← Insights
News Analysis25 July 20263 min read
LR
Lichia ReghuCo-Founder & AI Engineering Lead

The Model Is Only 20% of the AI Product

Cambridge-based CuspAI has raised $450 million at a $2.6 billion valuation to use AI for materials discovery. What stands out is not just the model. It is the system around it: generative design, simulation, synthesis planning, and experimental validation. That is where AI becomes a product.

We see the same pattern in document processing

Extracting text is the easy part. The hard part is preserving meaning.

Take tables. A weak pipeline flattens a table into a block of text and hopes the model can reconstruct it later. A reliable pipeline does more. It detects table boundaries and merged cells, preserves rows, columns, and repeated headers, and stores the original table as HTML, Markdown, or structured JSON with cell-level links back to the source page. It creates row-level chunks with the column headers attached, then indexes both the table summary and the individual rows, using hybrid retrieval across embeddings, keywords, and structured filters. When one row matches, it retrieves the full parent table rather than the fragment, and validates totals, dates, and relationships before the result is ever used.

In many cases, the same table should exist in two forms: structured storage for filtering, calculation, and validation, and vector storage for semantic retrieval and question answering. That is the engineering most AI demos leave out.

The same applies to agents

Generating an action is easy. Making sure it has the correct parameters, dependencies, permissions, retry behaviour, and audit trail is the actual product work.

A model can perform well while the system still fails, because a header was detached from its row, stale data remained indexed, the wrong table was retrieved, or an operation ran twice.

Model accuracy is not system reliability. The model may be 20%. The other 80% is building the path from messy input to a trustworthy business outcome.

AI EngineeringIntelligent Document ProcessingAgentic AIRAGAI Product Strategy
Thinking about your AI infrastructure?

Let's talk about how your business is built today — and how we'd make it AI-ready.

Fixed-fee. Delivered in weeks. No lock-in to any single provider.

Book a discovery call →
Related reading
EducationWhat Is Intelligent Document Processing?Most document workflows in financial services still rely on people reading things and typing them somewhere else. Intelligent Document Processing is the category of AI built to replace that work.News AnalysisWhy AI Agent Evaluation Fails SilentlyA conversation log can look completely clean and still represent a failure. Here is the evaluation gap most teams miss when they ship AI agents into production, and what we changed in our own pipelines because of it.News AnalysisWe Tested Headroom on a Real Agentic Workflow. Here Is What It Compressed.We ran the open-source token-compression tool Headroom against a real Kelriva agentic workflow instead of just reading the claims. Here is what the dashboard actually showed, and why the biggest source of waste had nothing to do with the model.
← Back to all insights