Keep the connection
For each later partner, you can use the same connection to a source system. The connector reads records into the company's Node. Each new partner receives only the views the company approves.
Axio Lattice connects data across companies and makes it ready for AI. Each company controls what it shares.
Each company keeps its raw records in its own Node. An approved view contains only the information the owner agrees to share.
Connectors read records from your source systems or exported files.
The Node is software that runs where you choose to store your records. Raw records stay there.
Rosetta is the translation layer. It combines a shared vocabulary with each company's own mappings. A mapping links a local field to a shared concept and its unit rule.
Lineage records where a result came from. It links the result to its source file, connector run, mapping version and approvals.
Axio runs the Coordination layer. It manages sign-in, identity, permissions and each company's sharing policies. Each company's gate applies those policies before a view leaves its Node. The Coordination layer holds the audit log and the views each team has permission to see. The audit log records access, changes and approvals.
Blueprints define a workflow's inputs, checks and results.
Tech transfer gives a Manufacturing Partner the process information it needs to make a Pharma Sponsor's medicine. This example follows the process handoff and the return of a direct measurement.
Viable cell density measures the number of living cells in a volume of liquid.
In a 2019 survey, transfers between companies took an average of 5.8 months longer than transfers within one company. The survey covered complete transfers of sterile products.
The sponsor approves a process view for the partner. The partner returns an approved cell-density result through its own gate. The sponsor compares the result with its target. The sponsor traces the result to its source. Each company keeps its raw records in its own Node. The sponsor needs separate permission to open the raw file.
Illustrative example. The sample values, fields, codes, versions and approval records show how Lattice handles a tech transfer.
Pharma SponsorManufacturing Partner
The Pharma Sponsor hands over the processThe Pharma Sponsor defines the cell culture step, the day-10 viable cell density target and the counting method under vcd. The field uses cells/mL. The process record stays in the sponsor's Node.
Today: Teams assemble process steps, targets and methods in separate documents.
Process package v3Awaiting approval
vcdRosetta maps vcd to VCD.01, then to the partner's Viable Cells (E6/mL). The unit rule links cells/mL with million cells/mL. If a term has no approved mapping, Rosetta flags a gap at its source line.
Efficiency: Approved mappings remove retyping and repeated term-by-term reading of documents.
Today: Teams read documents term by term and retype fields into local templates.
vcdVCD.01Viable Cells (E6/mL)Unit rule: cells/mL ↔ million cells/mLE6/mL means million cells per milliliter.
The sponsor approves the process package for the partner. Lattice records the approver, approval time and package version. The approved view passes through the sponsor's gate to the partner.
Efficiency: The approval record replaces requests for signed PDFs.
Today: Teams exchange signed PDFs and ask which version has approval.
Process package v3
The Manufacturing Partner returns the resultDuring the batch run, the Manufacturing Partner's lab system records a day-10 viable cell density of 24.6 ×10⁶ cells/mL. The connector reads the record into the partner's Node. Rosetta applies mapping v14 to link the result to VCD.01 and its source history.
Today: Teams copy the lab result into a results table.
Lab system2044R2_D10_CC.csv Source file stays hereConnector run: CR-01Measured resultAwaiting approval
Viable Cells (E6/mL)Reported value: 24.6 Day 10VCD.0124.6 ×10⁶ cells/mLThe partner approves the result and its supporting details for the sponsor. The approved view passes through the partner's gate. The raw file stays in the partner's Node.
Efficiency: Teams share the approved view without a separate export or a retyped certificate.
Today: Teams export results and retype them into a certificate of analysis.
Measured result
2044R2_D10_CCSource file stays here
The Pharma Sponsor checks the resultThe sponsor sees the result under its own vcd, beside the day-10 target from its process package. The meaning, unit and sample day match the process definition.
Today: Teams compare result names, units and sampling days in spreadsheets.
Pharma Sponsor viewvcd
The sponsor traces the result through mapping v14 and the connector run to 2044R2_D10_CC.csv. The sponsor needs separate permission to open the raw file.
Efficiency: The sponsor can trace the result without asking the partner to send source data.
Today: The sponsor emails the partner for the source record and its history.
Pharma Sponsor viewvcdLineage
CR-012044R2_D10_CC.csv Source access needs permission
Pharma SponsorNodeRaw records stay here
Manufacturing PartnerNodeRaw records stay hereProcess package v3Awaiting approval
vcdProcess package v3Awaiting approval
vcdPharma Sponsor fieldVCD.01Unit rule: cells/mL ↔ million cells/mLViable Cells (E6/mL)Manufacturing Partner fieldE6/mL means million cells per milliliter.Process package v3
Process view received
2044R2_D10_CC.csvCR-01Measured resultAwaiting approval
Viable Cells (E6/mL)Reported value: 24.6 Day 10VCD.01Mapping v14
Measured result
Source file stays here
Measured result
Pharma Sponsor viewvcd
24.6 ×10⁶ cells/mL
Source access needs permission
Illustrative example. The sample values, fields, codes, versions and approval records show how Lattice handles a tech transfer.
For each later partner, you can use the same connection to a source system. The connector reads records into the company's Node. Each new partner receives only the views the company approves.
Rosetta's shared vocabulary defines process steps, measurements and their relationships. For a result, those relationships include the sample, unit, test method and batch. Each company keeps its local field names.
Axio maintains the shared vocabulary. The Pharma Sponsor controls its mappings to that vocabulary. The Manufacturing Partner controls its own mappings. A subject expert at each company approves its mapping rules.
Teams can use an approved mapping for later records and other partners that use the same shared concept. Teams review the meaning, units and source version for each new use.
The partner renames Viable Cells (E6/mL) to Viable cell density (E6/mL). Rosetta holds the affected result for review because the new field has no approved mapping. The partner's subject expert reviews the meaning and unit before approving mapping v15.
The new mapping links the renamed field to VCD.01. The sponsor keeps its vcd mapping. The result from the transfer example keeps mapping v14 in its Lineage.
Each approved mapping change creates a new version. If a team processes a source record again, Lattice creates a new result with the selected mapping version. Lattice keeps the earlier result and its Lineage.
vcdVCD.01
Viable Cells (E6/mL)Mapping v14 approvedViable cell density (E6/mL)
A Blueprint defines a workflow's inputs, checks and results. The tech transfer Blueprint uses existing connections and Rosetta mappings to assemble process information and identify gaps. Teams adapt its checks and sharing rules for the next transfer.
Find answers about fit, data control, quality and setup.
Lattice connects a Pharma Sponsor with a Manufacturing Partner working on the same drug program. The sponsor develops the medicine. The partner makes or tests it. Their teams use approved views to exchange process information and review results.
Your existing systems stay the official sources of your records. Connectors read records from those systems or their exported files. Rosetta applies mappings for the views your teams use.
Process documents, batch records and certificates of analysis are part of your quality process. Lattice links the information for the agreed workflow to its source through Lineage. Your teams still review and approve the required documents.
AI can draft mapping proposals that a person at the company that owns the mapping must approve before use. Approved Rosetta mappings then apply the same rules every time. Lattice flags uncertain items for a person to review. Axio agrees the AI services and hosting with you in advance. Axio does not send raw sample values to hosted AI services.
Only approved views cross your Node boundary. Your sharing policy defines the fields, supporting details and recipients for each view. Raw records stay in your Node.
Axio's Coordination layer holds the shared vocabulary, approved views, source references and audit records. If Axio hosts your Node, that Node also holds raw records within your company's chosen boundary. The hosting agreement defines the locations and responsibilities.
Your company decides what to share and who can receive it. The Coordination layer applies permissions by company, program and role. Permission to see a result or its source reference gives no automatic access to the raw file.
Axio does not use your data to train or fine-tune models. Axio does not let third-party model providers train on it. Your records and results reach only the recipients you approve. Other customers can benefit from improvements to the shared vocabulary and connectors that contain none of your records or results.
The Node runs at your facilities, in your cloud environment or with Axio. Your company chooses the boundary that holds its raw records. Axio runs the Coordination layer in each arrangement. The hosting agreement assigns the operating and security responsibilities.
The Node starts an encrypted connection to the Coordination layer. Each end verifies the other's identity. Security review also covers stored data, administrator access and connections to outside services.
Data integrity means that records stay complete, accurate and traceable throughout their life. The Node keeps the original files. Lineage links results to their source files, connector runs, mapping versions and approvals. Lattice records corrections as new entries in the audit log. Earlier entries stay in the log.
Lattice includes controls that support validation for a defined use. Validation needs documented evidence that the installed system works for that use. Your quality team and Axio must establish the required evidence and procedures for your deployment.
The sponsor and partner keep their quality responsibilities. Their quality agreement defines how they divide and oversee the work. Lattice provides approved views and Lineage for their review. Authorized people approve investigations, process changes and product release.
Select a workflow, its source systems and the people who understand those records. Your technology and quality teams define Node hosting, connector access, mapping approval and sharing policies. Source access and data quality determine the setup work.
Lattice includes interfaces for authorized software agents. The Coordination layer applies the company's permissions to their actions. Your team reviews which interfaces, agent actions and human approvals the workflow needs.
The Node keeps records locally during an interruption. When the connection returns, the Node sends only approved updates. Your deployment plan must define storage limits, recovery and warnings for results that are out of date.
Raw records stay in your Node. Axio holds approved views, source references and audit records in the Coordination layer. After the work ends, Axio deletes the data it holds and confirms the deletion in writing. Axio agrees any retained records with you in advance.
Tell us about your organization and how you want to work with Axio on Lattice.