Company: A clinical-stage biotechnology company
Industry: Clinical-stage biotechnology / R&D
As programs matured, the organization found that early R&D decisions still depended on large Excel workbooks. Individual assays lived in lab sheets or instrument exports, but leadership reviews required a stitched picture: antibody developability readouts, functional assay tiers, and production-adjacent metrics side by side. Scientists exported CSVs, pasted into shared templates, and manually reconciled IDs and batch names. Each refresh of a program view meant repeating that cycle, and when two people edited different copies, IC50 summaries and curve-derived flags no longer matched.
Dose-response work made the friction especially visible. Teams moved concentration series into desktop graphing tools to fit four-parameter logistic curves, then copied parameters back into spreadsheets for ranking. That split meant traceability broke at the handoff: it was difficult to prove which raw points produced a given EC50, or to rerun the same fit after a data correction. For a clinical-stage portfolio, where traceability supports both science and quality conversations, the patchwork of files became a liability rather than a convenience.
Scale amplified the problem. Assay tables grew to hundreds of thousands of rows. Column renames and ad hoc schema tweaks in spreadsheets broke fragile lookups. One-off Python or R scripts helped a single analyst but did not give the broader team a repeatable, reviewable path. Version history in shared drives was uneven, and bulk imports sometimes corrupted formatting or silently dropped rows. The group needed one system that could hold large structured tables, preserve lineage, and expose program-level summaries without forcing everyone back into a master workbook.
They also wanted room to experiment responsibly with AI-assisted analysis: summarizing tables, suggesting QC checks, and connecting notebook-style exploration to governed execution. That only works if the underlying data model, permissions, and audit signals are clear. Spreadsheets and disconnected tools could not provide that foundation.
They partnered with Scispot to anchor multi-assay R&D data in lab sheets and program structures designed for aggregation. Master Data Summary capabilities became the hub for cross-program views: dynamic assembly of columns from multiple lab sheets, filters that mirror how scientists already think about cohorts and candidates, and color-coding and QC cues that surface incomplete or out-of-range rows before reviews. Instead of exporting a new stitched workbook every week, teams refresh a living summary that stays tied to source rows.
For dose-response and potency work, the Analyze Agent was extended with native four-parameter logistic (4PL) fitting and high-volume graphing. Scientists can run fits and inspect curves next to the tabular outputs that feed ranking and governance. Where the program required it, outputs could be checked against agreed reference curves or validation notebooks so statistical treatment stays consistent with internal standards. Keeping 4PL inside the same platform reduced the copy-paste loop that had separated raw data from published parameters.
Integrity features were matched to the volume and change rate of their assays. Column-header locking reduced accidental renames that used to break downstream formulas. Snapshots and version history made it easier to see what changed before a decision meeting. Bulk import paths were tuned for very large files so teams did not split data across fragments. Interactive QC visualization helped spot drift or plate effects without leaving the environment where the data is governed.
The roadmap also connected notebook-style work to execution controls: clearer boundaries between exploratory scripts and production-facing sheets, improved file sync and duplication with linkages intact, and governed use of AI-assisted analysis and external agents with explicit usage limits. The goal was not to replace scientific judgment but to ensure every automated suggestion or export could be traced to a defined dataset and permission set.
The organization shifted lead-selection compilation from a spreadsheet-heavy, multi-tool workflow toward program-level summaries and in-platform 4PL. Review cycles spend less time reconciling which file is canonical and more time interpreting biology. When questions arise about a parameter or a flag, teams can walk back to the same rows and fits instead of hunting across email attachments.
Repeatability improved because analysis paths live next to the data they consume. Onboarding for new scientists no longer starts with a tour of fifteen workbook tabs and three desktop tools; it starts with lab sheets, summaries, and documented workflows inside Scispot. That reduces single-person dependency and makes peer review of analytical choices practical.
Success measures for the engagement were defined with the team upfront: shorter time to compile lead-selection datasets for recurring reviews, adoption of Master Data Summary and 4PL workflows across programs, and validation of curve fits against agreed baselines where regulatory or quality discussions require it. Those metrics are tracked as the rollout continues so outcomes stay tied to validated production use rather than one-off demos.
The case for a unified assay and analytics layer is ongoing, but the direction is clear: fewer handoffs, stronger integrity at scale, and a platform that can grow with both pipeline complexity and governed automation.
Dynamic assembly, filters, and QC cues across lab sheets so teams review programs without exporting stitched workbooks.
Dose-response curves and tabular outputs in Scispot, with validation against agreed reference curves where the program required it.
Column locking, snapshots, bulk import for large files, and governed integrations to reduce corruption, loss, and one-off script drift.
