How AbleBlox works
A closed, white-label token exchange built on institutional-grade AI market-making infrastructure — engineered from first principles for loyalty currencies and tokenized assets.
You own the relationship.
We own the infrastructure.
AbleBlox is not a marketplace anyone can join. We partner exclusively with enterprise organisations — restaurant chains, fund managers, gaming studios, hospitality brands — who already have token assets and user relationships. We power their private exchange.
Your users never know AbleBlox exists. They see your brand, your colors, your exchange. We operate as silent infrastructure: matching orders, settling trades, collecting fees from you (not your users), and reporting back through our API.
What makes it closed
User Isolation
Every partner gets a separate user namespace. A Chipotle member cannot interact with, trade against, or even see AlphaFund investors. The environments are completely air-gapped at every layer.
Token Isolation
Each exchange only supports its partner's token type. Tokens cannot cross environment boundaries — by design and by enforcement.
Access Control
Partners control who accesses their exchange. Eligibility is enforced by the partner's identity and loyalty system — AbleBlox enforces it at the API level.
Brand Neutrality
AbleBlox branding is absent from all end-user surfaces in production. The exchange is presented entirely as the partner's product.
AI-driven Automated
Market Making
At the heart of AbleBlox is a proprietary AI-driven Automated Market Making (AI-AMM) engine, pioneered by Irene Aldridge — a recognised quantitative researcher and author with 20+ years of experience in market microstructure, high-frequency trading, and algorithmic exchange design.
Traditional automated market makers quote fixed spreads or use static bonding curves. AbleBlox's AI-AMM is fundamentally different: it continuously learns from order flow, member behaviour, and token-specific signals to quote adaptive bid-ask spreads that tighten when liquidity is healthy and widen intelligently to protect the pool during thin or anomalous conditions.
The result is a market that is always open — providing continuous two-sided liquidity to members even when there are no matching peer orders — while remaining capital-efficient and manipulation-resistant.
AbleBlox AI-AMM Signal Pipeline
The AI-AMM signal pipeline: raw data flows into the deep reinforcement learning agent, which continuously computes optimal bid-ask quotes for the AMM engine. Anomaly detection and inventory risk sub-models feed back into the core model. All matched trades feed back into the training loop, making the system progressively sharper over time.
Deep Reinforcement Learning
The core model is a deep RL agent trained on historical order flow across loyalty token markets. It learns to maximise long-run fee revenue while controlling inventory risk — the same optimisation problem faced by institutional market makers on public equity exchanges, applied to closed token environments.
Adaptive Spread Quoting
Rather than a static spread, the AMM recalculates its bid-ask on every incoming signal. Spreads tighten when order flow is balanced and liquidity is deep; they widen automatically when imbalance, thin depth, or anomalous patterns are detected — protecting the pool without manual intervention.
Manipulation Resistance
An embedded anomaly detection module monitors intra-session order patterns for wash trading, spoofing, and layering — behaviours common in thin markets. When anomalies are detected, the model widens spreads, flags the session for partner review, and can pause quoting automatically.
Continuous Online Learning
Every matched trade flows back into the model as a training signal. The AI-AMM improves continuously as it observes how its quotes affected actual trades — a closed feedback loop that makes each partner's exchange progressively better-calibrated to its specific member behaviour over time.
Pioneering AI-driven
market making
Irene Aldridge is a recognised quantitative researcher and practitioner in market microstructure and algorithmic trading. She has spent over two decades at the intersection of academic research and live market implementation — applying machine learning to the design of trading systems that are faster, fairer, and more robust than their predecessors.
Her work on high-frequency trading, market impact, and exchange design is widely cited in both academic literature and industry practice. At AbleBlox, Irene brings that foundational research directly to bear on a new problem: how to create deep, continuous liquidity for loyalty currencies and tokenised assets that have never had a secondary market before.
The AI-AMM engine she has designed for AbleBlox is not adapted from public-equity market-making software. It is built from scratch for the specific characteristics of closed token environments: lower trading frequency, concentrated member bases, strong redemption seasonality, and no external price anchor. These properties demand different mathematics — and a different model.
"Traditional AMMs use fixed formulas. Real liquidity provision requires a model that learns — from the market, from the members, and from its own mistakes." — Irene Aldridge, CEO, AbleBlox
B2B pricing — invisible to end users
What we charge
0.25% of gross trade value on every matched transaction. Billed to the partner monthly, based on reported volume.
This fee is paid by the enterprise partner, not the end user. Partners decide independently whether to absorb it, pass it through, or structure their own end-user pricing.
Example: Chipotle Rewards Exchange
Chipotle decides what, if anything, they show members about fees.