# Tiksom

> Tiksom is a reviewed AI agency in Manchester, United Kingdom, listed in the AIAgencies.eu register and last assessed 2026-08-06.

Canonical page: https://www.aiagencies.eu/agency/tiksom
Generated: 2026-09-17 · Curated by Gabor Kiss · No paid placements — inclusion, score, and rank are never sold.

## Facts
- Location: Manchester, United Kingdom
- Website: https://www.tiksom.co.uk
- Hourly rate: €20-40/hr (directory-listed)
- Team size: 10-49
- Founded: 2015
- Services: AI Development, AI Agents, AI Consulting
- Industries: SaaS & B2B, FinTech
- LinkedIn: https://www.linkedin.com/company/tiksomlimited/

## About
Tiksom is a Manchester software company building custom applications and product engineering for startups, with a separate AI practice covering agent systems and machine-learning features.

## Documented clients
Documented engagements: Myaigi, UNICEF, and Erasmus MC. All 3 carry a source URL on file.

## Curator assessment
Reviewed 2026-08-06 by Gabor Kiss against published evidence.

The Myaigi build is the reason this listing exists and it is documented properly: a multi-agent behavior-change backend on LangChain and LangGraph with RAG, a unified memory layer holding user context and history, and an internal Judge Agent enforcing consistency and safety across agents.

The write-up is explicit that the client specified the architecture problem—parallel agents, context switching, accuracy without hallucination—which is a more honest framing than most agency case studies attempt.

The wider portfolio is custom software and mobile work going back to 2015—maintaining and extending an online surveying system for UNICEF, feature work and regulatory compliance on an existing Erasmus MC system—so the AI practice sits inside a general product shop rather than defining it. Two further portfolio entries, Starryai and Typingflow, are presented as products without describing an engagement, so they are excluded here.

The weaker fit is a buyer who wants proof the agent system worked; nothing is published on what it improved, and it is the only AI project on the list.

### Key strengths
- Names its stack down to the framework—LangChain, LangGraph, RAG—which most listed agencies still will not do
- The Judge Agent pattern shows evaluation was designed in, not added after
- Building since 2015, so the engineering practice predates the AI positioning

### Good to know
- No outcome figures published for the agent system—ask what it improved and how that was measured
- One documented AI project; the rest of the portfolio is app and platform work

---
Source: AIAgencies.eu — the curated register of European AI agencies. Rate provenance: "agency-confirmed" = disclosed via submission or claimed profile; "directory-listed" = Clutch or two agreeing B2B directories (lone unverified sources are never written). Full method: https://www.aiagencies.eu/methodology · Rate index: https://www.aiagencies.eu/rates