Case studies/IES Limited
CASE 02Compliance-first content engine
4 brands→1 engine

Four regulated insurance brands, three regions, one content pipeline, and no external agencies in 2026.

IES Limited sells travel cover for people with pre-existing medical conditions, where one wrong word or one wrong statistic is a legal problem. Ovidius built an 11-stage pipeline of AI agents. For every brand, it researches, writes and checks compliance before a person reads the draft.

One trigger, every brandLive
Brand AUK
Brand BUK
Brand CAustralia
Brand DIreland
11stages, one pipeline
Brand voice, regional rules and SEO strategy assembled per brand
Compliance checked firstBuilt in n8nRuns inside Notion
Client
IES LimitedTravel insurance group
Scope
Four travel insurance brandsUK, Ireland and Australia
Problem
Slow, costly agenciesStrict compliance on every word
Built with
n8n, Notion, OpenRouterDataForSEO, Trustpilot API
Output
3 content types100+ topics generated
4→1
Brands, one system

Four brands and three regions on a single pipeline.

0
Agencies in 2026

External agencies and freelancers phased out, in the client's words.

11
Stages automated

From research to the compliance check, started by a single trigger.

100+
Topics generated

Across three content types.

01
The challenge

Agencies were too slow, and insurance copy cannot afford a wrong word.

IES needed weekly content across four brands, each with its own voice, its own regional law and its own SEO plan. External agencies and freelancers could not keep up. Every draft still had to clear banned terms, exact statistics and approved phrasing before it could go live.

Speed

Slow agency output

Agencies and freelancers could not keep up with weekly updates across four brands.

Risk

Regulatory compliance

Every piece had to avoid banned terms, get each statistic exactly right and use approved phrasing.

Complexity

Four brands, three regions

Each brand has its own tone of voice, its own UK, Australian or Irish rules, and localised SEO.

Scale

No engine to grow with

With manual work, quality was uneven, and publishing more meant paying more in proportion.

02
The solution

AI agents do the research, the writing and the compliance check, all from one trigger.

The pipeline runs in n8n and lives inside the Notion workspace the marketing team already used. Six of its eleven stages are below. A person sees the draft only after the compliance agent has passed it.

Researcher

Research agent

Search data from DataForSEO, customer reviews from Trustpilot.

Planner

Topic generation

Topics chosen per brand and region.

Planner

Brief assembly

Brand context pulled from Notion into each brief.

Writer

Content drafting

Drafts in the brand's own voice.

Compliance officer

Compliance checker

Checks every draft against that brand's rules.

The IES team

Human review

Reviews a draft that has already passed.

11 stages · 1 trigger · 3 content typesn8n · Notion · OpenRouter · DataForSEO · Trustpilot API
GOAL 01

Replace external agencies

Bring content production in-house with AI.

GOAL 02

Automate end to end

From research and briefs through drafting to compliance review.

GOAL 03

One system, four brands

Every brand from a single content engine.

GOAL 04

Compliance first

Every draft is checked before a person reviews it.

03
Deep dive

The compliance checker that turned a sceptic into a believer.

The agent checks every draft against the brand's rulebook. It must answer in a fixed JSON format: correct, or improvement needed, with each issue named. It cannot wave a draft through with a vague approval. Run it on the draft below.

Draft, Brand AAwaiting check

Covid-19

98% of claims paid

Checker outputJSON
// waiting for a draft

The two issues and the output format are from the project. Draft text is greyed out; the brand is withheld.

Banned termsProhibited language is flagged and replaced with the brand's approved phrasing.
Statistic verificationExact numbers such as claims-paid rates are checked against approved sources.
Forced structureThe answer must be correct or improvement_needed, so the agent cannot make up a vague approval.
Rules per brandUK, Australian and Irish requirements each have their own ruleset.
04
Deep dive

One system that rebuilds its prompt for each brand.

Each run looks up that brand's context in the Notion knowledge base. It then builds the prompt, the instructions the AI model writes from, out of four layers. Pick a brand to see what changes.

Dynamic prompt assembly

Brand infoMission, USP and target audience for Brand A
Tone of voiceThe style guide for Brand A
Compliance rulesLegal requirements for UK insurance copy
SEO strategyKeywords, competitors and targets in the UK market
Assembled at run time
brand: Brand A · region: UK
voice: Brand A style guide
rules: UK compliance ruleset
seo: Brand A keyword strategy
05
The results

From sceptic to believer, and no external agencies in 2026.

Manual effort

Targeted reduction

80%

less production workload

Brands

Three regions

4

on one unified system

Topics

Generated

100+

across 3 content types

Agencies

Needed in 2026

0

phased out entirely

Working with Owen and the Ovidius team has been a seamless experience; they are incredibly responsive and adapted quickly to our specific requirements. By leveraging their AI-driven content for our travel insurance sites, we've completely phased out our reliance on external agencies and freelancers in 2026 while significantly speeding up production.
Bhavek RughaniBhavek RughaniHead of Marketing, IES Limited
Stackn8nNotionOpenRouterDataForSEOTrustpilot API

Where do your rules slow your content down?

Bring the rulebook your writers check by hand. In a discovery call we work out which checks an agent can run first and what that changes for your team.

Book a Discovery Call30 minutes with a senior partner. You bring the process.