Why this rebuild matters
The Flowgear website is not a minor refresh project. By late March 2026, it had become the single most important marketing priority — nothing else in the marketing function was ranked above it. The rebuild carries direct commercial pressure: qualified leads need to improve, and the current site is not delivering them. Lead quality has been described internally as "shocking." The site generates activity, but not the pipeline that matches Flowgear's sales motion.
Part of the problem is product drift. Flowgear has evolved substantially — the V2 runtime, web apps, and MCP/AI integration represent a fundamentally more capable platform than the one the current site describes. Visitors are evaluating an outdated version of the product. That gap is not a copy problem alone; it is a positioning problem.
The second gap is commercial framing. Flowgear sells through a consultative, assessment-led motion — a structured five-step engagement where a Solution Architect evaluates a client's integration estate, baselines maturity, surfaces automation opportunities, and delivers a prioritised roadmap. That is a differentiated commercial offer. The current site does not position it clearly.
This rebuild must fix all four of the site's core jobs at once: clarify a complex platform so both business and technical buyers understand it quickly; support the assessment-led commercial motion; persuade both audience segments (business outcomes for C-level, technical depth for architects); and generate more qualified leads. None are optional.
Tech stack confirmed
The rebuild will be built entirely from source code — custom HTML/CSS/JS or a modern static framework — managed in a shared GitHub repository and deployed automatically on Vercel. No WordPress. No page builder. No Webflow.
- Framework: Next.js or Astro — decision to confirm Day 1 (see Six Decisions section)
- Styling: Tailwind CSS with Flowgear design tokens (turquoise, yellow, dark navy surfaces)
- Deployment: Vercel — zero-config for Next.js and Astro, automatic branch preview deployments for QA review
- Repository: Shared GitHub repo — GE and Flowgear dev team both have access
- CMS: Optional headless CMS (Sanity or Contentful) if Flowgear team needs to update copy without a developer; otherwise static MDX files are sufficient for MVP
Next.js vs Astro — the one decision to confirm
Next.js
- Most familiar to React developers
- Strong ecosystem and community
- Edge functions for dynamic features
- Easy CMS integration
Astro ✦
- Faster static output — lower JS overhead
- Better Core Web Vitals at 60+ page scale
- Framework-agnostic (use React components if needed)
- Purpose-built for content-heavy sites
Sitemap intelligence
The sitemap index at flowgear.net/sitemap_index.xml reveals the full current site structure. This informs both the redirect strategy and which content can be migrated vs built new.
| Sitemap file | What it contains | Rebuild action |
|---|---|---|
| integration-sitemap1–7.xml | 7 files of pairwise "Integrate X with Y" connector pages — high cardinality, thin content per page | 301 all URLs to relevant solution or connector hub pages. Not rebuilding. |
| department-sitemap.xml | All 8 department solution pages | Migrate all — confirmed live with rich content |
| technology-sitemap.xml | All 8 technology solution pages | Migrate all — confirmed live with rich content |
| use-case-sitemap1–2.xml | Use case pages (volume unknown without crawl) | Audit by traffic before deciding: migrate or redirect to solution framework pages |
| customer-story-sitemap.xml | Customer story pages — already exist on live site | Migrate top 3–5 stories; hub page is new |
| partner-sitemap.xml | Partner directory pages | Migrate or rebuild as new Partners hub |
| connectors-sitemap1–2.xml | Individual connector pages | Hub approach for MVP; individual connector detail pages Phase 2 |
| post-sitemap.xml | Blog posts | Classify by GSC traffic: migrate high-traffic, 301 or 410 the rest. Needs GSC access. |
| career-sitemap.xml | Careers pages | Migrate or redirect to About / Company page |
URL migration categories
Every URL on the live site falls into one of three categories. Classifying them requires GA4 and GSC access — do not cull anything without traffic data.
- Keep and 301 redirect: High-traffic posts aligned with MVP pillars. 301 to the closest equivalent page in the rebuilt site. These are the URLs that protect organic traffic through the transition.
- Consolidate: Thin or overlapping content covering the same topic. Merge into the appropriate pillar page. 301 all old URLs to the merged destination.
- Cull with 410: Outdated announcements, thin content with no backlinks and no measurable traffic. Return 410 (Gone) — this signals intentional removal to crawlers rather than a broken link.
Action required before launch: GA4 and GSC access to classify all existing URLs. Build the redirect map during Day 1–2 alongside the sitemap crawl. This is not optional — an unplanned launch creates 404s that damage the crawl health of the new site from day one.
Solutions: all 25 pages confirmed as MVP
The Flowgear website uses three frameworks for solutions pages — Department, Technology, and Application. All 25 pages across these three frameworks are included in the MVP. 24 of 25 exist on the live site. Most need Benefits and FAQ sections written — every live solution page currently has lorem ipsum in those sections.
By Department
8 pages| Page | Status | Build task |
|---|---|---|
| Finance | Live · Rich | Migrate + write Benefits & FAQ |
| Sales | Live · Rich | Migrate + write Benefits & FAQ |
| eCommerce | Live · Rich | Migrate + write Benefits & FAQ |
| Warehouse Management | Live · Rich | Migrate + write Benefits & FAQ |
| Human Resources | Live · Rich | Migrate + write Benefits & FAQ |
| Marketing | Live · Rich | Migrate + write Benefits & FAQ |
| Customer Support | Live · Rich | Migrate + write Benefits & FAQ |
| Manufacturing | Live · Rich | Migrate + write Benefits & FAQ |
By Technology
8 pages| Page | Status | Build task |
|---|---|---|
| ERP | Live · Rich | Migrate + write Benefits & FAQ |
| CRM | Live · Rich | Migrate + write Benefits & FAQ |
| eCommerce | Live · Rich | Migrate + write Benefits & FAQ |
| WMS | Live · Rich | Migrate + Benefits & FAQ; simplify URL + 301 |
| Payroll | Live · Rich | Migrate + write Benefits & FAQ |
| Logistics | Live · Rich | Migrate + write Benefits & FAQ |
| Customer Experience | Live · Rich | Migrate + write Benefits & FAQ |
| Data Storage | Live · Rich | Migrate + write Benefits & FAQ |
By Application
9 pages| Page | Status | Build task |
|---|---|---|
| Oracle NetSuite | Live · Thin | Enrich use cases + write Benefits & FAQ |
| Salesforce | Live · Rich | Migrate + write Benefits & FAQ |
| SAP S/4HANA | Live · Rich | Migrate + write Benefits & FAQ |
| Acumatica | 404 · Build New | Create from scratch — content source needed |
| Microsoft Dynamics 365 CRM | Live · Rich | Migrate + write Benefits & FAQ |
| Sage (Intacct) | Live · Rich | Migrate + write Benefits & FAQ |
| Shopify | Live · Rich | Migrate + write Benefits & FAQ |
| HubSpot | Live · Thin | Enrich use cases + write Benefits & FAQ |
| Microsoft Azure | Live · Thin | Enrich use cases + write Benefits & FAQ |
Full MVP scope
Work already in progress
The rebuild is not starting from zero. Substantial design and copy work has already been completed for the highest-priority pages. Four full-page mockups exist and are live for review.
What the mockups establish
- Homepage A: Assessment-led commercial motion — Integration Assessment as the primary CTA, platform proof as the supporting layer. Dual audience architecture with distinct hooks for C-level and technical evaluators.
- Homepage B: Platform-led — leads with product depth and capability before moving to the assessment offer.
- Pricing A: Consultative-led — frames pricing in the context of the assessment model. Builds commercial context before showing numbers.
- Pricing B: Price-forward — leads with the tiers and numbers directly, for buyers who already know what they want.
Flowgear marketing should select a preferred direction from each pair. Once one homepage variant and one pricing variant are selected, the build agent uses those as the design baseline for all remaining pages.
Pairwise pages + the connector tool concept
The live site carries significant URL volume from pairwise "Integrate X with Y" connector pages — confirmed as 7 separate sitemap files. The MVP scope deliberately does not replicate this pattern.
Why not rebuild them: Thin and near-duplicate content at this scale dilutes crawl budget from money pages. AI search (GEO) rewards verifiable depth over broad thin coverage. A site with 57 strong pages outperforms one with 500 thin ones in both traditional and AI search. All pairwise URLs will 301 redirect to the most relevant solution or connector hub page.
Phase 2/3: interactive connector pairing tool
Rather than rebuilding hundreds of static pairwise pages, we're proposing an interactive drag-and-drop connector pairing tool — embedded in the Connectors hub — as the Phase 2/3 replacement. Users select two connectors; the tool dynamically surfaces relevant use cases, integration patterns, and links to related solution pages.
This replaces hundreds of static thin pages with one intelligent, maintainable tool. It also creates a more compelling product-adjacent UX that reflects how Flowgear actually works — visually, not just descriptively.
Comparison and alternative pages
Recommend one comparison hub page — "How Flowgear Compares" — rather than individual versus-competitor pages at MVP. One well-structured page with FAQPage schema is easier to rank and maintain than ten thin pages. Post-MVP: add individual /flowgear-vs-[competitor]/ pages for competitors with proven search volume — likely Zapier, Workato, and Boomi / SAP Integration Suite.
2-week AI-assisted build plan
With agentic AI handling all build, copy, and schema work simultaneously, the critical path is decision speed and review throughput — not production time. AI builds all 57–62 pages in parallel. Humans review in priority order. The bottleneck is not the build; it is Day 1–2.
All of the following must be confirmed before a single page is generated. Every day this slips, the go-live date slips by the same amount.
- Framework: Next.js or Astro — GE can recommend; Flowgear dev team must confirm for long-term maintenance comfort
- CTA hierarchy per page type — Trial, Assessment booking, or Sales call: one primary per page type, confirmed by Pashnee
- Acumatica content source — who provides content or approves AI-generated use cases?
- Customer stories cleared for web — how many? Can GE access the customer success PDF now?
- Redirect mapping sign-off — 7 integration sitemaps need a 301 destination per URL pattern; GE runs Screaming Frog crawl on Day 1 (30 min)
All pages generated simultaneously across three batches. The AI agent:
- Migrates and reformats existing rich use case content into new design templates
- Writes all Benefits sections using the existing use cases as evidence
- Writes all FAQ sections from stakeholder direction and product research
- Generates
FAQPage,Organization, andSoftwareApplicationJSON-LD schema on all relevant pages - Sets all meta titles (under 60 chars) and descriptions (under 160 chars)
- Maps all 301 redirects from old URLs to new destination URLs
Homepage, Pricing, Integration Assessment, Platform hub · V2 Runtime, Web Apps, AI and MCP, Hybrid/Deployment · Infrastructure (Free Trial, Book Assessment, Contact, Privacy, Terms) · About · Customer Stories hub
All 8 Department pages (Finance, Sales, eCommerce, Warehouse Management, HR, Marketing, Customer Support, Manufacturing) · All 8 Technology pages (ERP, CRM, eCommerce, WMS, Payroll, Logistics, Customer Experience, Data Storage)
All 9 Application pages (5 rich migrate, 3 enrich, 1 new) · Connectors hub · Trust and Security · Blog hub · Partners hub · Individual customer stories (top 3–5)
Estimated 8–12 minutes per page for a clean AI build. 60 pages across 4 review sessions of ~3 hours each.
- Per-page QA checklist (AI pre-validates; human confirms): no lorem ipsum · CTA matches agreed hierarchy · no unverified claims or fabricated metrics · meta title/description within character limits · FAQ schema present · mobile layout correct at 375px · 301 redirect entry present for all migrated URLs
- Day 8 AM: Batch A — core commercial, highest priority, most scrutiny
- Day 9 AM: Batch B part 1 — all 8 Department pages
- Day 9 PM: Batch B part 2 — all 8 Technology pages
- Day 10–11: Batch C — all Application pages + supporting pages
- Session 1 (Day 12, 90 min): Core commercial pages — Homepage, Pricing, Integration Assessment, Platform, About, CTA hierarchy across the site
- Session 2 (Day 13, 90 min): Solutions pages — 3–4 pages from each framework; confirm voice, accuracy of use cases, and Benefits copy
- All revision requests resolved same day. Buffer: amends from Flowgear that require AI rebuild are completed by end of Day 13.
Morning: Redirect validation (spot-check 20+ old URLs) · Schema validation in Google Rich Results Test (5–10 pages) · Cross-browser check (Chrome, Safari, Firefox; mobile and desktop) · GA4 conversion events firing · GSC submit for indexing
Afternoon: Go-live. Monitoring active for 48 hours post-launch.
Risk register
Five known risks, all manageable with early action. The single most important mitigation is booking Day 12–13 Flowgear sign-off sessions on Day 1 — not Day 11.
| Risk | Likely delay | Mitigation |
|---|---|---|
| Framework decision not locked Day 1–2 | 1 day per day of slip | Frame as the single hardest dependency in the kickoff. GE comes with a recommendation. Flowgear dev confirms same day. |
| Flowgear sign-off sessions not pre-booked | +2–3 days | Book Day 12–13 calendar slots on Day 1 — not Day 11. Non-negotiable. |
| Acumatica content not provided by Flowgear | +0.5 days | AI generates from public Acumatica documentation; Flowgear confirms accuracy in the Day 13 review session. Page is not a launch blocker. |
| 301 redirect mapping reveals unexpected URL count | +0.5–1 day | Screaming Frog crawl on Day 1 gives exact count. Add to Day 14 morning buffer if needed. |
| Customer stories not cleared for web by launch | Hub launches as shell | Customer Stories hub launches with 2–3 approved quotes as placeholder. Full story migration in Week 3 post-launch. Does not delay Day 14. |
Visual design direction
The rebuild must feel like a product, not a marketing website. The visual language should be adjacent to the V2 console — dark, operational, and precise. Reference benchmarks: n8n (dark mode and product-adjacent UX), Workato (show-don't-tell, looping screen captures, motion to show integration as live and working).
Core principles
- Dark mode first. Deep navy backgrounds — not a light site with a dark variant. Visitors should feel they are entering the product's world, not a separate marketing artefact.
- Show, don't tell. Looping UI captures, annotated screenshots, and interactive elements replace text-only explanations. Integration is shown as live, working, and measurable — not as a cloud diagram with arrows.
- Operational, not abstract. Every diagram and illustration must show real product behaviour. No stock illustrations. No generic iStock workflow arrows.
Colour tokens
Key page patterns
Hero section: Large dense headline — outcome-led, not feature-led. A looping or animated product capture directly in the hero (V2 console, workflow canvas, or MCP in action). Trust strip immediately below fold: logo carousel of known customers and connectors. The four existing homepage and pricing mockups demonstrate this pattern.
Solutions pages (Department / Technology / Application): Full-width dark header band with turquoise left-border accent on the page title. Use case cards in a grid with framework-coloured left borders (Finance = turquoise accent, Operations = blue). An "Integrate with" panel at the bottom of each page — connector logo grid linking to the connectors hub. Interactive tab or filter on solutions hub pages — switch between Department / Technology / Application view.
Connectors hub: Searchable, filterable grid of all connectors. Category filters matching the Technology framework (ERP, CRM, WMS, etc.). Future home of the drag-and-drop connector pairing tool in Phase 2.
Visual elements that differentiate from a generic iPaaS site
- Inline workflow diagrams built in SVG/CSS — not stock illustrations. Animated data flow lines between system icons (subtle pulse, as shown in the connector tool concept above).
- Real metric callouts from customer stories — not fabricated statistics. Named customers with company, role, outcome, and metric in a citation-ready format.
- Code window mockup for V2 Runtime and developer-facing sections — shows a real config or workflow snippet, not a screenshot of empty IDE.
- Assessment progress stepper on the Integration Assessment page — the five-step model (Discover, Assess, Identify, Shape and Prioritise, Deliver) rendered as an interactive stepped component. This pattern already exists in the homepage mockups.
Design system foundation already built
The four existing mockups (Homepage A, Homepage B, Pricing A, Pricing B) establish the complete design language: token set, typography scale, component patterns, grid system, and dark mode conventions. The build agent references these as the visual baseline for all remaining pages — no design system work is required before the Day 3 build begins.
Responsive breakpoints: 375px (mobile minimum) · 768px (tablet) · 1280px (desktop) · 1440px (wide)
Typography: Inter across all weights (400, 500, 600, 700) — already loaded in all mockups via Google Fonts CDN.
Icons: Lucide icon set via CDN — consistent stroke weight, no external image dependencies.
Why GEO matters for Flowgear
AI assistants — ChatGPT, Perplexity, Gemini, Copilot — are increasingly used for platform evaluation queries: "what integration platform should I use for SAP and Shopify?", "best iPaaS for mid-market ERP integration", "Zapier alternatives for enterprise." These queries return generated answers, not a list of blue links. Flowgear should appear in those answers.
GEO (Generative Engine Optimisation) requires the site to be structured so that AI models can extract, trust, and cite it: answer-first content structure, FAQPage schema on every relevant page, entity consistency (consistent brand name, product names, and customer names across all pages), citation-ready proof (real company names and metrics), and freshness signals.
Flowgear already has strong GEO ingredients
- Real customer outcomes with company names and scale (Coricraft, 70 locations; others in the case study bank).
- The five-step assessment model is a named, distinctive process — exactly the kind of structured differentiator AI models cite when recommending vendors.
- Named technical differentiators: V2 runtime, MCP Server, lazy evaluation. These are citable specifics, not generic claims.
- The iPaaS category is well-represented in AI training data — Flowgear needs to be the answer for mid-market and enterprise in the African and emerging-market context.
MVP GEO actions
FAQPageschema on Homepage, Pricing, Platform, and all 25 solution pages — minimum viable structured data.OrganizationandSoftwareApplicationschema on Homepage with accurate product description, founding info, and service area.- Customer proof in structured, citation-ready format: name, company, role, outcome, metric. Do not bury this in long paragraphs — make it extractable.
- Consistent entity naming across all pages: always "Flowgear" (not "flowgear" or "FLOWGEAR"), always "V2 Runtime" (not "v2" or "new runtime").
How do you know the MVP worked?
Define success criteria before launch — not after. These are the five signals that tell you whether the rebuild improved the commercial function of the site.
Percentage of trial signups and assessment bookings from target ICP — measured against current baseline. This is the primary metric. "Better leads" means this number goes up.
Maintain or grow impressions and clicks for commercial intent queries within 90 days of launch. A drop in the first 30 days signals redirect or indexing errors — investigate immediately.
Assessment bookings and trial signups as a percentage of organic sessions. Set a pre-launch baseline from GA4 so post-launch comparison is credible.
Flowgear cited in AI answers for two to three target queries within 60 days of launch. Track manually in ChatGPT and Perplexity monthly.
Zero 404s on launched URLs. All deprecated URLs returning 301 or 410. Verify in GSC within the first 14 days post-launch. A clean redirect map is a prerequisite for any of the other metrics to be reliable.
Six decisions before build can start
A build schedule cannot exist until these six are locked. They are not sequential — all six can be resolved in a single focused 90-minute stakeholder meeting on Day 1. Every day they are not locked, the Day 14 go-live slips by the same amount.
Framework: Next.js or Astro
GE recommends Astro for 57–62 pages at this scale — better Core Web Vitals output, less JS overhead. Next.js is valid if Flowgear's dev team is more comfortable with React for long-term maintenance. This must be confirmed before the build scaffolding is created.
Owner: Flowgear dev + GECTA hierarchy per page type
Which primary action per audience segment? Trial, Assessment booking, or Sales call — one per page type, confirmed. Without this, every page is designed to ambiguous intent and conversion tracking is meaningless.
Owner: Pashnee / MarketingAcumatica: content source
The one application page that does not exist on the live site. Does Flowgear want this page at launch? If yes: who provides content, or who approves AI-generated use cases derived from public Acumatica documentation?
Owner: FlowgearCustomer stories: how many are cleared?
Customer story pages exist on the live site and are confirmed in the sitemap. How many are cleared for migration to the rebuilt site? Can GE access the customer success PDF now? If none are cleared by Day 14, the hub launches as a placeholder.
Owner: Christine / PashneePartners hub scope
Partner pages exist on the current site. Should the rebuilt Partners hub replicate the current partner directory, or be rebuilt with different positioning? This affects whether the partner page is a Batch A or Batch C build.
Owner: FlowgearBlog carryover: GSC access needed
Which existing blog posts are high-traffic and should be migrated vs 301'd vs culled with 410? This requires GSC access. Without it, the blog is treated as "migrate all" as a safe default — which may bring thin content into the new site.
Owner: GE + GA4/GSC access from FlowgearOnce these six are confirmed, a delivery plan follows: page-level scope, owners, dates, and a QA gate before launch. The copy and design work for homepage and pricing is already substantially complete — four variants are live and ready for selection. The rebuild is not starting from zero.
Ready to move forward?
Share this plan with your team or save it as a PDF. Once the six decisions are locked, reply and we'll turn this into a delivery schedule with page-level owners and dates.
Print / Save as PDF