AI-native marketing & development · Riga, Latvia

AI first, in everything we build.

We build custom AI tools and the web platforms they live on — chatbots grounded in your own data, automation workflows, online stores and lead-gen sites. Ads, SEO and analytics sit on top as the growth layer. One senior team, every number measured to the euro.

196
AI skills in production
31
Connected tools & MCPs
85+
Builds shipped
48h
Fastest ship on record
TRUSTED BY 50+ BRANDS ACROSS THE BALTICS, EU, CANADA & USA
LIAAMS SolutionsFlightdelaysVadītāju SkolaWunderLatvija RažoPremium GreenhousesSanta ZvejniecePurus PetTech RecruitmentAMBK GroupeCarMillionaire MindLielzemes PirtsAlma AromaDTSLume DrinksBaltu KoksLIAAMS SolutionsFlightdelaysVadītāju SkolaWunderLatvija RažoPremium GreenhousesSanta ZvejniecePurus PetTech RecruitmentAMBK GroupeCarMillionaire MindLielzemes PirtsAlma AromaDTSLume DrinksBaltu Koks
What we do

Three layers, one team

AI development leads, the web platform carries it, and the growth layer brings the traffic. Most clients start with one layer and add the next once the numbers hold.

01 · AI development

Chatbots grounded in your own knowledge base (RAG), WhatsApp and web assistants that answer around the clock and qualify enquiries, automation workflows across your CRM, e-mail and internal systems, and in-house AI training for your team. Everything plugs into the tools you already run — Pipedrive, HubSpot, Slack, Notion, e-mail, POS and ERP.

02 · Web development

Online stores, lead-gen sites, marketplaces and internal tools — built AI-first, which is why they ship in days and weeks, not quarters. Shopify, Lovable, React or fully custom: we pick the stack that fits the job, and wire every build for measurement from day one.

03 · Growth layer

Google and Meta ads, SEO and AEO, and analytics we set up ourselves so the numbers can be trusted. Reporting is a dashboard you can open any day of the month, not a slide deck at the end of it.

The stack our own agents read and write (31 tools & MCPs)
  • Google Ads
  • Meta Ads
  • GA4
  • Microsoft Clarity
  • SE Ranking
  • Shopify
  • Supabase
  • Notion
  • Slack
  • Linear
  • HubSpot
  • Apollo
  • Figma
  • Canva
  • Lovable
  • our own Hub MCP server
The work

Every build, in full

85+ builds shipped. 16 of them are written up below and 5 more are named on the bench — what each one does, the AI functionality inside it, the stack it runs on and what you can check for yourself. Not AI in the pitch deck: AI in production, including the 196 codified skills our own delivery runs on.

Our products · 5

Products we own and keep shipping — the fastest way to see how we build.

WhatsApp Sales Machine

Our productBuilt in 6 weeks
A WhatsApp AI bot that sells while you sleep
What it does
  • Replies to customers 24/7 in native Latvian, in under 3 seconds
  • Auto-qualifies leads — industry, company size, budget — and hands only ready-to-close deals to the team, with a lead score in the CRM
  • Runs whole transactions in the chat: inventory lookup → order → payment link (Stripe, Latvian banks), price calculators, calendar bookings with reminders, returns with auto-approval rules
  • Plugs into the tools a business already runs — Pipedrive, HubSpot, Slack, Notion, email, POS/ERP — instead of replacing them
  • Demo walks through 12 industry scenarios end to end: from web-dev sales and agriculture quick orders to vet bookings, freight quotes and insurance calculations
The AI part

LLM over WhatsApp Cloud API with RAG on the client's own data — it answers from the business's inventory, calendar and price list, not from vibes.

  • ▸RAG over the client's own inventory, price list, calendar and documents — answers are grounded, not generated from memory
  • ▸Native Latvian conversation, sub-3-second response time, around the clock
  • ▸Lead qualification and scoring written straight into the CRM
  • ▸Intent recognition that routes a chat to order, booking, quote or return without a menu tree
  • ▸Transactional execution: the model triggers real actions — stock lookups, payment links, bookings, refunds under set rules
  • ▸Escalation to a human the moment a deal is ready or the model is out of its depth
Built with
  • WhatsApp Cloud API
  • OpenAI / Gemini models
  • RAG + vector search over client data
  • Supabase (Postgres) for state and conversation history
  • n8n automation workflows
  • Stripe and Latvian bank payment links
  • CRM connectors — Pipedrive, HubSpot, Slack, Notion, POS/ERP
Checkable
  • ✓The full scroll-driven demo runs inside this OS — open the WhatsApp Bot app and try all 12 scenarios
  • ✓Also live on clickscale.agency under AI development

Tveris

Our productBuilt in 1 week
AI meeting transcription, tuned for Latvian
tveris.app
What it does
  • Records and transcribes meetings, generates formal minutes
  • Optimised for Latvian — a low-resource language most transcription vendors handle badly
  • Pay-as-you-go pricing; Baltic expansion planned
The AI part

Solving speech-to-text for a low-resource language is a harder problem than English.

  • ▸Speech-to-text tuned for Latvian, where off-the-shelf models degrade badly
  • ▸Speaker-attributed transcript from a raw meeting recording
  • ▸Formal minutes generated from the transcript — decisions and actions, not a wall of text
Built with
  • Web app + landing, shipped in a week
  • Speech-to-text model tuned for Latvian audio
  • LLM summarisation layer for minutes
  • Usage-based (pay-as-you-go) billing
Checkable
  • ✓Project created 4 Aug → landing page live 11 Aug

Atklāj Latviju

Our productBuilt in 4 weeks
AI route planner on an automated data pipeline
What it does
  • Tourism discovery platform: interactive map (Mapbox + PostGIS) as the primary interface
  • AI-assisted personalised itinerary planner
  • Automated nightly import pipeline from Google Places with progress tracking in the admin
  • SEO/AEO-optimised object pages
The AI part

The AI feature rides on real data infrastructure — a nightly ingest pipeline, not a wrapper around a prompt.

  • ▸Personalised itinerary planning from the traveller's own interests, time budget and start point
  • ▸Routes built over live geodata — the plan only contains objects that are actually in the database
  • ▸AI-assisted object descriptions written for both search engines and answer engines (SEO/AEO)
Built with
  • Mapbox GL interactive map as the primary interface
  • PostGIS geospatial queries on Postgres
  • Nightly Google Places import pipeline with progress tracking in the admin
  • LLM itinerary planner over the imported dataset
  • SEO/AEO-optimised object pages

Auto Scout

Our productBuilt in 2 weeks
AI deal scoring and negotiation strategy for used cars
aito-shop.lovable.app
What it does
  • Paste a used-car listing URL or set search criteria
  • Market price comparison, deal scoring, and a generated negotiation strategy
The AI part

Unstructured web input → structured judgement → an action the user can take. The pattern most businesses actually want.

  • ▸Listing extraction: a pasted URL becomes structured vehicle data — model, year, mileage, condition signals
  • ▸Deal score against comparable market listings
  • ▸Generated negotiation strategy — the arguments and the walk-away number, not just a verdict
Built with
  • Lovable + React + TypeScript
  • Listing fetch and parse pipeline
  • LLM extraction into a structured vehicle schema
  • Comparison and scoring logic over market listings

ChatBot Companion

Our productBuilt in 4 weeks
Embeddable knowledge-base chatbot
What it does
  • Customisable chat widget that drops into WordPress
  • Answers from a client knowledge base, connected through to an LLM
  • Running live on clickscale.agency inside the FAQ and web-building sections
The AI part

Grounded in the client's knowledge base — it answers from documents, not vibes.

  • ▸Retrieval over the client's own documents and FAQ before the model answers
  • ▸Answers cite what they came from, so a wrong answer is traceable to a wrong document
  • ▸Tone and behaviour configurable per client without touching the widget code
Built with
  • Embeddable JavaScript widget — drops into WordPress or any site
  • Knowledge-base ingestion + retrieval layer
  • LLM answering over the retrieved context
  • Per-client configuration for branding and tone
Checkable
  • ✓Deployed on our own front door — try it on clickscale.agency

Client builds · 8

Client-owned builds. Described by functionality and sector; the client's name stays theirs to share.

The Wed Thread

Client buildBuilt in 3 months · still shipping
US wedding marketplace with an AI planning core
thewedthread.com
What it does
  • Two-sided marketplace for the US wedding market
  • Couple side: AI Wedding Builder plus a conversational thread that plans the whole wedding
  • “Your wedding brain” — persistent AI memory that holds every detail about the couple and feeds it into every later step
  • Inspiration, moodboards and vendor outreach in one workspace
  • Vendor side: profiles, private availability calendar, direct couple↔vendor messaging, vendor inbox
  • Guest-list model that handles individuals, couples and families with formal titles
The AI part

A persistent memory surface: the product remembers the couple, not just the session.

  • ▸AI Wedding Builder — turns a short brief into a structured wedding plan
  • ▸Conversational planning thread that keeps context across sessions and devices
  • ▸Persistent couple memory store that every later feature reads from — budget, style, guest count, decisions already made
  • ▸Inspiration and moodboard suggestions drawn from what the couple has already chosen
  • ▸Vendor outreach drafted from the couple's own brief, not a blank message box
Built with
  • React + TypeScript front end
  • Tailwind CSS design system
  • Supabase — Postgres, auth, row-level security, storage
  • LLM planning layer with a persistent memory store
  • Relational guest-list and vendor-availability model
Checkable
  • ✓Three months of continuous commits
  • ✓Both sides of the marketplace shipped

CoreWatch

Client buildIn production · still shipping
Ops platform for an aviation passenger-rights operation
What it does
  • Claims: intake, processing and per-airport work queues, with attribution held against each claim
  • A field-agent network run by airport — availability, objectives, commission terms, monthly settlements and locked payroll runs
  • Recruitment and onboarding: candidate pipeline with interview summaries pulled from Google Docs, onboarding modules, questions and per-person progress
  • Outreach tracking with monthly totals and invoice snapshots
  • Operational intelligence: TAF and SIGMET logs, airspace and world-hazard feeds, per-airport weather outlooks that raise alerts for the agents on the ground
  • Seven scheduled dispatch endpoints running the daily admin brief, agent digests, calendar sync, weather alerts and the weekly recruitment report
  • Latvian, English and Russian throughout
The AI part

Admin questions are answered in SQL the page shows and can re-run — every question is logged, one row each.

  • ▸Natural-language questions over the operational database: the answer arrives as a summary, the rows, and the SQL that produced them, with a re-runnable history
  • ▸Claim-brief generator where the EU-261 figures are ground truth the model may not recompute — each flag becomes an open check, and citations are never invented
  • ▸Knowledge-base importer that turns pasted spreadsheet rows into draft entries, shape-checked in code and unpublished until an admin ticks them
  • ▸Agent assistant that shows each agent their own station live — today's alerts, nearby hazards and flights
  • ▸Guardrails written into the spec: the model answers only from the material it is given, never states a compensation amount it was not handed, and the on-screen conversation history is never sent back to it
Built with
  • TanStack Start + React + TypeScript
  • Supabase — Postgres, auth, row-level security
  • Drizzle migrations
  • Cron-driven dispatch endpoints for briefs, digests and alerts
  • Google Calendar, Google Docs and Outlook connections
  • Queued transactional email with send state, unsubscribe tokens and suppression
  • Aviation weather and hazard feeds (TAF, SIGMET, airspace)
Checkable
  • ✓221 migrations and 50 test files in the repository
  • ✓34 database views, all with security_invoker on — checked against pg_catalog, not assumed
  • ✓The database contract is kept as a dated snapshot re-derived from the live catalog, with the SQL that produced each number

WakeSurf Map

Client buildIn build 2026 · still shipping
Custom marketplace for wakesurf spots, clubs and coaches
wakesurfmap.com
What it does
  • Map-first marketplace: spots, clubs, coaches and events on a Mapbox map, each with its own page and search
  • Booking flow end to end — availability, reservation, payment, rescheduling, an .ics calendar file per booking, and reminders
  • Three dashboards: coaches manage services, availability, earnings and their Google Calendar; partners run their spots, coaches, bookings and stats; admin runs everything above it
  • Payouts and club billing through Stripe, with the settlement side visible in the admin
  • Club claiming and spot registration, then verification before anything goes live
  • Admin surfaces for translations, SEO, gallery, messages, accounts and bulk spot import
The AI part

Eleven server-side AI modules, and not one of them publishes anything on its own.

  • ▸Trip planner that streams an itinerary built from semantic retrieval over spots and clubs, with the coaches, events and forecasts that hang off that shortlist — lexical ranking stays as the fallback
  • ▸Vector embeddings (1536-dim) over the catalogue, so retrieval matches meaning rather than keywords
  • ▸Spot enricher that turns a sparse spreadsheet row into a drafted description and structured details — boat brands, skill levels, languages, FAQ
  • ▸Event extractor that reads a championship or federation page and returns structured event drafts
  • ▸Review summariser condensing published reviews into three pros and three cons
  • ▸SEO generator writing location landing-page content from the spots actually in that area
  • ▸Everything the models produce is a draft: imported spots land as pending and events as draft, and an admin edits every field before it publishes
Built with
  • TanStack Start + React + TypeScript
  • Mapbox GL for the map interface
  • Supabase — Postgres, auth, row-level security, SSR sessions
  • Stripe payments, payouts and club billing
  • Google Calendar integration for coaches, ICS export per booking
  • Lovable AI Gateway — Gemini 3 Flash for generation, text-embedding-3-small for retrieval
  • TipTap editor for editorial content, i18n with an admin translation surface
Checkable
  • ✓125 routes and 63 migrations in the repository
  • ✓Every AI surface writes drafts only — nothing it generates reaches the public site unreviewed

Purus Pet

Client buildIn build since July 2026 · still shipping
B2B pet-care brand platform with an agent-readable catalogue
puruspet.com
What it does
  • Brand and catalogue site for a professional pet grooming and care range — 27 products across grooming, paw and nose, ear and eye, dental, odour control and specialty ranges
  • Three partnership models explained and priced separately: private label, white label and resell
  • Partner tracks written for each buyer — veterinarians, local pet stores, e-commerce sellers and new businesses
  • Sample ordering runs through Shopify while the marketing site, catalogue and B2B pages stay local; the cart re-syncs when a shopper returns from checkout
  • Enquiries and Shopify orders flow into Odoo through edge functions, so sales never retypes a lead
  • Latvian and English, with campaign attribution (UTM, gclid, fbclid and the rest) held across the visit
The AI part

The catalogue is exposed as an MCP server, so an AI agent can search the range and quote specs directly.

  • ▸Public MCP server over the B2B catalogue with four tools: list the ranges, search by keyword, category or partnership model, fetch full product specs, and explain the partnership models
  • ▸The agent-facing snapshot is generated from the same product file the site renders, so the two cannot disagree
  • ▸Specs an agent can answer from: benefits, ingredients, how to use, packaging, MOQ and the spec-sheet PDF
Built with
  • TanStack Start + React + TypeScript
  • Tailwind CSS
  • Shopify Storefront API — headless cart and checkout
  • Supabase for forms, auth and storage
  • Odoo integration via Supabase edge functions
  • MCP server (@lovable.dev/mcp-js) over the product catalogue
  • Transactional email templates, GA4 dataLayer and first-touch attribution
Checkable
  • ✓54 routes across catalogue, partner and service pages

SINERGIA

Client build2026 · in production
Industrial-symbiosis platform for the Latvian circular economy
sinergia.lv/en
What it does
  • Resource marketplace where one business lists a surplus material and another takes it as an input — every listing is submitted, audited, then published
  • By-product classifier that says what a given material can actually become: animal feed, a food ingredient, bioethanol feedstock, and on what basis
  • Resource mapping and a participant map, plus cross-sector collaboration, market research and experience stories
  • Accelerator programme with its own call, applications and participant administration
  • Member area for submitting and managing your own resources; admin over resources, applications, participants, users, FAQ, accelerator entries and a change log
  • Training, news and useful materials, the whole site in Latvian and English
  • Queued transactional email with send state, unsubscribe tokens and suppression
Built with
  • TanStack Start + React + TypeScript
  • Supabase — Postgres, auth, row-level security
  • Leaflet map of resources and participants
  • i18next across a fully mirrored LV/EN route tree
  • Email outbox with queue processing and delivery state
  • Audit trails on both resources and editorial content
Checkable
  • ✓79 routes and 42 migrations in the repository
  • ✓No model anywhere in it — the by-product classifier is curated domain knowledge, not a generated answer

Asais Pipars

Client build2026 · in production
Headless Shopify storefront with a heat-matching quiz and a generated catalogue
asaispipars.lv/en
What it does
  • Products come live from Shopify while the storefront, story and catalogue stay local — Shopify handles cart and checkout only
  • Every product carries a heat rating and the peppers behind it, with a reference showing each pepper's Scoville range and which products use it
  • Quiz that recommends products by how close their heat sits to what the visitor picked
  • Wholesale catalogue generated as a PDF in both languages — category and heat breakdowns, pairing stats and wholesale terms — handed over in exchange for an e-mail
  • Stockist map for where to buy, plus newsletter capture
  • Latvian and English, every route mirrored
Built with
  • TanStack Start + React + TypeScript
  • Shopify Storefront API — headless cart and checkout
  • Supabase for catalogue leads, newsletter and store credentials
  • pdf-lib + fontkit generating the wholesale catalogue in the browser
  • Mapbox GL for the stockist map
Checkable
  • ✓39 routes across a fully mirrored LV/EN tree
  • ✓The PDF catalogue is generated from the same product data the store renders, so the two cannot disagree

Prefab marketplace

Client buildBuilt in 6 months · still shipping
Multi-country marketplace with an AI localisation layer
prefabmarket.eu/en/
What it does
  • B2C marketplace for modular and prefab homes
  • Role-based access for admins, manufacturers and public users
  • Multi-language routing with a backend translator model doing content localisation
  • Supply-side automation: manufacturers with incomplete profiles are hidden from live listings and nudged to finish, with a nudge counter in the admin
  • Architecture built for country expansion
The AI part

A translator model in the backend localises content across markets automatically.

  • ▸Backend translator model localising manufacturer content into every market language automatically
  • ▸Listing quality gate — incomplete supply-side profiles never reach the public marketplace
  • ▸Automated nudge sequence pushing manufacturers to complete their profiles, counted in the admin
Built with
  • Lovable + React + TypeScript
  • Supabase — Postgres, auth, role-based access and row-level security
  • Multi-language routing (i18n) with a backend translation model
  • Admin panel for moderation, nudges and listing state
  • Architecture parameterised per country for expansion
Checkable
  • ✓Six months of continuous iteration

Quote-to-invoice automation

Client buildBuilt in 5 months
Industrial e-commerce across three tax regimes, for a Baltic wood-pellet manufacturer
What it does
  • Live pricing calculator computing delivery fees from Latvian postal codes; pallets, Big Bags or pickup
  • Customers in the EU, UK and USA select their country — VAT and taxes calculated to their regime
  • Bank-transfer orders auto-generate a PDF invoice and email it to the customer
  • Admin panel a non-technical client operates: prices, product availability, restock notifications, incoming applications
The AI part

Quote → order → invoice with no human in the loop, including US customers.

  • ▸The whole quote-to-invoice path automated end to end — no human keys in an order
  • ▸Rules engine handling three tax regimes (EU, UK, USA) so the customer sees the right price at checkout
  • ▸Built AI-first: five months from first line to a system the client's own staff run unaided
Built with
  • Postal-code-driven delivery pricing calculator (pallets, Big Bags, pickup)
  • Multi-regime VAT/tax engine — EU, UK, USA
  • Automated PDF invoice generation + transactional email
  • Admin panel for prices, stock, restock notifications and incoming applications

Internal tools · 2

The tools the agency runs on. No links — client data stays inside — but the architecture is on the table.

Agentic Command Center

Internal toolBuilt in 6 weeks · in production
The dashboard the agency actually runs on
What it does
  • Single view over clients, the skills library (196 codified skills), connected tools and live automation workflows
  • Prospect scoring system
  • Multi-step outreach engine: email sequences with open/click tracking, human-task steps interleaved with automated ones, cron-driven sequence automation
  • Gated behind auth + row-level security — the preview shows the front door; client data stays inside
The AI part

We didn't buy an AI stack — we built the one we run our own delivery on.

  • ▸196 codified agent skills across 10 categories — audits, ads copy, SEO, design, security, ops — so every team member gets the same output quality
  • ▸31 tools and MCP servers the agents read and write directly: Google Ads, GA4, Meta, Shopify, Supabase, Notion, Slack, our own Hub MCP
  • ▸AI prospect scoring feeding the outreach engine
  • ▸11 scheduled routines and 4 live workflows running research, monitoring and reporting without a human starting them
Built with
  • React + TypeScript dashboard
  • Supabase — Postgres, auth, row-level security
  • Own MCP server exposing the hub to the agents
  • Cron-driven sequence automation, email open/click tracking
  • Claude agents running the codified skills library
Checkable
  • ✓196 codified skills across 10 categories, queried from this system's own database
  • ✓31 connected tools & MCPs, 11 scheduled routines, 4 live workflows

Reseller Hub

Internal toolBuilt in 3 months
AI-assisted commerce operations for second-hand apparel
What it does
  • Centralised product data, media and sales history driving purchasing decisions
  • Generates multi-language listing titles and descriptions
  • Profit and ROI calculated per unit; streamlined marketplace listing creation
  • AI image processing on product photos
  • Sourcing panel that surfaces profitable resale opportunities
The AI part

AI applied to a workflow with a measurable unit economic: time and margin per listing.

  • ▸Multi-language listing titles and descriptions generated per item
  • ▸AI image processing on product photos — marketplace-ready shots without a studio
  • ▸Sourcing recommendations ranked by expected margin, from the platform's own sales history
Built with
  • Central product, media and sales-history database
  • LLM copy generation per marketplace and language
  • AI image-processing pipeline for product photos
  • Per-unit profit and ROI calculation
  • Marketplace listing export

Open source · 1

Built in the open, methodology and changelog published with it.

Packaging Navigator

Open sourceBuilt in 48 hours
EU packaging-regulation data engine, built in 48 hours
package-stamp-guide.lovable.app
What it does
  • Compliance hub + cost calculator for the EU Packaging & Packaging Waste Regulation (in force 12 Aug 2026)
  • Structured dataset for all 27 member states: EPR tariffs per material, competent authority, legal basis, registration costs, deposit-return status, reporting cadence
  • Full-cost calculator that keeps estimates separate from hard data
  • Three-step action guide that outputs a per-country compliance plan; CSV + PDF export
  • Public methodology page and changelog
The AI part

Built with Claude Code — ten reviewed pull requests merged in two days, each commit citing the primary source it verified.

  • ▸Agent-driven research: each of the 27 country files assembled from primary sources, with the source URL and check date carried per fact
  • ▸A seven-state provenance model the agent has to fill in — an unverified fact cannot silently become a verified one
  • ▸Agent-generated per-country compliance plan from the user's own packaging inputs
  • ▸Adversarial self-audit: we pointed the agent at our own build, and shipped its findings publicly
Built with
  • Lovable + React + Vite + TypeScript
  • Tailwind CSS
  • Structured per-country dataset with per-fact provenance
  • Client-side cost calculator, CSV + PDF export
  • Claude Code as the build pipeline — reviewed pull requests, not vibes
Checkable
  • ✓27/27 country files; 18 verified against primary sources, 9 explicitly flagged unverified with the reason
  • ✓Seven-state per-fact provenance model — every fact carries its own source URL and check date
  • ✓Unknown rates are never treated as zero; partial totals are labelled a floor
  • ✓We ran an adversarial audit against our own build and shipped the fixes publicly

Also on the bench · 5

Shorter builds — one line each, and the piece of the stack worth naming.

eSIM e-commerce

Payments, two-tier admin, referral tracking, sales analytics, CMS, and a CRM capturing cruise line + departure port at checkout for segmented advertising

Stack: Payments, referral attribution, CMS, CRM hand-off to the ads platforms

White-label interactive video widget platform — video funnels for lead capture, review widgets, agency multi-tenant

Stack: Embeddable widget, multi-tenant agency accounts

Window & door configurator

Live SVG schematic that redraws in real time and outputs a full spec for sales; email or WhatsApp submission. Built in ~2 weeks

Stack: Real-time SVG rendering, spec export, email + WhatsApp submission

Community Connect

Event platform behind Latvia's first Lovable community meetup: applications, approvals, SSO attendee networking, calendar invites, automated reminders

Stack: SSO, calendar invites, automated reminder sequences

ClickScale OS

The site you're using right now — faux-desktop with draggable, resizable windows. We built our own OS as a homepage

Stack: React + TypeScript + TanStack Start, Tailwind, Three.js for the 3D modes

Web development

Platforms with the numbers attached

Three builds where the result is measurable, taken from the client's own analytics and platform data.

E-commerce / Home & Garden · 6 weeks

Premium Greenhouses

A four-figure product needs a storefront that earns trust before the price tag appears. We rebuilt the Shopify store mobile-first with a clean, high-trust UI and a shorter checkout — sales more than doubled after relaunch.

  • Shopify
  • Liquid
  • JavaScript
  • Klaviyo
  • Google Analytics
Conversion:
+156%
Load time:
1.2s
Traffic:
+340%
Lead Generation / Legal · 4 weeks

FlightDelays.eu

Flight compensation is legal small print most passengers give up on. We turned it into a guided, mobile-first claim flow that answers “am I eligible?” in seconds — with steady growth in qualified leads since launch.

  • Lovable
  • React
  • Tailwind CSS
  • Supabase
  • AI Integration
  • Google Analytics
Conversion:
+234%
Load time:
0.9s
Traffic:
+420%
Forestry / Lead Generation · 1 week

Baltu Koks

A landing page where an interactive forest-value calculator does the selling: visitors get an instant valuation, the client gets a qualified lead. Built, wired to Google Ads and CRM automation, and live in 7 days.

  • Lovable
  • React
  • Google Ads
  • Zapier
  • Microsoft Clarity
Conversion:
+280%
Load time:
0.8s
Traffic:
+450%
Growth layer

Ads, SEO & analytics on top

What we build needs traffic and proof that it converts. The growth layer runs on the same AI-first workflow — and on tracking we set up ourselves, so the numbers can be trusted.

Meta Ads (Facebook & Instagram)

Full-cycle Facebook and Instagram ad development, testing and optimization, focused on leads, sales and predictable ROI. We work with clear strategy, data and regular testing — not guesswork.

Best suited for: E-commerce, local businesses, B2C

Google Ads

Reach high-intent buyers at the exact moment they're searching for your products or services. Expert Google Ads management that delivers qualified leads and measurable ROI.

Best suited for: Service businesses, e-commerce

LinkedIn Ads

Connect with professionals and decision-makers through strategic LinkedIn advertising. Generate high-quality B2B leads with precise targeting that converts.

Best suited for: B2B, professional services

AEO Optimization

Position your brand in the AI-powered future of search. We optimize your digital presence so ChatGPT, Perplexity, Claude and Gemini recommend your business when users ask for solutions.

Best suited for: Forward-thinking brands

Keyword Research & Strategy

Uncover high-value opportunities with comprehensive keyword and competitor analysis. The foundation every other SEO and content decision builds on.

Best suited for: Anyone starting or refreshing SEO

SEO Audit

Comprehensive technical and on-page audit with prioritized recommendations. Identify what's holding rankings back and what to fix first for fastest growth.

Best suited for: Existing sites stuck or losing rankings

Analytics & Tracking

Make data-driven decisions with proper tracking across every channel. GA4, GTM, server-side events and conversion APIs — no more flying blind.

Best suited for: Teams running ads

Selected results — every figure from the platform itself:
Lead Generation / Google Ads
3,490 applications at €16.76 per lead
Impressions:
806K
Clicks:
77.2K
Conversions:
3.49K
Education / Training Services
157× ROAS lead-gen system on Meta
Applications:
106
Cost / app:
€6.20
Convert to client:
90%
Solar Panel Sales & Installation
Regional Meta lead-gen with controlled CPL
Qualified leads:
123
Cost / lead:
€8.29
Impressions:
97.1K
E-commerce / Shopify
118K organic impressions from SEO fundamentals
Impressions:
118K
Clicks:
2.87K
Avg. CTR:
2.4%

We don't publish price lists — scope and price are agreed in a conversation, once we understand your situation and goals.

Process

How a collaboration runs

No packages off the shelf. We start with your data, agree on the goals we'll be measured against, and only then build.

  1. 01
    Discovery & Audit · 1–2 weeks

    We map your funnel, ad accounts, tracking and competitors before a cent is spent. You leave with a clear gap analysis and the opportunities worth chasing.

  2. 02
    Strategy & Roadmap · 3–5 days

    A costed plan tied to your numbers — the channels, budgets and KPIs we'll be measured against, agreed before we build.

  3. 03
    Build & Launch · 2–4 weeks

    Campaigns, conversion tracking and the pages they land on go live together, wired to clean data so the results are trustworthy from day one.

  4. 04
    Scale & Optimize · Ongoing

    We pour budget into what converts and cut what doesn't — refined continuously and reported weekly in numbers you can check yourself.

Clients

What our clients say

27 reviews, unedited — named clients and public Google reviews, Latvia and abroad.

“SIA AMBK definitely recommends cooperation with Click Scale Agency. The services provided are not only of high quality, but are also fulfilled within the agreed deadlines. It is also nice that the entrusted channels are managed proactively and provide quality recommendations, without special invitations. If you are 'stuck' with a company that does not achieve the planned goals and talks things over with slogans such as 'well, you need to increase the budget…' — we recommend contacting Click Scale Agency. In our case, these professionals adapt to the company's requirements, perform work systematically, start with the most painful points, and in the end this results in much more stable and predictable indicators of incoming lead flows. We wish Click Scale Agency to continue achieving good results, always be able to maintain a personal approach to customer needs, and not stop growing.”
Kristaps
AMBK Group
“A very positive experience! It is definitely the best agency I have ever worked with. I was pleasantly surprised by how much they go into depth and really want to understand the situation and needs before anything is created — nothing is done formally or in a hurry. Professionalism, responsibility and genuine interest in the result. I definitely recommend!”
Gatis Kokins
Purus Pet
“Fantastic cooperation! Very knowledgeable and responsive team. Perfect support especially for beginners — they know how to explain clearly and teach everything that's necessary. I really appreciate the teamwork with clients. I highly recommend them to anyone looking for partners in both online store creation and marketing. Thank you for what you have already done!”
Marta Dreimane
Kai Ora Cosmetics
“Cooperation with Click Scale Agency has been very valuable to us. During the consultations we received not only practical advice on advertising, but a broader perspective on marketing in general. It felt like they really delved into our business and thought about how to help us specifically, rather than offering ready-made templates. That gave us confidence and helped us make safer decisions. We're very satisfied with the results and with the cooperation process itself.”
Lauris Lauskis
Millionaire Mind
“Professional and personal support in achieving marketing goals. Kristaps always chooses quality over quantity, ensuring that each project is implemented at the highest level. Our cooperation brings real and positive value to the company.”
Oskars
DTS.lv
“We are very happy that we chose to work with this team. You truly understand my needs and help develop Alma Aroma marketing activities.”
Aelita
Alma Aroma
“Very professional team — they delve into the specifics of the business and look for solutions, not just “fulfill the task”. Pleasant to work with — everything is clear, fast and with results. Able to find potential clients with the lowest possible budget.”
Kaspars Cveiģelis
Vadītāju Skola
“Great agency that deeply understands the technical aspects of Meta ads, in our case. Very clear communications and delivery on time of projects assigned.”
Lauris Kalnins
Wunder
“Click Scale helped with advertising support and implementation of quality projects for the Blueberry Nest advertising campaign!”
Ērika
Melleņu Ligzda
“Great experience working with Google Ads projects. Works hard, communicates effectively, and the cooperation went smoothly.”
Anita
Latvija Ražo
“Click Scale Agency quickly and professionally solved our task — to reach the specific audience and register on our e-commerce page. I highly value Kristaps and his team's knowledge in digital marketing.”
Patriks Grīva
Google review
“I have worked with Click Scale on several projects. What sets them apart in the Latvian market is their focus on concrete business results, not empty marketing metrics. Communication is clear, deadlines are met, and they are not afraid to honestly say when an idea won't work. If you are looking for an agency that understands that marketing is an investment, not an expense, this is the right team. I recommend them without a doubt.”
Emīls
Google review
“Working with this company was a fantastic experience. While their project management is top-tier — always on time and incredibly responsive — it is their technical knowledge that truly sets them apart. They possess exceptional expertise in SEO and Digital Advertising. They didn't just execute tasks; they helped us clarify our project vision and applied a data-driven approach to our ads that yielded great results. If you are looking for a team that is proactive, flexible, and masters the digital marketing landscape, look no further.”
Kalvis Ceizins
Google review
“The entire website design turned out elegant, sophisticated, and truly aligned with my brand vision. They not only created the website, but also helped me rethink and refine my branding strategy along the way. The final result was exactly what I was hoping for.”
Casper Zhu
NeoSpace
“Absolutely recommend. Always positive and proactive in communication, responsive to all and any needs, highly responsible project management, always true to set deadlines. If you still don't have a clear vision of your project and deliverables — they're there to help formulating it.”
Anna Jekabsone
Google review
“Looking for a reliable digital marketing partner? I wholeheartedly recommend Click Scale Agency! 🚀 My experience with Mārtiņš Āriņš started 5 years ago, and he has been an irreplaceable assistant in the world of Meta advertising. With Mārtiņš, everything is always precise and responsible — even when everything seems technically complicated, he finds a solution. Now that Mārtiņš has joined forces with Click Scale Agency, the power is even greater. They don't just run ads, but look at the whole business as a whole — data, strategy and real sales results. I feel confident that my marketing is in the hands of professionals. Thank you Kristaps, Dainis and Mārtiņš! We are a great team!”
Ināra Pomere
Google review
“Their service made our daily work easier and increased revenue. Responsive people, fast reactions, easy communication. We do less and get more. I recommend them to anyone whose sales volume has grown to the point where it demands automation.”
Inese Valtere
Sajūti Dārzu
“Together with Click Scale Agency, we carried out the development/redesign of the project — Google Ads advertising in accordance with the company's guidelines and goals. I highly recommend this company; we received even more than we expected, starting with the first communication and continuing with constant involvement and interest from Click Scale Agency. The company not only made changes and maintained them, but also provided in-depth training so that we can successfully follow the analytics and make changes to it in the future. Very personal attitude, thank you very much!”
Linda Brodiņa
Google review
“We are working with Mārtiņš and Kristaps on one big project! Cool cooperation so far — organic content from our side and paid advertising from Click Scale Agency. The guys are very interested in providing the best service to the client, and I feel it from my side as a cooperation partner. It's a pleasure to work together!”
Ilze Švarcbaha
Google review
“Digital marketing has become such a broad discipline that it is impossible to know and follow all the news — however, after a conversation with Kristaps, you will always feel that it was not only professional, but also motivating. The Click Scale Agency team is truly interested in the growth and results of its clients and partners. I recommend it! You will always receive support and your marketing goals will become a reality.”
Barbara Smildzere
Google review
“Reliable and powerful partners in business! I've experienced several agencies firsthand — both as an employee and by working with them. Latvia is currently full of couch experts who call themselves marketing specialists, so finding a good company right away isn't always possible. Kristaps and Mārtiņš actively share their advice on what kind of commercial to film, bring new ideas, and friendly conversations easily arise on other topics too. The guys are genuinely interested in the results of the commercials and they dig into the nuances of your company. Click Scale Agency is understanding and accommodating, even in an emergency. Overall a very pleasant experience, and a good impression of Kristaps and Mārtiņš as people.”
Miķelis Vīks
Google review
“They work fast, professionally and with genuine interest in hitting the client's targets. Communication is clear, and the results exceeded expectations.”
Jēkabs Sipkovs
Google review
“Great experience on Google Ads projects. Easy to work with, communication is effective and understandable — you always keep the feeling that you know what's being done and why.”
Kaspars Misiņš
Google review
“Thank you for the collaboration! Everything we set out to do worked. See you on the next projects.”
VER-TEC
Google review
“I really appreciate the company's high sense of responsibility and communication skills, even though I started the cooperation with minimal funds, covering basic needs. A company with which you feel understood and heard, where every idea has a solution — plus the extensive experience the company shares. I'm glad I found you, and thank you for creating the feeling that you care about my business and its development in marketing as if it were your own. 💯”
Google review
“A chance meeting turned into a collaboration — from a small conversation about what we each do, we ended up auditing the website (user experience) of the company I run. Initially, as a non-specialist, it was difficult for me to understand the terminology and recommendations, but during a remote conversation we quickly understood what was what. I really like the lead generation tool offered. Click Scale Agency specialists listen, delve deeper, and offer creative solutions.”
Mārīte Mundure
Google review
“Thank you so much for your work — the result was better than I could have imagined. I really liked the positive mindset in communication and the approach to work: you look for solutions, not excuses, and the work is done with maximum responsibility and within the agreed deadlines.”
Google review
Team

Who does the work

A small senior team — the people you speak with on the first call are the same people who build and run the work.

Kristaps Krankelis · Co-Founder & CEO

With over 9 years of experience in performance marketing, Kristaps shapes Click Scale's strategic direction, combining data-driven approaches with practical understanding of what actually works across different markets. Focus on precise marketing and measurable results defines every campaign's structure and the decisions we make.

Mārtiņš Āriņš · Co-Founder & Strategy Director

With a deep understanding of both technology and business strategy, Mārtiņš ensures every project is built on solid technical foundations while driving measurable business results. His commitment to excellence and innovation helps our clients stay ahead in the competitive digital landscape.

The specialists
  • Artūrs · AI Specialist
  • Uno · Web Developer
  • Beāte · AdWords Specialist
  • Uģis · Meta Ads Specialist
8 + AI employees
Questions

Asked most often

Are results guaranteed?

No — and anyone who guarantees results in marketing is most likely lying. We guarantee transparent work, data-driven decisions, continuous optimization, and honest communication about what's working and what isn't.

Do you conduct an audit before starting collaboration?

Yes. We almost always start with an audit of existing data, ads, website, or funnel. This helps identify where the real problems are and where the fastest growth opportunities lie.

How often will I receive reports?

Depending on the project: weekly summaries, monthly detailed reports, real-time dashboard access (if needed).

How quickly can we see initial results?

Depends on the channel: paid ads — first signals within 1–2 weeks; SEO — 2–4 month perspective; complex funnels — gradually. We always tell you what to realistically expect.

Is the collaboration for a fixed term?

Usually — yes. Digital marketing isn't a one-time event. The minimum collaboration period is typically 2–3 months to realistically evaluate results.

Will I have a single point of contact?

Yes. You'll always have a specific person responsible for your project and communication. No being passed around from one person to another.

Let's talk about what AI can do in your business.

Whether you need an AI assistant, a web platform, or growth on top of both — the next step is a short introductory call. No charge, no obligation. We discuss your business and goals, tell you honestly whether and how we can help, and only then talk scope and price.

hello@clickscale.agency·LinkedIn·Riga, Latvia