation has "Off-the-Shelf", "Expert Workforce" and a list of its public benchmarks.See the research Source
• Your websiteContact page: a buyer form ("What are you interested in?": "Request a Custom Dataset", "Browse our Off-the-shelf Datasets", "Enterprise AI Consulting", "End-to-end Enterprise AI Implementation"); below it: "If you're looking for our Experts site (contract jobs), please refer to experts.afterquery.com." and "All messages seeking Expert roles are automatically discarded."; no stated response time.Source

Build a status view in the expert dashboard: application stage, assessment result (pass, not yet, waitlist by domain), tasks awaiting review with an expected date, and the next payout date. Publish two service levels on the experts site: application decisions within a set number of days, and reviews within a set number of days, with a monthly on-time rate. Send automatic updates when a project pauses.

Visible timelines turn silence into trust. Experts who know where they stand stay active and refer peers, which raises supply without extra sourcing spend.

• Missing a published service level

  Start with conservative targets and publish the real on-time rate.

• Expert reviewsAnnotation Academy (a seller of AI-evaluation training) analysed 44 Reddit threads and 327 comments, mostly May-September 2026: AfterQuery is "a real platform" and 16 accounts reported being paid; 11 accounts reported no update for weeks or months after applying, 7 described task reviews taking weeks or months, 12 described paused or cancelled projects, and 6 reported Stripe payout problems in some countries.Source
• HiringHead of Growth (posted July 29, 2026): "Design and execute the strategy for sourcing expert independent contractors across professional domains"; "Forecast supply needs based on customer and project demand"; "Measure expert satisfaction and lead the team in ensuring experts are unblocked and satisfied with their experience".Source
• HiringMarketing Lead: "own AfterQuery's expert marketplace strategy, acquisition, activation, matching, retention, and quality", "setting the metrics and cadences that track staffing speed, conversion, and quality", and "go-to-market and brand strategy for technical and enterprise audiences". A Software Engineer - Marketplace will "Build end-to-end product experiences across expert acquisition, onboarding, assessments, matching".Source

Create a page per domain, in the professional's language: the kind of work (for example, writing legal rubrics or grading clinical reasoning), credentials wanted, typical time commitment, how pay is set, and a story from a contributor who agreed to be quoted. Link each page to the matching application path. Use the pages in professional associations, alumni groups and conference outreach.

Specialists respond to work described in their own terms. Each page also ranks for searches like "AI training work for attorneys" that generic pages miss.

• Promising work that pauses

  Show waitlists honestly by domain.

• #1 Give experts a front door on afterquery.com instead of a line under the buyer form that says their messages are discarded

• HiringStrategic Projects Leads in Coding, Healthcare, Legal, AI Safety, Consulting, STEM, Engineering (MechE / EE / IE) and Life Sciences & Chemical, each to "Lead and coordinate highly specialized contributors"; the Legal lead will "Recruit and lead practicing attorneys and specialists across practice areas and jurisdictions"; the AI Safety lead recruits "red teamers, security researchers, trust & safety practitioners".Source
• Your websiteSitemap (34 URLs, checked November 9, 2026): 15 blog posts and 9 knowledge pages, all knowledge pages dated April 8, 2026 with no author; no page for experts, no dataset catalogue, no industry or comparison pages, no security or trust page.Source
• HiringGrowth Associate: "grow our expert network from 100K to hundreds of thousands and beyond"; "Build, optimize, and scale automated expert sourcing pipelines"; "Expand AfterQuery's Conference footprint"; "Expand Campus Partnership efforts with elite universities and MBA/PHD programs".Source

Create /enterprise/legal, /finance and /banking pages (adding healthcare when a case study exists). Each one covers the workflows where models break, the evaluate-curate-post-train-deploy path for that industry, the matching case-study result in plain language, and the expert depth available (practising attorneys, credit analysts). Use the pages in outbound and as leave-behinds.

Enterprise buyers search and forward by industry. A page in their language shortens the outbound conversation and gives champions something to send internally.

• Overclaiming results

  Quote only published numbers, with links.

• HiringHead of Enterprise GTM (posted November 9, 2026): "Personally drive outbound prospecting, relationship development, pipeline generation, and strategic account acquisition"; experience "selling complex, high-value solutions to frontier AI labs, Fortune 500 companies"; "Identify the highest-value customer use cases for custom evaluations, data generation, model training".Source
• Your websiteFor Enterprises page: "Services across the full evaluation, training, and deployment lifecycle" (Evaluate, Curate, Post-train, Deploy with forward-deployed engineers); custom benchmarks can be re-run against new models, checkpoints and vendors; four case studies (Legora, NVIDIA, Hebbia, The Raine Group); a "Trusted by" strip whose logos carry no names.Source
• Your websiteSitemap (34 URLs, checked November 9, 2026): 15 blog posts and 9 knowledge pages, all knowledge pages dated April 8, 2026 with no author; no page for experts, no dataset catalogue, no industry or comparison pages, no security or trust page.Source

Publish /trust covering data handling for each line (lab projects, enterprise deployments, data partnerships): access controls, the de-identification and re-identification process, NDAs, retention and deletion, contractor confidentiality, and current certifications or audits (state only what is true). Offer a security pack on request, and link it from the enterprise pages and the data-partnership start page.

Answering the standard questions upfront shortens security review and reassures data partners. Few data vendors explain de-identification in plain language.

• Claims that outrun practice

  Publish only controls already in place, reviewed by counsel.

• Your websiteData Partnerships page: "Your company's data, licensed with care."; operating, winding-down and closed companies are welcome; personal details removed or pseudonymized and a re-identification test before delivery; six steps from inquiry to payment; a "Refer and earn up to $85k" button. The /data-partnerships/refer form gives no amount, eligibility or payment timing, only that referrals are tracked so "payment is ensured if we move forward with the partnership"; that page is not in the sitemap.Source
• Your websiteSitemap (34 URLs, checked November 9, 2026): 15 blog posts and 9 knowledge pages, all knowledge pages dated April 8, 2026 with no author; no page for experts, no dataset catalogue, no industry or comparison pages, no security or trust page.Source
• HiringHead of Enterprise GTM (posted November 9, 2026): "Personally drive outbound prospecting, relationship development, pipeline generation, and strategic account acquisition"; experience "selling complex, high-value solutions to frontier AI labs, Fortune 500 companies"; "Identify the highest-value customer use cases for custom evaluations, data generation, model training".Source
• Competitor researchRivals' own homepages: Mercor's top navigation has "Experts" with "Find work" and shows "$90/hr average contracted rate", "$4M+ daily payouts" and "5M+ experts"; micro1 links "Join as an expert", "Refer & earn" and "Interview Prep", plus a data-partnership referral program; Turing has "Join expert network" in its header and "Off-The-Shelf Datasets"; Surge AI's navigation has "Off-the-Shelf", "Expert Workforce" and a list of its public benchmarks.See the research Source

Package a "benchmark re-run" add-on for every custom eval: quarterly or per-release runs, a failure report by category, and a recommendation for the data or environment that would close each gap. Offer it at handover and price it as a subscription. Each report becomes the Engagement Manager's agenda for the next expansion.

Recurring re-runs turn project revenue into renewals, and each failure report points to the next dataset to sell.

• Benchmark contamination over time

  Refresh a share of tasks with each run.

• Your websiteFor Enterprises page: "Services across the full evaluation, training, and deployment lifecycle" (Evaluate, Curate, Post-train, Deploy with forward-deployed engineers); custom benchmarks can be re-run against new models, checkpoints and vendors; four case studies (Legora, NVIDIA, Hebbia, The Raine Group); a "Trusted by" strip whose logos carry no names.Source
• HiringRevenue hiring: a GTM Operations Lead to "Architect CRM and pipeline hygiene" and "Own analytics & attribution", including "ROI on sales events (conferences, dinners, etc.)", on Attio and Ramp; an Engagement Manager as "primary point of contact for key frontier AI labs" who will "Identify and close upsell, cross-sell, and new use-case opportunities"; an Associate to "Create and maintain sales collateral".Source

Assign each article to a domain lead or research scientist as named reviewer, add an updated date, add one worked example from AfterQuery's own work, and end with a link to the matching product or catalogue page. Add two new articles a month on questions from sales calls.

Named experts and fresh dates are signals that search engines and AI assistants weigh, and buyers trust a page more when an expert signs it.

• Experts lack time

  Ask for a 20-minute review, not a draft.

• Your websiteSitemap (34 URLs, checked November 9, 2026): 15 blog posts and 9 knowledge pages, all knowledge pages dated April 8, 2026 with no author; no page for experts, no dataset catalogue, no industry or comparison pages, no security or trust page.Source
• HiringTechnical Content Writer (Contract, posted September 21, 2026): "Write and adapt content for AfterQuery's brand and founder social media accounts, including X and LinkedIn", about 10-15 hours a week to start.Source
• HiringStrategic Projects Leads in Coding, Healthcare, Legal, AI Safety, Consulting, STEM, Engineering (MechE / EE / IE) and Life Sciences & Chemical, each to "Lead and coordinate highly specialized contributors"; the Legal lead will "Recruit and lead practicing attorneys and specialists across practice areas and jurisdictions"; the AI Safety lead recruits "red teamers, security researchers, trust & safety practitioners".Source

Give every active expert a referral link, with a bonus paid when the referred expert's first task is approved (more for domains in short supply). Show the referrer the status of each person they referred. Add referral asks to project kick-offs, and give campus ambassadors their own codes.

Referred specialists arrive pre-vetted by peers, and paying on the first approved task rewards quality, not sign-ups.

• Referral fraud

  Pay only after identity verification and an approved task.

• Referring into a waitlist

  Open referrals only in domains with work available.

• #8 Show every expert where they stand: application status, review times and payout dates, published as service levels

• HiringGrowth Associate: "grow our expert network from 100K to hundreds of thousands and beyond"; "Build, optimize, and scale automated expert sourcing pipelines"; "Expand AfterQuery's Conference footprint"; "Expand Campus Partnership efforts with elite universities and MBA/PHD programs".Source
• Competitor researchRivals' own homepages: Mercor's top navigation has "Experts" with "Find work" and shows "$90/hr average contracted rate", "$4M+ daily payouts" and "5M+ experts"; micro1 links "Join as an expert", "Refer & earn" and "Interview Prep", plus a data-partnership referral program; Turing has "Join expert network" in its header and "Off-The-Shelf Datasets"; Surge AI's navigation has "Off-the-Shelf", "Expert Workforce" and a list of its public benchmarks.See the research Source
• Expert reviewsAnnotation Academy (a seller of AI-evaluation training) analysed 44 Reddit threads and 327 comments, mostly May-September 2026: AfterQuery is "a real platform" and 16 accounts reported being paid; 11 accounts reported no update for weeks or months after applying, 7 described task reviews taking weeks or months, 12 described paused or cancelled projects, and 6 reported Stripe payout problems in some countries.Source

Add required domain, expert count and start window as mandatory fields at the scoping stage in Attio. Turn the probability-weighted pipeline into a weekly supply forecast by domain. Recruiting campaigns for a domain start when forecast demand exceeds active supply, and waitlisted experts in that domain are reactivated first.

Projects staff faster, experts wait less, and growth spend goes to domains that will actually have work.

• Recruiting for deals that slip

  Weight by stage probability and recruit to a waitlist first.

• HiringHead of Growth (posted July 29, 2026): "Design and execute the strategy for sourcing expert independent contractors across professional domains"; "Forecast supply needs based on customer and project demand"; "Measure expert satisfaction and lead the team in ensuring experts are unblocked and satisfied with their experience".Source
• HiringRevenue hiring: a GTM Operations Lead to "Architect CRM and pipeline hygiene" and "Own analytics & attribution", including "ROI on sales events (conferences, dinners, etc.)", on Attio and Ramp; an Engagement Manager as "primary point of contact for key frontier AI labs" who will "Identify and close upsell, cross-sell, and new use-case opportunities"; an Associate to "Create and maintain sales collateral".Source
• Expert reviewsAnnotation Academy (a seller of AI-evaluation training) analysed 44 Reddit threads and 327 comments, mostly May-September 2026: AfterQuery is "a real platform" and 16 accounts reported being paid; 11 accounts reported no update for weeks or months after applying, 7 described task reviews taking weeks or months, 12 described paused or cancelled projects, and 6 reported Stripe payout problems in some countries.Source

For each target account, have experts write 20 to 30 tasks typical of its workflow (for example, credit memos or contract review). Run current models on them and send a two-page summary of where they break, with an offer to scope a full evaluation. Reuse tasks across accounts in the same industry to keep the cost down.

It shows AfterQuery's method instead of describing it, and it naturally leads to the Evaluate stage of the enterprise offer.

• Generic tasks miss the prospect's reality

  Use public filings and job posts to tailor five of the tasks.

• Your websiteProducts page: 12 offerings, including Off-the-shelf Data ("Request access" goes to the contact form), Loss Analyses, Custom Evals and Professional Domains with "300,000+ verified practitioners"; no prices and no page for any individual dataset.Source
• HiringHead of Enterprise GTM (posted November 9, 2026): "Personally drive outbound prospecting, relationship development, pipeline generation, and strategic account acquisition"; experience "selling complex, high-value solutions to frontier AI labs, Fortune 500 companies"; "Identify the highest-value customer use cases for custom evaluations, data generation, model training".Source
• Your websiteFor Enterprises page: "Services across the full evaluation, training, and deployment lifecycle" (Evaluate, Curate, Post-train, Deploy with forward-deployed engineers); custom benchmarks can be re-run against new models, checkpoints and vendors; four case studies (Legora, NVIDIA, Hebbia, The Raine Group); a "Trusted by" strip whose logos carry no names.Source

For every post, publish a chart and a three-line takeaway from the founder account, a longer LinkedIn version aimed at enterprise buyers, and a thread for researchers. Link back to the post and its buying block. Re-share older results when a related model or benchmark is in the news.

Labs and enterprise buyers follow researchers and founders more than company pages, and a measured result shares well.

• Overstating results

  Every chart links to the method in the post.

• #6 End every research post with the dataset or service behind the result

• HiringTechnical Content Writer (Contract, posted September 21, 2026): "Write and adapt content for AfterQuery's brand and founder social media accounts, including X and LinkedIn", about 10-15 hours a week to start.Source
• Your websiteResearch page: 15 posts from October 2025 to November 8, 2026; public benchmarks SpreadsheetBench 2 and Legora BAR; the newest benchmark, CrossSearch, has only a blog post. The GDPval case study used AfterQuery's "Off-The-Shelf Office Agent Training Dataset" (GDPval 35.3 to 46.7, against a teacher's 49.5) and ends: "Get in touch here to access our off-the-shelf GDPval, Office Agent, and agentic post-training datasets".Source
• HiringIn every job post: "We serve every frontier AI lab with the mission of delivering the best data to power the best models"; "Forbes reported that we could be YC's fastest unicorn, reportedly raising at a $3.2 billion valuation"; backed by Altos Ventures, BoxGroup and Y Combinator. 41 open roles, 40 in San Francisco.Source

Build partnerships with that group: a standard, counsel-approved licensing process for closing companies, a fast valuation from a sample, and a referral fee for advisors (aligned with the referral terms). Publish a page for advisors that explains de-identification and how proceeds flow to the estate. Track each advisor as a channel.

One advisor sees many closures a year, so each partnership is a repeatable source of unique, multi-year professional data that rivals cannot easily reach.

• Consent and ownership questions

  Counsel-approved checklist before any transfer; exclude data the estate cannot license.

• Reputational sensitivity

  Publish the de-identification process and decline unclear cases.

• #3 Put the terms behind "Refer and earn up to $85k" on the referral form itself
• #11 A trust page that explains how AfterQuery handles enterprise and partner data, before security review asks

• Your websiteData Partnerships page: "Your company's data, licensed with care."; operating, winding-down and closed companies are welcome; personal details removed or pseudonymized and a re-identification test before delivery; six steps from inquiry to payment; a "Refer and earn up to $85k" button. The /data-partnerships/refer form gives no amount, eligibility or payment timing, only that referrals are tracked so "payment is ensured if we move forward with the partnership"; that page is not in the sitemap.Source
• HiringData Acquisition Lead: "Personally lead negotiation of AfterQuery's highest-value, most complex enterprise data-purchase agreements" and "set offer prices, deal structures and walk-away points in a market with little public pricing data", ideally deals of "seven figures or more". A Quality Control Lead will "Own quality metrics and KPIs, reporting on error rates, root causes, and improvement trends".Source

Publish a twice-yearly index: a fixed, refreshed task set per professional domain, scored for current frontier and open models, with failure categories explained by the experts who wrote the tasks. Keep most tasks private to avoid contamination, and release a sample. Give each domain a landing page and a "run this on your workflow" call to action that leads into custom evals.

A recurring index makes AfterQuery the reference for professional-grade AI, recruits experts who want their field represented, and creates enterprise demand for custom versions.

• Lab customers dislike public rankings

  Report failure categories, and give labs a preview before publishing.

• Contamination

  Keep most tasks private and rotate them.

• #16 Open enterprise outbound with a free mini loss analysis on the prospect's own workflow

• Your websiteProducts page: 12 offerings, including Off-the-shelf Data ("Request access" goes to the contact form), Loss Analyses, Custom Evals and Professional Domains with "300,000+ verified practitioners"; no prices and no page for any individual dataset.Source
• Your websiteResearch page: 15 posts from October 2025 to November 8, 2026; public benchmarks SpreadsheetBench 2 and Legora BAR; the newest benchmark, CrossSearch, has only a blog post. The GDPval case study used AfterQuery's "Off-The-Shelf Office Agent Training Dataset" (GDPval 35.3 to 46.7, against a teacher's 49.5) and ends: "Get in touch here to access our off-the-shelf GDPval, Office Agent, and agentic post-training datasets".Source
• Competitor researchRivals' own homepages: Mercor's top navigation has "Experts" with "Find work" and shows "$90/hr average contracted rate", "$4M+ daily payouts" and "5M+ experts"; micro1 links "Join as an expert", "Refer & earn" and "Interview Prep", plus a data-partnership referral program; Turing has "Join expert network" in its header and "Off-The-Shelf Datasets"; Surge AI's navigation has "Off-the-Shelf", "Expert Workforce" and a list of its public benchmarks.See the research Source

Package a six-week pilot: week 1-2 a custom eval on one workflow, week 3-4 a targeted dataset, week 5-6 a post-trained model or agent compared against the baseline, ending in a go/no-go report. Publish a price band and a sample report. Convert successful pilots into deployment and re-run contracts.

A fixed scope and price fit enterprise budgets and procurement, and a measured before-and-after result sells the larger engagement.

• Pilots pull capacity from lab projects

  Cap the number of concurrent pilots and staff them separately.

• #10 Industry pages for enterprise buyers in legal, finance and investment banking, each built on a published case study
• #11 A trust page that explains how AfterQuery handles enterprise and partner data, before security review asks

• Your websiteFor Enterprises page: "Services across the full evaluation, training, and deployment lifecycle" (Evaluate, Curate, Post-train, Deploy with forward-deployed engineers); custom benchmarks can be re-run against new models, checkpoints and vendors; four case studies (Legora, NVIDIA, Hebbia, The Raine Group); a "Trusted by" strip whose logos carry no names.Source
• HiringHead of Enterprise GTM (posted November 9, 2026): "Personally drive outbound prospecting, relationship development, pipeline generation, and strategic account acquisition"; experience "selling complex, high-value solutions to frontier AI labs, Fortune 500 companies"; "Identify the highest-value customer use cases for custom evaluations, data generation, model training".Source
• Your websiteSeries A announcement (April 9, 2026): "$30 million Series A at a $300 million valuation, led by Altos Ventures"; "surpassed $100 million in revenue run rate"; "nearly 100,000 verified practicing professionals"; the money goes partly to "expanding our expert network".Source

Show every expert a rolling quality score per domain, the three most common reviewer notes on their recent tasks, and what it takes to reach the next tier (priority matching, higher-paid projects, reviewer roles). Feed the same score into matching, so the best experts see work first.

Visible standards raise quality at the source, cut review load, and give strong experts a reason to stay rather than drift to a rival.

• Experts dispute scores

  Show the tasks and notes behind the score, and offer an appeal.

• #8 Show every expert where they stand: application status, review times and payout dates, published as service levels

• HiringData Acquisition Lead: "Personally lead negotiation of AfterQuery's highest-value, most complex enterprise data-purchase agreements" and "set offer prices, deal structures and walk-away points in a market with little public pricing data", ideally deals of "seven figures or more". A Quality Control Lead will "Own quality metrics and KPIs, reporting on error rates, root causes, and improvement trends".Source
• Expert reviewsAnnotation Academy (a seller of AI-evaluation training) analysed 44 Reddit threads and 327 comments, mostly May-September 2026: AfterQuery is "a real platform" and 16 accounts reported being paid; 11 accounts reported no update for weeks or months after applying, 7 described task reviews taking weeks or months, 12 described paused or cancelled projects, and 6 reported Stripe payout problems in some countries.Source
• HiringMarketing Lead: "own AfterQuery's expert marketplace strategy, acquisition, activation, matching, retention, and quality", "setting the metrics and cadences that track staffing speed, conversion, and quality", and "go-to-market and brand strategy for technical and enterprise audiences". A Software Engineer - Marketplace will "Build end-to-end product experiences across expert acquisition, onboarding, assessments, matching".Source

Find the countries with the most stalled payout set-ups or abandoned onboardings at the payment step. Add a second payout provider for those countries, and show the supported methods by country before an expert applies. Prioritise countries with deep pools of in-demand credentials.

Removing a payout blocker opens supply that is already interested, at a lower cost than sourcing new experts elsewhere.

• Compliance in new jurisdictions

  Start with one or two countries after counsel review.

• Expert reviewsAnnotation Academy (a seller of AI-evaluation training) analysed 44 Reddit threads and 327 comments, mostly May-September 2026: AfterQuery is "a real platform" and 16 accounts reported being paid; 11 accounts reported no update for weeks or months after applying, 7 described task reviews taking weeks or months, 12 described paused or cancelled projects, and 6 reported Stripe payout problems in some countries.Source
• HiringHead of Growth (posted July 29, 2026): "Design and execute the strategy for sourcing expert independent contractors across professional domains"; "Forecast supply needs based on customer and project demand"; "Measure expert satisfaction and lead the team in ensuring experts are unblocked and satisfied with their experience".Source