Content Strategy & Go-to-Market Positioning
Built for the 14,000: Launching an AI Legal Associate for South Africa's Underserved Law Firms
Project description
Lexora AI is a, LegalTech startup: an AI-powered legal assistant combining legal research, document drafting, case summarisation, contract review, and compliance support in a single platform, built specifically for advocates, sole legal practitioners, and small law firms in South Africa. I wrote the campaign strategy, content plan, and performance model in six phases, one problem to solve.
THE PROBLEM
On the surface South Africa's legal AI market looks like it has no room left. With LexisNexis and Juta serving the country's largest firms and academic institutions. Consumer tools like Legal Interact's My AI Lawyer serve the public directly. The obvious conclusion is that the category is either covered or irrelevant to anyone outside the Big Six law firms. This conclusion is wrong, and the below listed research gaps reveal why.
FIVE RESEARCHED GAPS
THE STRATEGY
Before writing any content, I built the strategic architecture the rest of the campaign would be built on. Four decisions defined everything that followed.
Strategic Campaign Decisions
Position around the underserved middle, not the whole market
"Built for the 14,000, not the 21" became the single sentence every downstream decision had to serve, messaging, keyword priority, even which case studies to feature.
GEO as a core pillar, not a keyword afterthought
Generative Engine Optimisation was built into the campaign objectives, the content architecture, and the KPI framework from the start, not bolted onto one content cluster after the fact. Every pillar article was built with a quick-answer block, structured extractable content, and schema markup as a production standard, not an optional pass.
Expand the research past the brief
The original brief targeted attorneys. I conducted independent research into how the South African legal industry is structured surfaced a second, sharper opportunity: advocates, who are legally barred from forming firms and are consequently talked past by every existing competitor. I added this as a documented expansion track, not a replacement for the core campaign.
Build one connected campaign, not six deliverables
Every phase of this project was built to transition naturally into the next: personas built in Phase 1 drove the content plan in Phase 2, which drove the SEO brief in Phase 3, the social calendar in Phase 4, the paid media plan in Phase 6, and the KPI framework measured in Phase 5 traced back to the SMART targets set in Phase 1.
Framework 1: Research & Positioning Architecture
Built around three personas representing structurally distinct buyers: a sole practitioner attorney, a small-firm managing partner, and, a practising advocate. Each persona was built from real South African market data: Legal Practice Council firm-size statistics, Law Society of South Africa practitioner counts, and documented technology-adoption barriers, not invented assumptions.
Campaign SMART targets set at the strategy stage:
Framework 2: GEO-Integrated Content Architecture
Every pillar article was built to a shared structural standard: a 150–200 word quick-answer block extractable by AI systems, structured lists and comparison tables instead of unstructured prose, a prompt-aligned FAQ section, and Article/FAQPage schema markup at the technical level — applied as a production requirement across every piece, not a specialist add-on for a
Framework 3: Full-Funnel Campaign Design
The campaign was mapped end-to-end across a 12-week core programme plus a pre-launch and launch-moment phase: content and SEO/GEO built the organic engine, social media carried distribution and trust-building, paid media accelerated qualified traffic at each funnel stage, and every channel fed one shared measurement framework.
EXECUTION
Content Strategy & Market Research
Before any content existed, I built the research foundation: South African legal-AI market analysis, a six-player competitive matrix, three audience personas, a seven-cluster keyword architecture, and the GEO strategy framework, compiled into a 16-tab strategic model.
Deliverables produced:
Market and competitive analysis grounded in verified, named sources (no fabricated statistics)
Three personas: sole practitioner attorney, small-firm managing partner, practising advocate
Seven-cluster keyword and search-intent architecture
GEO strategy integrated from objectives through measurement
Full risk register and source appendix
Editorial & Content Planning
With positioning locked, I built a 55-item content inventory spanning LinkedIn, paid search, SEO blogs, landing pages, and GEO posts, then restructured it around a genuine product-launch arc (pre-launch anticipation, a concentrated launch moment, then a 12-week always-on programme), rather than a flat, undifferentiated content calendar.
Deliverables produced:
55-item master content inventory, categorised by phase, format, funnel objective, and target persona
Full dated editorial calendar: pre-launch teaser sequence, 48–72 hour launch-moment plan, and 12-week post-launch programme
Three fully developed content briefs (LinkedIn, landing page, long-form SEO/GEO blog) with headline options, structure, CTA, and GEO tactic specified for each
A dedicated asset and visual brief mapping every piece to its required visual, format, and specs
SEO & GEO On-Page Optimisation
I selected the campaign's actual Week 1 pillar article, "Why Case Law Research Still Eats Your Week", as the subject for full on-page optimisation, rather than a generic or hypothetical page, so every recommendation ties back to a real piece already in the content plan.
Deliverables produced:
Primary and secondary keyword research validated against South African market sizing, with all estimates explicitly flagged for validation against a live keyword tool
Optimised meta title (multiple A/B-test variants), meta description, and URL structure
Full header hierarchy and content architecture built on an answer-first model
A written quick-answer block, structured for direct AI-system extraction
JSON-LD schema markup specification, including Article, FAQPage, and E-E-A-T implementation
A technical SEO audit framework covering Core Web Vitals, mobile-first standards, and crawlability
Social Media Campaign
Built as a standalone, platform-specific campaign layered on top of the editorial calendar, with its own audience segmentation, content pillars, visual asset briefs, and engagement tactics tailored to each persona.
Deliverables produced:
Platform strategy and audience segmentation by persona
Content pillar framework and a dedicated social content calendar
Individual post briefs with format, copy direction, and visual specification
Paid social strategy and hashtag framework
Community management guidelines
Integrated Advertising Strategy
A full paid media plan layered across Google Search, LinkedIn, Display, and video, engineered to work with the organic content engine rather than duplicate it, with competitor-conquesting campaigns, remarketing sequences, and a funnel-stage budget model.
Deliverables produced:
Channel architecture across Google Search Ads, LinkedIn Sponsored Content, Google Display Network, and video
Full creative concepts mapped to top-, middle-, and bottom-of-funnel stages
Budget allocation by channel and funnel stage, with CPC/CTR/CVR/ROAS targets
Audience targeting strategy, including LinkedIn job-title and company targeting
A dedicated competitive-conquesting strategy against LexisNexis, Jutastat, and Clio
Full attribution and UTM tracking setup, including multi-touch attribution modelling
Performance Reporting & Analysis
I built a full KPI framework tied directly back to the Phase 1 SMART targets, then simulated 12 weeks of campaign performance against realistic industry benchmarks, and wrote the resulting analysis exactly as I would for a real retrospective, including a "what didn't work" section, not just a highlight reel.
RESULTS
Simulated data. The figures below were generated using a documented simulation methodology benchmarked against real SaaS and legal-marketing industry standards (see the full Performance Report for methodology). They demonstrate how I structure, interpret, and act on campaign data, not actual campaign performance.
Organic search and LinkedIn distribution outperformed expectations; the mid-funnel visitor-to-demo conversion underperformed, pointing to email nurture and landing page optimisation as the clearest priorities for a follow-on campaign.
A few specific findings worth surfacing on their own:
The POPIA compliance checklist was the campaign's highest-performing single asset, an 8.2% gated-download conversion rate, well above the 5–15% benchmark, confirming that compliance anxiety is a genuine, high-intent lead driver in the South African legal market, not just a defensive content topic.
The "14,242 vs 21" stat card was the top-performing social post by reach, the specific, named statistic outperformed every general-messaging post in the campaign, reinforcing that the positioning research itself, presented plainly, is stronger marketing than any creative treatment layered on top of it.
The advocate-specific quote card generated the highest share count of any post in the campaign, direct evidence that the advocate research track (Section 3, Decision 3) wasn't just analytically interesting; it produced the single most shared piece of content in the plan.
GEO citations reached 2 of the 5+ target by Week 12, below target, but establishing a real baseline; GEO is documented in the research as a compounding channel over 6–12 months, and the simulation was built to reflect that honestly rather than overstate early traction.
IMPACT
Three things this project demonstrates:
Full-funnel operating capability:
Strategy, content, SEO/GEO, social, paid media, and performance analysis, built as one traceable system, where every downstream deliverable pulls directly from the research and personas built in Phase 1, rather than six disconnected exercises.Original strategic thinking beyond the brief:
The advocates research wasn't asked for. I found it by studying how the South African legal profession is actually structured, and it became one of the project's two strongest results (the highest-share-count social post in the simulated campaign). Good strategy work finds the brief's blind spots, not just its stated requirements.Discipline under a verification-first standard:
Every statistic in the research phases is sourced and named. Every simulated figure in the performance phase is clearly labelled as simulated. That distinction, knowing exactly which numbers in your own work are real and which are illustrative, is a discipline, not a detail.
WHAT I LEARNED
Running a full campaign solo, end to end, surfaces things that working on a single channel doesn't.
Trust content outperforms growth-hack content. The two best-performing assets in the simulation, the POPIA checklist and the advocate quote card, both worked because they were specific and evidence-based, not because they used a clever format. Specificity is the growth hack.
GEO compounds; it doesn't spike. Two citations against a target of five, by design, isn't a failure, it's an honest baseline for a channel that the research itself describes as a 6–12 month compounding play. Reporting a leading indicator honestly is more useful than dressing it up as a finished result.
A campaign is only as strong as its weakest handoff. The funnel's soft point wasn't any single channel, organic and social both beat expectations, it was the mid-funnel handoff from visitor to demo booking. That's an integration problem between content, landing pages, and email nurture, not a channel-performance problem, and it's the clearest argument for why full-funnel visibility matters more than channel-level optimisation alone.