Implementation Details

Many times, our 3rd party integrations are defined during the pre-sales process or early during the implementation process with our mutual clients. Below you will find a summary of our integration process. Lucy's development team follows an agile process and requests we provide approximately 8-12 weeks to complete new integrations. 

Implementation & Integration Process for Lucy

  1. Pre-Kickoff & Planning

    • Develop a custom project plan using insights from sales and client persona sessions.
    • Align implementation milestones with the Lucy Sprint calendar.
    • Assign accountability, define stakeholder roles, and set expectations.

  2. Discovery & Requirement Gathering

    • Conduct detailed discovery sessions to understand client needs, provider requirements, available data types, and system requirements.
    • Identify key stakeholders (Developers, TPM, CSM, EPM) and establish project scope.

  3. Project Planning

    • Define objectives, deliverables, and timelines.
    • Identify risks, obtain necessary access credentials (e.g., API tokens), and finalize the integration roadmap.

  4. Implementation & Integration

    • Developers execute tasks based on Jira tickets.
    • The team will first complete an exploratory task (referred to as a Spike) in which any API or technical documentation is reviewed and the integration's architecture defined.
    • Following the exploratory task, the team will complete the execution and build or connect to the endpoint or service as defined in the process above. 
    • QA testing ensures integration accuracy; adjustments are made as needed.
    • Once validated, the integration is deployed within a week or the next available release.

  5. Post-Implementation & Maintenance

    • CSM supports user engagement, adoption, and ongoing improvements.
    • Continuous monitoring ensures system performance and data integrity.

Key Goals of Integration:

  • Enhance efficiency by streamlining data access, saving our mutual clients' time when looking for an answer and focus on more important tasks.
  • Improve decision-making through unified data insights and drive increased usage of our clients' valuable data.
  • Ensure a seamless user experience with robust integrations.

General Guidelines

Items to consider while making repository connections

  • Determine repository of content to connect with Lucy (Box, SharePoint, etc.).
  • Identify and curate folder(s) with appropriate content (e.g. provide external access as needed).
  • Confirm desired source structure and organization within Lucy.
  • Determine specific file types for indexing.
  • Post processing, adding, moving or deleting files within folders will be automatically reflected in Lucy (continue using folders as normal).
  • Lucy will automatically update as you make changes.

Items to consider while making API connections

  • Determine the appropriate mechanism of delivering content (API, JSON, RSS, etc.).
  • Identify and confirm appropriate content access (e.g. API token).
  • Confirm desired metadata structure for Lucy or specific file types for indexing.
  • Confirm range of historic data for indexing and frequency of required updates. 
  • Confirm API limits and cohesiveness with update architecture. 
  • Lucy will automatically query for delta or changes based on the connection mechanism. 

Lucy's Indexing Process

Metadata extraction and enrichment process

  • We classify the document to identify the optimal parsing techniques.
  • Create metadata to support answers.
  • Process images for embedded content.
  • Extract metadata from the documents and images.
  • Enrich metadata based on Lucy knowledge graph.
  • Store enrichments as metadata in Lucy.
  •  Leave original content at source during the extraction of the metadata.

Common Questions to Consider

Supporting unstructured data integrations
  • What kind of content is available to serve our mutual clients?
  • Would you be willing or able to provide a demo of the content as it appears in the original source?
  • What format is the content in? Files, Websites, Videos, Dashboards, etc.
  • How much content does our mutual client have access to?
  • How is the content organized?
  • Are there access controls applied to the content or do all users see the same set of content?
  • Is there an API available where we can list and read this content?
  • If not, is there an underlying repository available such as an S3 that we can connect to?
Supporting unstructured data integration
  • Does the tool or dashboard allow deep-linking to the content?
  • Does the tool or dashboard allow embedding of content?
  • If the tool or dashboard allows embedding, is there a preview image available we can show while it loads?
  • What sort of ways do you want to slice this content in search results?
  • Lucy supports subject extraction such as "Where is the X dashboard?"
  • Lucy also supports market recognition such as "Where is the X France dashboard?"
  • For each dashboard, we will need the following details here. 
  • Please provide any other keywords or tags that are deemed critical for search.
Was this article helpful?