.png)
Your Sales Pipeline Is Lying to You
Reps don't update the CRM, your forecast is optimism in a spreadsheet, and your board wants a real number. AI just changed how fast you can fix all three.
A RevOps War Story
The Problem
I’ve worked with healthtech companies that run the core operations of their business on Google Sheets. We’re not talking about a small shop either. This was an established company, with day-to-day operations living in spreadsheets, while the commercial deal motion runs through their CRM that grew messier every year. When we audited that CRM, what we found was shocking.
The setup made it impossible to calculate foundational counts such as churned customers, active customers and revenue by rep. Surprisingly, this is extremely common. On the surface, you’d think running a sales team is straightforward: get the prospect > convert them to a paying customer > done.
So why is this so hard to map correctly into a software system? Are these tools unnecessarily complex?
It’s Never the Tool
Like all data problems, it rarely comes down to the tool. People and processes create the bottlenecks. Similar to Conway’s Law, a CRM usually mirrors the organization and adoption habits of the people and processes inside a company. And when the process is broken, the pipeline lies.
The Confession
Dashboards Aren’t Enough Anymore
As foundational LLMs continue to strengthen, I noticed a massive shift in the market in early 2026. As a boutique AI consultancy, selling Business Intelligence systems and dashboards is no longer enough. Data warehouses and dashboards are still necessary in enterprises and mid-sized businesses, especially in industries like healthcare with heavy regulation, strict data security needs and metrics that have to be 100% correct when it comes to clinical outcomes.
However, many early stage healthtech startups and scale-ups aren’t in a position to consider data warehousing yet. Their platform database is often enough for now. What they need urgently is a way to find where they’re leaking revenue, and analysis on where to focus to earn more of it.
What Startups Actually Need
Especially for healthtech SaaS and subscription businesses, a solid Revenue Operations and CRM implementation cures the pain sales teams feel when they don’t know how the company and team are actually performing in the market. Another buzz term I’ve been hearing lately is Go To Market (GTM) AI Engineering. The catch is that this is challenging to pull off. Once it’s set up though, the intelligence gained from quality data can be exponential for a business’s growth.
Why Now
The Disciplines Are Converging
As Joe Reis frequently talks about, the disciplines of software engineering and tech are converging. What used to be separate camps (software engineering vs. data engineering) are blending together. AI is lowering the skills barrier, which creates more robust solutions that pull from multiple disciplines to solve a problem faster.
Data That Drives Action
The goal now is data that produces an action. Historical reporting on a dashboard is the bare minimum. What can we do with data, software and AI engineering that tells a CEO to act now? Data that powers entire workflows can replace a task core to a company’s product or service. It’s insane what AI is unlocking in the healthtech space. I’ve been talking with founders building centralized data layers directly on platforms like HubSpot and ChurnZero instead of a traditional data warehouse, and they’re using Claude Code to build it themselves in record time.
Is it textbook Kimball modeling with perfect entity resolution and history tracking? Not always. But I’m not seeing the negative impact for these founders (at least not yet). They’re building, deploying and scaling quicker than ever, and getting to decisions faster in high growth mode.
Revisiting the RevOps Problem
The convergence set the stage. It opened my mind to the discipline of RevOps and GTM Engineering. Staying on the traditional RevOps side for a moment, let’s go back to the company from the beginning.
What Is a Deal, Anyway?
How is a healthtech company able to count its deals when it hasn’t defined what a deal is in its process?
Confusion often starts when you don’t separate a lead from a deal (opportunity). Is this prospect really an opportunity, or should they still be considered a lead? What determines when they convert to a qualified opportunity?
Many sales teams can’t answer these questions. Then they implement those loose rules in their CRM, and it turns into a complete mess.
When Leads and Deals Blur Together
The count of leads blends into the count of deals. Now you’re scratching your head over which records belong in the pool of leads vs. the pool of deals. There’s no field to flag which is which, and that’s a direct result of not defining it conceptually before mapping it into the CRM’s data model. If mapped correctly, these fundamental metrics would be easy to calculate.
However, getting sales teams to slow down and implement correctly is a huge challenge. They just want to do what they do best: CLOSE.
These are simple examples, but I’ve seen WAY more complex cases, like determining which customers are part of a certain line of business, or when to create a deal based on complex contract specifics. Now when investors ask how many active customers you have, or how many leads vs. deals, you can’t answer correctly or quickly. Yikes!
How We Fix It
Steinert Analytics serves healthtech startups in exactly this capacity. First we assess and define a roadmap to either correctly implement or re-architect what’s broken in your RevOps process. Then we implement and support you through adoption. Getting a handle on these numbers shows you where your sales process is losing revenue, so you can optimize, take action and drive more growth. Most of these engagements involve Salesforce or HubSpot.
Aiming for the Future of Sales as a Boutique Consultancy
Maybe you can tell, but we’ve been hard at work diving into RevOps head on. Everything above is traditional RevOps on a mainstream enterprise CRM. But there’s a new monster in the market, a sub-discipline showing up on RevOps teams.
Enter GTM AI Engineering
GTM data platforms are common in enterprises, but they rarely fit the startups and healthtech scale-ups in our Ideal Customer Profile. A startup seldom has enough marketing data sources to justify a data warehouse, let alone a CRM set up robustly enough to track accurate data like the problems described earlier.
GTM AI Engineering is a different thing. I recently had a coffee chat with one of the most elite CEOs I know. We’ve been connected for years, and he runs a GPS IoT logistics startup. I know it’s not healthcare, but the principles hold for any healthtech subscription product or service.
He’s a mega AI adopter, and his market insights blew my mind. Rather than keeping his sales team on Salesforce, he did something entirely different.
Off Salesforce in 3 Days
Using Claude Code, he migrated the entire company off Salesforce into a more integration and AI-friendly CRM in 3 days (way cheaper too!). From there, agents pull every rep’s call transcripts, calendar meetings and emails, and give him daily updates on every deal.
The Agent Does the Nagging
Each morning, an agent flags exactly which deals a rep has let go stale and drafts the email telling them. He approves it and it sends. The nudge comes from an objective system, not from the boss breathing down their neck, and adoption follows.
A Forecast Built to Be Beaten
His forecasting is just as sharp. He’s tuned the model to be deliberately pessimistic, scoring every deal from the actual conversations rather than its pipeline stage. A deal he’d personally call done gets booked at a discount for risk, and his team has beaten the forecast every quarter since.
He also has a marketing analyst using the same Claude Code setup to build forward-looking reports on deal activity: what may close in Q4, future projections, why a deal may not close, how to handle the stakeholders on a specific deal, and how to speed up sales velocity. Truly next level stuff, built in record time on a Claude Max plan.
The Economics
Theoretically, an AI GTM/RevOps system like this can cut a 30-person sales ops organization to 5 people. A $27K CRM bill drops to around $5K. Those are rough numbers, but you get my point.
The Honest Caveat on a GTM AI System
The Risk of Metric Sprawl
One thing I challenged this CEO on was metric sprawl. If analysts use non-deterministic AI to quickly build sales intelligence reports, metrics that should match can end up calculated slightly differently. He agreed.
Lock the Definitions, Flex the Precision
Here’s how I see it. The definitions have to be locked: what a deal is, what an active customer is, and which metric the team is judged on. That’s the entire lesson of the war story above. Once those are nailed down, small differences in how a report calculates them matter far less at a startup. Sales revenue may differ slightly from finance’s number, but I’d bet this AI system is pretty darn accurate, and that’s enough to fly: get the insight, decide with confidence, act faster.
In the enterprise, mid-sized businesses and regulated industries, metric clarity defined across the org in a data warehouse is non-negotiable. For startups scaling quickly, it’s a much smaller pain.
How Steinert Analytics Is Responding to the Market Shift
The Unlock
These insights have been brewing in my mind for months. I just couldn’t put my finger on exactly what I was working toward until a few days ago. The direction has always been to move data closer to revenue, and this was the real unlock.
The Announcement
We’ve already taken on clients for traditional CRM implementations, migrations and re-architectures, and we’ve brought on certified HubSpot engineers to help scale this offering.
We also just brought on a proper GTM AI Engineer to help early stage healthtech scale-ups run leaner, faster and more confidently with their sales teams. We’re running on the cutting edge of RevOps and GTM AI Engineering, and this is the formal announcement.
Steinert Analytics’ managed analytics services for healthcare companies remain a core offering. We’ve simply extended into RevOps as well.
Side note: I recently gave an Innovate New Albany TIGER Talk on why so many data teams and initiatives fail. It goes deeper on the people and process side of this problem if you want more.
I’d love your feedback. And if you’re a healthtech sales leader and this resonates, book a Pipeline Velocity Audit here and let’s find out if your pipeline number is real.
Follow me on LinkedIn for more on RevOps and Data Strategy.
Keep crushing it and changing patient lives in the healthcare space. Appreciate the time, thank you.
Comments