Skip to content
AllFixAI

Work

Engagements, described honestly

We publish engagement archetypes rather than client logos. Most of our work sits under NDA, and we will not invent results to fill a page. What follows is the real shape of the problems we are hired to solve, the approach we take, and what each engagement includes.

WebSEOCloudflare

Marketing site replatform

B2B services company with a decade-old CMS nobody wanted to touch.

The problem

The marketing team could not publish without a developer, page speed was poor on mobile, and a legacy plugin surface made the site a standing security concern.

The approach

  1. 01Audited every URL and mapped old paths to a new information architecture
  2. 02Modelled content as typed collections so pages are assembled from reusable sections
  3. 03Built the front-end as a component library with a documented design system
  4. 04Moved hosting to Cloudflare Workers with caching and image optimisation at the edge
  5. 05Rebuilt conversion paths around a shorter, better-validated enquiry form
Scope
Design system, 12 page templates, CMS configuration, migration, analytics, launch support.
Team
1 designer, 2 engineers, 1 SEO lead
Timeline
8–10 weeks
Discuss a similar problem
SoftwareIntegration

Internal operations platform

Operations team running a multi-step fulfilment process in spreadsheets.

The problem

Every order was re-keyed three times, ownership of each step was unclear, and reporting required manual consolidation at month end.

The approach

  1. 01Mapped the real workflow on a whiteboard before writing any code
  2. 02Built a single operational console with role-based approvals and a full action log
  3. 03Integrated the CRM and accounting tool so records reconcile automatically
  4. 04Added an exception queue so failures are visible instead of silent
  5. 05Delivered reporting directly from the operational database
Scope
Platform build, two integrations, data migration, training, 90-day support.
Team
2 engineers, 1 product lead
Timeline
12–16 weeks
Discuss a similar problem
AI & Automation

AI document workflow

Back-office team manually reading and re-typing structured documents.

The problem

High document volume created a persistent backlog, and transcription errors propagated into downstream systems.

The approach

  1. 01Measured the baseline: documents per day, minutes per document, current error rate
  2. 02Built extraction against a strict output schema with validation rules
  3. 03Added an evaluation set and regression tests so prompt changes are measurable
  4. 04Shipped a review interface so a human approves low-confidence results
  5. 05Set cost and quality alerts before widening rollout
Scope
Workflow automation, extraction pipeline, review UI, evaluation harness, runbook.
Team
1 AI engineer, 1 engineer, 1 process owner (client side)
Timeline
6–9 weeks
Discuss a similar problem
SEOWeb

Search recovery and rebuild

Company whose traffic collapsed after a rushed replatform by a previous vendor.

The problem

Thousands of URLs were dropped without redirects, canonical logic was contradictory, and templates rendered key content client-side only.

The approach

  1. 01Reconstructed the historical URL set and prioritised redirects by value
  2. 02Fixed canonical, sitemap, and structured data implementation in code
  3. 03Moved critical content server-side and rebuilt internal linking to money pages
  4. 04Rebuilt thin pages around topics where genuine expertise existed
  5. 05Established monthly reporting tied to enquiries, not keyword counts
Scope
Technical audit, code remediation, redirect strategy, content briefs, monthly reporting.
Team
1 SEO lead, 1 engineer
Timeline
3 months initial, then ongoing
Discuss a similar problem
CybersecurityCloud & IT

Security readiness programme

Software vendor facing customer security reviews it could not answer.

The problem

No inventory of assets, inconsistent MFA, unpatched dependencies, and backups that had never been restored in anger.

The approach

  1. 01Built the asset and data inventory, then reviewed external exposure
  2. 02Enforced SSO and MFA and rebuilt the permission model on least privilege
  3. 03Added dependency, secret, and container scanning to CI with blocking gates
  4. 04Verified backups by performing a full restore into a clean environment
  5. 05Wrote the incident response plan and ran a tabletop with the leadership team
Scope
Assessment, remediation, CI security gates, restore test, IR plan and exercise.
Team
1 security lead, 1 engineer
Timeline
6–8 weeks
Discuss a similar problem
Cloud & ITWeb

Cloud and edge migration

Company running production on a single legacy virtual machine.

The problem

Deployments were manual and risky, there was no staging environment, and a hardware failure would have meant days of downtime.

The approach

  1. 01Documented the existing system and captured dependencies before touching anything
  2. 02Created an infrastructure-as-code baseline with reproducible environments
  3. 03Moved workloads in phases with a tested rollback at each stage
  4. 04Added monitoring, alerting, and an on-call escalation path
  5. 05Moved the public site to Cloudflare Workers with edge caching and a WAF
Scope
Assessment, landing zone, phased migration, CI/CD, observability, managed operations.
Team
1 platform engineer, 1 engineer
Timeline
10–14 weeks
Discuss a similar problem

A note on our publishing policy

You will not find client logos, star ratings, or percentage improvements on this site. Client work is confidential, and metrics we cannot verify with you in a call are marketing rather than evidence. Ask us for references during a conversation and we will arrange them where confidentiality allows.

Next step

Tell us what needs fixing

Send us the problem in plain language. You will get a considered reply with a recommended next step — not a sales sequence.

Start a conversationhello@allfixai.com

We reply to every enquiry within one business day.