The CRM integration was always in the plan.

It wasn’t a late addition. Without it, the Weekly Index Score was just a number. With it, you could say: your market perception improved twelve points the quarter your renewal rate went up eight percent. That’s a product people pay for.

We just kept pushing it out.

[VISUAL: roadmap slide with the “Salesforce integration” milestone slipping quarter to quarter — the receipts]

Why It Kept Slipping

Across four beta customers, every path to Salesforce hit a wall:

  • IT needed a formal security review.
  • The Salesforce admin was buried under a hiring-freeze backlog.
  • Legal wanted sign-off on what data we’d read.

None of it was unreasonable. CRM data is sensitive. But four customers in, not one had completed the integration. We’d built the product around a loop nobody had closed.

[VISUAL: simple diagram — four customer logos, four red Xs on the “CRM connected” step]

The LMS Pivot

This is where having access to the codebase mattered in a way I didn’t expect.

Thought Industries was the LMS. Every customer was already emitting data we could query: course completions, feature adoption, engagement, certification pass rates. I could open the API, see what was there, and design around what actually existed — not around an assumed Salesforce connection we couldn’t close.

We didn’t write a spec and wait. We checked, then designed.

[VISUAL: side-by-side — left, the Salesforce data we wanted; right, the LMS data we already had. Same shape, different source.]

Instead of “your competitor is gaining ground on feature adoption messaging,” the product could say:

Learners in your advanced certification track are churning at onboarding — and your competitor is positioning heavily on that exact job.

Tighter story. We controlled both sides of it.

Two Products, One Repo

The pivot didn’t just change the UI — it changed the shape of what we were shipping.

We split the work into a monorepo with two products:

  • Customer Value Agent — the original. Market perception, Weekly Index Score, content generation. Now tied to LMS engagement.
  • Learning Value Agent — the agentic LMS itself. Native AI inside the learner experience, running on data every customer already had.

[VISUAL: architecture sketch — monorepo with two app entry points, shared infra below; arrow from LVA → CVA labeled “Salesforce unlocks ROI loop”]

The bet: LVA could land on its own with data customers already had. Then, once Salesforce was connected, CVA would close the ROI loop on top of it.

LVA was the foot in the door. CVA was the upsell.

Two products, one codebase — and a path to ship that didn’t depend on us solving the integration problem first.

What We Cut, What We Kept

[VISUAL: BEFORE — dense raw signal grid with eight categories, competitor confidence scores, the full JTBD breakdown]

Cut: the raw signal breakdown. Competitor confidence scores. The full JTBD grid power users had loved.

[VISUAL: AFTER — simplified LMS-tied view: Weekly Index Score, content generation, behavior-linked recommendations]

Kept: Weekly Index Score. Content generation. Recommendations — now tied to LMS behavior.

The honest tension: power users felt the product got dumber. New users finally understood what to do. Both were true.

Where That Left Us

Four beta customers, paying. Real, but fewer than expected. The CRM integration never landed. The LMS pivot was promising but recent. The ROI loop wasn’t proven at scale.

That was the state of things heading into the question of whether this became a real product.

This is part three of a four-part series on building the Customer Value Agent at Thought Industries. Part four: Four Paying Customers and a Product That Didn’t Ship.