Most pharmaceutical manufacturing facilities today are sitting on more data than they have ever had, and yet the people who need to make decisions are still waiting for answers. The problem is rarely a shortage of data — it is the distance between the data and the person who needs it, and the chain of steps required to close that distance. Getting an answer means navigating disconnected systems, finding the right subject matter expert, waiting for them to pull the data, and receiving a report long after the moment to act has already passed. It is a quiet operational problem that rarely surfaces in boardroom conversations but plays out on the plant floor every single shift. To understand what that gap looks like in practice — and what it looks like when it closes — consider what Itanta Analytics demonstrated at this year's ProveIt conference, working with a real-world biopharma scenario that put exactly this challenge on the table. The Challenge Behind a Paper-Based Biopharma Operation Point Three Therapeutics (PTT) is a growing biopharmaceutical operation producing an enzyme through fed-batch fermentation — a virtual factory scenario presented by ProveIt as part of their enterprise challenge at this year's conference. The setup will feel familiar to anyone working in life sciences: single-use equipment, controlled batch production, and documentation requirements that shape every decision made on the floor. But beneath that familiar surface lies a challenge that extends far beyond PTT. Data flows in from equipment through MQTT streams, but without a PLC control layer to consolidate and contextualise it, that connectivity remains raw and largely unused. Processes are defined on paper, and batch records are assembled manually. The deeper problem is that when something goes wrong — a deviation, a yield drop, an unexpected temperature trend — the entire investigation depends on the availability of one person: the SME who knows which parameters to examine, which batch to compare against, and what the process documentation actually requires. When that person is on another shift, unavailable, or pulled across competing priorities, the investigation stalls and the decision waits. In a GMP environment, that wait carries a real cost — in product, in compliance exposure, and in the confidence of the team that eventually has to make the call. What Changes When You Connect the Silos The starting point for PTT was not a replacement of existing infrastructure but a connection — one that finally brought everything already running into a single environment where it could be used together. Prompt-to-Insights (PTI) by Itanta Analytics connected directly to PTT's live MQTT streams, pulling real-time process data into a unified layer alongside historical batch records and operational documentation, without requiring any manual stitching of systems or a parallel engineering project to make it work. The data that had always existed across the operation became immediately accessible to any team member who needed an answer — not just to the SME or the analyst, but to anyone on the floor or in the office asking a legitimate operational question. "It connects to the data, pulls everything together, and lets you explore it instantly. What's really powerful is that you don't have to be a data scientist. You just need to have a question." — Amy Williams, Skellig Automation The SOP Layer — Where Pharma Gets Different What separates PTI in a pharmaceutical context is its ability to hold process definition and process performance in the same analytical space. When an SOP is uploaded into PTI alongside live operational data, the system no longer simply reads what a parameter value was — it understands what that value should have been at that specific phase of the process. A reading that appears normal in isolation becomes a flagged deviation when read against the SOP specification for that fermentation stage. The institutional knowledge that typically lives in the heads of experienced SMEs, built over years of process development and stored in filing cabinets, becomes a live and active part of every analysis. PTI understood not just what happened during a batch, but what should have happened — and that distinction fundamentally changes how production teams investigate deviations, assess performance, and prepare for audits. The Kind of Analysis PTI Makes Possible The best way to understand what PTI unlocks is to look at the questions it can actually answer — the kind that currently sit in someone's inbox waiting for an SME to respond. In the PTT environment, questions like these were answered directly, in seconds, without a single manual query or system login: - Show me the details of batches executed in the single-use bioreactor. - What batch was running on the Tangential Flow Filtration (TFF) unit on a specific date and time? - What phase was the Tangential Flow Filtration (TFF) unit in at a given timestamp — chase, flush, or processing? - Show the trend for weight for a specific material or batch in the single-use mixing system. - Show the temperature trend for a specific instrument or tag in the single-use mixing system for the current or selected batch. Each of these is a question that would ordinarily require an SME to know which system to open, which parameters to pull, and how to interpret what comes back. PTI answers all of them — in whatever format makes most sense for the question being asked, whether that is a table, a direct text response, or a graph — without any pre-configuration or technical knowledge required. "You've probably done the most with this specific data compared to what I've seen so far." — Amy Williams, Skellig Automation Filters, Comparisons, and Decisions That Don't Wait Once everything is connected, the analysis becomes something any team member can drive without needing the expertise of an SME or the availability of a data analyst. PTI's filter capability allows teams to work through their data with real precision — narrowing by batch ID, time range, operating conditions, or a specific piece of equipment across multiple runs — and because this happens through a plain-language prompt rather than a technical query, the answer comes back in seconds rather than hours. The real value becomes most visible when batches are placed side by side: weight trends, agitator behaviour, and temperature stability across two or more runs appear in a single view, and the differences that would previously have taken an expert hours to isolate become immediately apparent to anyone looking at the screen. Because PTI was built to understand plant hierarchy, batch structure, and process context, the analysis it returns is manufacturing-aware and traceable rather than generic — it knows what a bioreactor is, what a fermentation phase means, and what a deviation looks like in that specific context. The insights generated through this exploration are saved automatically and become the foundation of live dashboards built from real analysis, so there is no manual configuration required and no rebuilding from scratch when the next reporting cycle arrives. There is no learning curve, and no longer any dependency on the one person who knows where everything lives. "For pharma, this is a huge use case. Everyone wants to understand golden batches and compare how phases behave. I was impressed. I think what you've built is really cool, and people should definitely check it out." — Amy Williams, Skellig Automation Turn Data Into Decisions. Instantly! Frequently Asked Questions How do I analyse batch data in a biopharma facility without a data scientist? Prompt-to-Insights (PTI) by Itanta Analytics allows production teams to query batch data — trends, phase comparisons, equipment behaviour, and deviation analysis — through plain-language prompts, with no coding or data science background required. How can pharma teams compare batch performance and identify golden batches? PTI allows teams to compare weight trends, temperature profiles, agitator behaviour, and phase execution across multiple batches in a single view, through a single prompt. Combined with SOP upload, PTI can identify which batches performed closest to defined process standards. Can AI tools work with paper-based pharma operations that don't have full automation? Yes. PTI connects to live MQTT streams, historians, and existing data sources without requiring a fully automated plant infrastructure — designed to work with the data environment manufacturers already have.