Aristral Loop · Agentic SEO
Three tiers. One closed loop. SEO that ships while you sleep.
Aristral Loop is a closed-loop agentic SEO system — autonomous agents that crawl, research, draft, judge, schema, publish and re-optimise on a goal you set once. Frontier-LLM-powered, run by a Bristol team, flat-fee monthly. GEO and Local SEO included on every tier.
What agentic SEO actually means
Goal-level instruction. Tool-using agents. No prompt babysitting.
Agentic SEO is the deployment of autonomous LLM-backed agents that plan, execute and iterate across the full SEO lifecycle with goal-level rather than prompt-level instruction. The agent perceives, plans, calls tools, observes the result and adjusts — without a human approving each step.
If you have to log in and click generate, you have AI-assisted SEO. If you set a goal once a quarter and the system runs the seventeen steps from crawl to re-optimisation on its own — with human review only where it’s load-bearing — you have agentic SEO.
| Category | Reasoning loop | Owns execution | Target surface |
|---|---|---|---|
| Automated SEO | None | Yes (deterministic) | Google SERP |
| AI SEO | Per-prompt | Partial | SERP + AI Overviews |
| Programmatic SEO | None (data-driven) | Yes (template render) | Long-tail SERP |
| GEO / AEO | None (content spec) | No — it's a discipline | LLM citations |
| Agentic SEO | Multi-step, persistent, tool-using | Yes (autonomous) | SERP + LLM citations |
Pricing
Three tiers of Aristral Loop
Flat, monthly and inclusive. No per-task overages. No required add-ons. GEO and Local SEO included on every tier.
Solo
Owner-operated businesses, single-site SaaS, solo consultants.
≈ $625/month · ex-VAT
- ✓10 production articles per month
- ✓GEO + Local SEO included
- ✓Monthly GSC + GA4 + AI Overview report
- ✓Publishes to WordPress, Webflow, Sanity or Next.js MDX
- ✓48-hour first-draft turnaround
- ✓Schema validated in CI before push
- ✓Cannibalisation pre-flight on every brief
- ✓Monthly rolling, cancel any time
“Done-for-you agentic SEO, including GEO, for the price of three SaaS subscriptions.”
Start Solo →Sits below Jasper Pro ($69/seat, no SEO loop), Frase Scale ($299 with required add-ons) and Surfer Peace of Mind ($299, tool-only).
Growth
Agencies, multi-location operators, 10–50 person SaaS.
≈ $1,520/month · ex-VAT
- ✓50 production articles per month
- ✓GEO + Local SEO + multi-location support
- ✓Automation bolt-ons: review responder, GBP automation
- ✓Bi-weekly strategy call with the team
- ✓1 custom comparison page per quarter
- ✓Internal-link engine over your full sitemap
- ✓Slack channel for async questions
- ✓Monthly rolling after a 30-day pilot
“Agency-grade output at SaaS pricing. Fixed fee, no overage surprises.”
Start Growth →Sits in the band of AirOps Pro (opaque task-based pricing, no UK delivery), MarketMuse Strategy ($499) and Clearscope Business ($399) — but flat-fee and execution-included.
Enterprise
Multi-market brands, agency resellers, regulated SMB.
≈ $5,050/month · ex-VAT
- ✓300 production articles per month
- ✓Multi-market (UK + UAE + KSA + custom)
- ✓White-label option for agency resellers
- ✓Weekly strategy + dedicated Slack channel
- ✓API access for CMS integrations
- ✓1 custom comparison page per month
- ✓Dedicated evaluation reports, weekly
- ✓3-month minimum, then monthly rolling
“Half the cost of an in-house head of growth, ships before they'd finish onboarding.”
Start Enterprise →A UK head of growth runs £70k+ a year loaded — about £5,833/month. Surfer Enterprise starts at $999 with no execution. We deliver both the strategy and the work.
Why flat-fee, not task-based. Task-based and credit-based pricing in this category is documented to swing monthly bills by three to four times without warning. Flat-fee monthly is a deliberate choice — your finance team can forecast spend, and you stop paying surprise overage invoices for work you didn’t request.
Included on every tier
The features below are not add-ons
They’re the floor. If a tier doesn’t have it, we built it wrong.
GEO (Generative Engine Optimisation)
Every page ships with a definition-first passage under sixty words, FAQ schema mirroring on-page text verbatim, citable statistics, and the schema density that AI engines reward. Our own May 2026 analysis confirms AI Overviews fire on 80% of US and 85% of UK SERPs in this niche — citation strategy is the dominant SERP play, not classical rank tracking.
Local SEO
Google Business Profile schema, NAP consistency audit, LocalBusiness JSON-LD on every location landing page, citation building across UK and US directories, and review-response automation for multi-location operators.
Schema validated in CI
Five schema failure classes account for most rich-result losses at scale. We run the Rich Results Test in continuous integration before any page pushes to your CMS — not in production where a bad markup deployment can wipe rich results across thousands of URLs.
Cannibalisation pre-flight
Before any draft starts, the brief agent runs a semantic similarity check against your live sitemap. If a proposed page overlaps an existing page above a tuned threshold, we re-scope the brief or update the existing page instead.
Real fact-checking
The fact-check agent doesn't trust web search snippets — it runs a real retrieval pass against primary sources, then a higher-tier judge model evaluates the draft against the retrieved evidence. Hallucinated statistics get caught here, not after publication.
Cost guardrails
Every client account has a hard monthly budget cap enforced inside our orchestration layer. When usage hits 80% of budget, the router auto-degrades to lighter-weight processing for non-critical steps and sends an alert.
How Aristral Loop works
Transparency on shape, not on stack
Cost discipline — why your bill stays predictable
Each task class gets the cheapest frontier model that can handle it. Lightweight models run crawl triage, schema generation, link extraction and formatting. Mid-tier frontier models run briefs, drafts, SERP analysis and fact-check setup. The highest-tier judge model only runs on ambiguous SERP-intent calls and schema-at-scale reviews. Documented savings from this routing discipline alone are 40–70% on token spend.
Human-in-the-loop — load-bearing, not theatre
We gate four things on a human: YMYL pre-publish review, brand-voice sign-off on the first five drafts per new client, schema mass-changes, and disavow file commits. Everything else runs autonomously. We cut the theatre gates — per-keyword approval, per-link approval — because they add latency without adding judgement.
Failure modes we guard against
Google core update wipeouts, hallucinated statistics, schema errors at scale, content cannibalisation, indexing refusal, prompt drift on silent model updates, and runaway API costs. Aristral Loop guards each one.
What we don’t do
Who shouldn’t hire us
Buyer’s guides that genuinely arm buyers to evaluate the publisher are rare. Here’s where Aristral Loop is not the right call.
| Gap | Why it matters | Who should look elsewhere |
|---|---|---|
| 15-person team, not a 50-person agency | You get direct access to a co-founder at every tier. That's a feature for some, a ceiling for others. | Enterprises that need a large account team and a formal SLA. Talk to AirOps or a Big Four digital arm. |
| We will not auto-publish YMYL content | Legal, medical and financial drafts always pause for human review. Adds 24–48 hours to those specific pages. | Teams that want fully unattended publishing for regulated verticals. We won't take that engagement. |
| We will not guarantee rankings | Any agency that does is either lying or planning a Webmaster Guidelines violation that will cost you more than the retainer. | Buyers shopping on guarantees. Find a different agency — ideally not one in this category. |
The data this page is built on
Methodology is the moat
Numbers and claims above trace to specific sources, documented here.
Verified keyword and SERP data
In May 2026 we ran a multi-market SERP and keyword analysis across US, UK, Germany and Canada — 47 relevant keywords in the AI SEO category, depth=20 for organic positions, AI Overview presence, People Also Ask questions and Local Pack signals. Headlines: 0 keywords in the “moat” bucket (KD > 45), 8 quick-wins (KD ≤ 20, volume ≥ 30), 10 mid, 32 long-tail. AI Overviews fire on 80% of US and 85% of UK SERPs. Full methodology and raw data available to clients under NDA.
Competitor pricing — fetched live
The pricing comparisons on this page (Jasper, Surfer, AirOps, Frase, MarketMuse, Clearscope, NeuronWriter, Scalenut, Koala, Outranking) were fetched directly from each vendor's pricing page in May 2026. SaaS pricing in this category shifted materially every three to four months in 2025 — we re-confirm before each client kickoff.
Failure-mode evidence
HubSpot lost 70–80% of organic traffic between November 2024 and Q2 2025. Business Insider lost 55% over the same window. The March 2025 Core Update hammered low-quality programmatic SEO; the August 2025 Spam Update explicitly hit programmatic pages and spun content. AI Overviews are factually correct ~9 times in 10 — meaning agentic systems inherit a 10% factual error rate unless the fact-check agent runs real retrieval.
Academic provenance for GEO
Generative Engine Optimisation as a term was coined in the November 2023 paper from Aggarwal, Murahari, Rajpurohit, Kalyan, Narasimhan and Deshpande at Princeton, Georgia Tech, Allen AI and IIT Delhi. The paper introduced GEO-bench (10,000 queries, 9 optimisation methods) and found that adding statistics lifts AI citation visibility by 41%. That finding shapes our content briefs directly — every passage carries citable numbers where the SERP rewards them.
Who’s running this
A fifteen-person team led by three co-founders
New engagements go to a co-founder directly — not a sales pod, not an account-manager funnel.
Taha Bilal
Co-founder
Heads up new client engagements and product. Reachable directly at admin@aristral.com.
Amelia Grace Gascoigne
Co-founder
Co-founder at Aristral, Bristol.
Huzaifa Jan Asim
Co-founder
Co-founder at Aristral, Bristol.
Selected clients include Executive Homes Realty — a Canadian residential real estate brokerage running SEO and AI-powered content infrastructure on Aristral Loop.
Aristral is registered at 172 Gloucester Road, Bristol BS7 8NU and delivers in UK timezone. Aristral Loop targets US and UK primary markets with Germany, Canada and MENA as tertiary coverage. Pricing is GBP-native, with USD-equivalent shown for US buyers.
FAQ
Real questions from real discovery calls
What is agentic SEO?+
Agentic SEO is the use of autonomous LLM-backed agents that plan, execute and iterate across the full SEO lifecycle with goal-level rather than prompt-level instruction. The agent perceives, plans, calls tools, observes the result and adjusts — without a human approving each step.
Is this just AI writing with extra steps?+
No. Aristral Loop runs an end-to-end pipeline: it crawls competitor SERPs, pulls keyword and ranking data, drafts and judges its own output, generates schema, picks internal links from a representation of your existing site, pushes to your CMS, then monitors rankings and AI Overview citations and re-optimises. The writing is one of seventeen steps.
What if a Google core update wipes my rankings?+
March and August 2025 core updates wiped large programmatic plays. We mitigate four ways: similarity pre-flight against your live sitemap, schema validated in CI against the Rich Results Test, IndexNow plus Google Indexing API only where permitted, and a nightly evaluation against a hundred-page golden set that flags regressions before they hit traffic.
Do you guarantee rankings?+
No. Any agency that guarantees rankings is either lying or planning to commit a Webmaster Guidelines violation. We guarantee delivery cadence, evaluation discipline and full reporting access. Rankings are downstream of execution quality and Google's algorithm.
What does 'GEO included' mean?+
Generative Engine Optimisation is the discipline of getting cited inside AI-generated answers (Google AI Overviews, ChatGPT web search, Perplexity, Bing Copilot). We ship every page with definition-first passages under sixty words, FAQ schema, citable statistics and the schema density that AI engines reward — on every tier, not as an add-on.
Who actually writes the content — the AI or a human?+
Both. Drafts are generated by a frontier LLM against a structured brief produced by the research agent. Drafts pass through a higher-tier judge model against a per-client golden set. The first five drafts per new client also go through a human voice-calibration review. After calibration, human review concentrates on YMYL content, brand-sensitive launches and schema mass-changes.
How is this different from Surfer, AirOps or Jasper?+
Surfer and Jasper are content tools — you still operate them. AirOps uses opaque task-based billing with no UK delivery layer. Aristral Loop is done-for-you on a flat monthly fee, ships within seven days of kickoff, includes GEO and Local SEO at every tier, and gives you full reporting access.
Can I cancel any time?+
Yes. Monthly rolling on Solo and Growth. Enterprise has a three-month minimum because of the dedicated golden-set build and the multi-market provisioning, then rolls monthly. Email admin@aristral.com and we wind it down at the end of the current billing month.
Do you handle YMYL content — legal, medical, financial?+
Yes, with a hard human-review gate before publish. YMYL pages run a regulated-language check, a citation-quality check against primary sources, and a final human sign-off. We will not auto-publish YMYL content even at Enterprise tier.
What CMSes do you publish to?+
WordPress (REST API), Webflow (CMS API), Sanity and Next.js MDX via Git commits to your repo. Other CMSes on request — most modern CMSes have an authenticated POST endpoint we can wire in within a day.
Start with a 20-minute call
No pitch deck. We’ll look at your current SEO data together, run a live SERP query against your target market, and tell you which tier fits — or tell you to hire someone else.