Workshop Output Strategic Opportunity Report

Where AI can meet LI-COR + Onset across every part of the business.

A synthesis of the ideas the team generated — clustered into strategic themes, prioritized by impact and effort, and sequenced into a practical roadmap leadership can act on. Anchored to the digital commerce pillar Kristin is leading.

52ideas surfaced
5functional pillars
7strategic themes
May 2026workshop date
CA PI ED OE BP LI-COR + ONSET
52
Ideas generated by the team across all functional areas
7
Cross-cutting strategic themes that emerged from the synthesis
8
Lighthouse opportunities identified as highest-leverage starting points
3
Implementation horizons: Now (0–3 mo), Next (3–9 mo), Later (9+ mo)
Executive Summary

What the team told us, in one paragraph.

The LI-COR + Onset team sees AI as the lever to move from individual-level adoption to enterprise-level capability — and to make the digital commerce experience match the precision of the instruments behind it. Across 52 ideas, three patterns stood out: the website acting as a salesperson (proactive outreach, compatibility mapping, an AI advisor, a responsive purchase pathway), turning proprietary depth into a moat (73K research papers, instrument service history, internal documentation — all activated as AI inputs), and operating at the speed Kristin wants (collapsing 8-hour product pages, automating quote and order entry, unifying data that today lives in personal hubs).

The strongest opportunities aren't technology projects in isolation — they're places where AI unlocks something the team has been wanting to do for a long time. This report organizes those opportunities so leadership can decide where to invest first.

Strategic Themes

Seven cross-cutting themes emerged from 52 raw ideas.

When we clustered the workshop output, the same big ideas kept showing up under different functional banners. These are the strategic threads worth pulling.

Proactive Customer Engagement
8 ideas · Customer Acquisition

From personalized outreach based on what a customer already owns, to anonymous-visitor retargeting and SMS — the team wants the relationship to start before the buyer ever opens a chat window.

The Website as Salesperson
7 ideas · Customer Acquisition

Kristin's exact framing. Compatibility mapping, an AI advisor that coaches and educates, a customer-responsive purchase pathway, improved search, checkout UX that adapts to who's standing in front of it.

The Instrument as AI Surface
5 ideas · Product Innovation

An AI agent on the instruments themselves — built-in support, data interpretation, education, and direct purchase of consumables & services. Plus quick access to complete service-life history and triage to subject-matter experts.

Proprietary Knowledge Activation
6 ideas · Brand & Mission

The 73,000-paper citation corpus is a moat. Mining publications and funding opportunities, auditing content for combined brand position, competitive analysis agents, citation-driven positioning. Depth becomes leverage.

Unified Data & Decision Intelligence
7 ideas · Operational Excellence

Kristin's "AI is being adopted individually, not enterprise-wide." AI-ready datasets, integrated info access, financial summarization, SRO analysis, in-house knowledge from multiple sources — the unification layer Tom and Kristin are already building toward.

Workflow & Ops Automation
10 ideas · Operational Excellence

ISO routing & approvals, Tuesday meeting slide generation, supply-chain sourcing, order entry through form builders, requirement & test-criteria development from product-ideation transcripts. The "give-everyone-an-extra-half-day-a-week" theme.

Enterprise Learning & Enablement
8 ideas · Employee Development

LI-COR University for career development, just-in-time learning replacing "laborious onboarding reading sessions," internal product training, interactive instrument-use training, easier access to internal business information.

Top Voted Use-Cases

What the team chose to bet on.

After generating 52 ideas, participants voted on the ones they most wanted LI-COR + Onset to pursue. These seven rose to the top — and a clear pattern emerged about where the team's energy is most concentrated.

01
Customer Acquisition & Retention

Proactive, personalized outreach based on purchase history — identify customers who bought "X" and surface upgrades or things to consider next.

02
Customer Acquisition & Retention

Map compatibility and solutions to match customers in their journey.

03
Product & Service Innovation

An AI agent on the instruments themselves — built-in support, data review and education, and the ability to purchase consumables, professional services, etc.

04
Customer Acquisition & Retention

A personal AI advisor on the web that coaches and educates.

05
Brand, Market & Mission

Mine and find research publications and funding opportunities that shape future product directions and sales strategies.

06
Customer Acquisition & Retention

A customer-responsive purchase pathway — the website behaves like a salesperson.

07
Customer Acquisition & Retention

Improved search across e-commerce and the website.

The pattern

Five of the seven top-voted ideas live in Customer Acquisition & Retention. The team's energy is pointed squarely at the commerce experience — the website, the buyer's journey, and the proactive moment that turns a past purchase into the next one. This validates Kristin's framing exactly: the workshop is the commerce pillar of an enterprise AI strategy, and the room agrees on where to start.

Two ideas reach beyond pure commerce — the AI agent on instruments (which extends the customer relationship into the product itself) and research-publication mining (which turns the 73K-paper corpus into product & sales leverage). Both are differentiators no off-the-shelf tool can replicate, and both deserve their own track even as commerce takes the lead.

Vote distribution across pillars
Customer Acquisition · 5 votes Product · 1 Brand · 1
Prioritization

Impact vs. effort: where the leverage really sits.

Each dot is a strategic theme, placed by the value it can unlock and the effort required to get there. Hover for detail. Quick Wins are where to start; Strategic Bets are where to plan.

Quick Wins
Strategic Bets
Fill-ins
Reconsider
Effort & Complexity →
Low
High
Low
High
↑ Impact
Proactive Customer Engagement
High impact, modest effort — data is in CRM & order history already
Workflow & Ops Automation
Slide gen, ISO routing, order entry — quick to deploy across staff
Unified Data & Decision Intelligence
Foundational — replaces personal GenSpark hubs, unlocks everything else
The Website as Salesperson
Kristin's signature vision — meaningful build but compounding payoff
The Instrument as AI Surface
Transformational — multi-year, deeply differentiated
Proprietary Knowledge Activation
The 73K papers as a moat — needs ingestion work but no peer can match
Enterprise Learning & Enablement
LI-COR University & JIT learning — fast to start, internal payoff
Proactive Customer Engagement
Workflow & Ops Automation
Unified Data & Intelligence
Website as Salesperson
Instrument as AI Surface
Proprietary Knowledge
Enterprise Learning
Lighthouse Opportunities

Eight ideas worth a deeper look.

If leadership only acts on a handful of opportunities in the next year, these are the ones we'd build a business case around first. Each is anchored in a real idea the team named — six of them appear in the top-voted seven, plus two strategic stretches Kristin flagged on the call.

— Lighthouse 01 · Canvas focus
A dynamic, role-aware page builder

A tool that renders a custom page and value proposition dynamically based on who the visitor is — new end-customer, returning lab manager, distributor, or internal customer-service user — and what they're trying to do. Pulls from every content source LI-COR + Onset owns (inventory, specs, product attributes, sales trends, customer metadata) so the same shop URL feels different to every visitor. This is the use case the team chose to take through the AI Canvas.

High impact Mid+ effort Now → Next
→ View AI Canvas Deep Dive
— Lighthouse 02
Proactive, personalized outreach (#1 voted)

Agents that scan order history and identify the right "next conversation" with each existing customer — a consumables refill, a related instrument, a relevant new release. The highest-voted idea in the room, and a natural sibling pilot that runs alongside the dynamic page builder.

High impact Low effort Now (0–3 mo)
— Lighthouse 03
Compatibility & solution match engine (#2 voted)

A researcher describes their study and the system returns the ideal instrument configuration, with compatibility validation against eddy covariance, HOBOnet, and the rest of the catalog. Solves "what fits with what" without a sales-engineering call — and feeds directly into the role-aware page builder.

High impact Medium effort Next (3–9 mo)
— Lighthouse 04
AI agent on the instruments themselves (#3 voted)

Built-in support, data interpretation, education, and direct purchase of consumables & professional services — turning every deployed instrument into a relationship surface. The most differentiated long-term play and the one no commerce platform can replicate.

Transformational High effort Later (9+ mo)
— Lighthouse 05
Research publication & funding mining (#5 voted)

Mine the 73,000-paper corpus continuously to surface emerging applications, funding opportunities that shape product strategy, and citation trends that refine positioning and messaging. The team's depth becomes a strategic asset, not a marketing footnote — and a source the dynamic page builder can cite.

High impact Medium effort Next (3–9 mo)
— Lighthouse 06
AI-enhanced site & e-commerce search (#7 voted)

Semantic + technical search that handles scientific terminology, unit conversions, and application-first queries. Table stakes but foundational — every other commerce idea works better when search works. Fastest near-term win to ship.

Medium impact Low effort Now (0–3 mo)
— Lighthouse 07
Autonomous product-page generation

Kristin's "8 hours per page × 50 products" pain. A multi-agent pipeline ingests spec sheets, manuals, and photos and produces complete, SEO-ready product pages — copy, tech specs, applications, cross-sells, citations — in minutes. The backstage complement to the dynamic page builder.

High impact Medium effort Now (0–3 mo)
— Lighthouse 08
Unified commerce data agent ("Ask the Business")

The enterprise version of what Kristin is building individually in her GenSpark hub. A natural-language agent connected to pipeline, orders, inventory, catalog, and web analytics. Answers commerce questions in plain English — and lays the foundation for the broader enterprise AI build-out.

Foundational High effort Next (3–9 mo)
← Back to Lighthouse Opportunities
Canvas Deep Dive · Lighthouse 01

The Lighthouse idea the team chose, under the microscope.

After generating the use-case list, the team chose to take "a tool that responds by application and builds a custom page by pulling from all content sources" through the AI Use-Case Canvas — a structured exercise that pressure-tests an idea across thirteen dimensions before resources get committed. Here's what surfaced, in the team's own words.

High-Impact AI Use Case
A dynamic, role-aware page builder — render a custom page and value proposition based on who the visitor is and what they're trying to do.
POC 6–8 weeks MVP 10–12 weeks Mid+ cost High impact
Problem
What specific problem or challenge does this use case aim to solve?
The e-commerce experience is the same for everyone — new or returning, end-customer or partner — and customers don't always know where to start. The team wants the site itself to do more of the discovery work, on behalf of each visitor.
  • Help customers discover products more efficiently
  • The ecomm experience remains the same no matter if it's a new or a return customer
  • Customers are overwhelmed today and don't know where to start
  • Customers may overlook related products
  • Some customers know what to buy/replace — but some don't
Users
Who is the end user and how will they interact with the solution?
A multi-audience tool — the page renders differently depending on which kind of user lands on it.
  • End-customers
  • Distributors and partners
  • Customer services
  • Professional Services
Data
What input data is needed for this use case?
The five data sources the team named — combining what's known about each product with what's known about each customer.
  • Inventory and specs
  • Product attributes (existing and missing)
  • Past sales trends
  • Product taxonomy
  • Customer metadata
Missing product attributes flagged as a gap that needs closing before the page can render reliably.
Integration
Where does the use case fit within the existing infrastructure and how will humans interact with it?
Sits on top of the existing systems of record — Salesforce on the customer side and Drupal on the content/web side. The dynamic page is rendered into the Drupal stack; the personalization signals come from Salesforce.
  • Salesforce
  • Drupal
Risks
Any compliance, reputational or competency risks to address?
  • Downselling by accident
  • Incorrect recommendations
  • Tech integrations and connectivity
Stakeholders
Who are the people or units impacted by this use case?
A broad stakeholder map — every customer-touching function plus product. Strong cross-functional ownership will be the make-or-break.
  • Customers
  • Partners / Distributors
  • Customer Services
  • Sales / Marketing
  • Ecomm
  • Product
Value Prop
How does this use case add value to the business?
Two clean headline outcomes the team named — better customer experience, paired with measurable commercial lift.
  • Better CX
  • Conversion rate and revenue
AI Solution
What types of AI (GenAI, NLP, ML, CV) are being applied?
A multi-component AI stack — content indexed in a vectorized DB and retrieved via multimodal RAG, LLMs assembling the page copy and value prop, and a recommender driving the product mix. This is meaningfully beyond GenSpark territory.
  • Vectorized DB → Multimodal RAG
  • LLM Integrations
  • Recommender Systems
  • Data pipelines for reporting
  • Other integrations
Flow noted by team: Vectorized DB → Multimodal RAG
Resources
What human, data, infrastructure or technical resources are required?
  • Tech team (Drupal)
  • Ecomm team
  • Data / Reporting team
  • Sales Science / Technical team
Challenges
What are potential roadblocks?
  • Localization
  • Existing tech stack / integrations
  • Access to customer metadata
  • Competing priorities
  • Access to institutionalized knowledge / memory
  • Buy-in on reliability
  • Customer adoption / acceptance
  • Adaptability
Metrics & KPIs
How will success be measured?
A four-metric scoreboard the team named — three commercial signals (conversion, attachment, AOV) and one operational signal (call-centre volume) that tells you whether the page is doing its job of helping customers self-serve.
  • Conversion rate
  • Attachment rate
  • AOV (average order value)
  • Call Center volume
Costs & ROI
What are the costs of development and maintenance? When can we expect an ROI?
The team sized this as a Mid+ investment — not a quick-and-cheap pilot. The build touches Drupal, Salesforce, content infrastructure, and a vector + RAG layer, with real engineering effort across Tech, Ecomm, Data, and Sales Science. ROI will be tied to the conversion / attachment / AOV scoreboard above.
  • Mid+ cost envelope
  • Engineering effort across four named teams
  • Infrastructure: vector DB, LLM & RAG layer, recommender system
  • ROI tracked via conversion, attachment, AOV & call-centre deflection
Timeline
What is the expected timeline for implementation?
A clean three-phase rollout the team agreed on — proof-of-concept first, MVP next, then expansion.
  • POC: 6–8 weeks
  • MVP: 10–12 weeks
  • Expanded product: 14 weeks onwards
A note on this canvas. The thirteen fields above were filled in by the LI-COR + Onset team during the workshop — captured here in their own words. The next move is to take this canvas into a POC scoping conversation with the Tech (Drupal), Ecomm, Data, and Sales Science teams — with Kristin and the enterprise AI team as the executive sponsors. A 6–8 week POC is the explicit first checkpoint.
Implementation Roadmap

A practical sequencing: Now, Next, Later.

Built to balance momentum with foundation-laying. The Now bucket gives the team early wins and confidence. Next builds the platform layer. Later is where the visionary bets — the instrument-as-AI-surface — land.

01
Now · 0–3 months
Build momentum
Low-effort, high-trust wins — plus the POC for the canvas pick.
  • Dynamic page-builder POC (6–8 weeks) — the canvas focus
  • Proactive personalized outreach pilot (#1 voted)
  • AI-enhanced site & e-commerce search (#7 voted)
  • Autonomous product-page generation (Kristin's "8-hour-page" pain)
  • Tuesday meeting slide generation & financial summarization
  • Internal information access agent (replaces multiple GenSpark hubs)
02
Next · 3–9 months
Build the platform
Foundations that unlock the bigger ambitions.
  • Dynamic page-builder MVP (10–12 weeks) + expanded product (14 weeks +)
  • Compatibility & solution-match engine (#2 voted)
  • Unified commerce data agent — "Ask the Business"
  • Research publication & funding mining (#5 voted)
  • JIT learning & LI-COR University career-development pathways
  • Customer-services & partner-facing rollout of the page builder
03
Later · 9+ months
Build the future
The transformational bets — multi-year ambitions.
  • AI agent on the instruments themselves (#3 voted)
  • Full instrument service-life history at every touchpoint
  • Agentic quote & configurator across eddy covariance + HOBOnet
  • Enterprise data unification across pipeline, orders, & analytics
  • AI-forward product positioning as a market differentiator
  • Real-time customer success signals from deployed instruments
All Ideas

Every idea the team surfaced.

All 52 ideas, preserved in the team's own language. Filter by pillar to see what came up where.

Filter:
What's Next

Five concrete moves for leadership.

Suggestions for turning this report from a document into a decision.

1

Green-light the dynamic page-builder POC in the next 30 days

The team chose this use case for the canvas — and the canvas above gives you a 6–8 week POC plan. Stand up a small Tech + Ecomm + Data + Sales Science working group, pick the first audience to personalize for (returning end-customers is the cleanest first cut), and use the POC to prove the ROI before committing to MVP.

2

Name an AI champion across each commerce function

Workshop participants are the earliest believers. Name one champion per function — marketing, sales, customer success, e-commerce ops, IT — so adoption isn't bottlenecked at Kristin or Tom, and so wins compound across the org.

3

Commission a unified-commerce-data scoping sprint

This is Kristin's strategic unlock — moving from individual GenSpark hubs to a shared, governed enterprise capability. A 2–3 week scoping engagement now sets up the platform investment in the Next bucket without delaying anything in the Now bucket.

4

Frame the instrument-as-AI-surface vision externally

This is the long-term differentiator — and the funder, partner, and investor story. Even before implementation, the ambition itself is a positioning asset. Build a one-pager. Start the conversations with Battery Ventures and the rest of the board.

5

Set a 90-day check-in

Revisit this report with the team in three months. What landed? What didn't? Which ideas surprised you? Treat this as a living artifact, not a one-time deliverable — and the natural moment to scope the next engagement.