Welcome to the Sturdy Blog
News and Resources
The latest from Sturdy — product news, insights, and resources.
Sturdy's MCP Server: One Call. Every Source. Already Resolved.
Another Step to Unlocking AI Outcomes: Resolve the Data First
The bottleneck is not your AI model. It’s the data it has access to. Sturdy’s MCP server delivers pre‑resolved, canonically organized context so your LLM can reason over it instead of guessing around it.
Another Step to Unlocking LLM Outputs: Resolve the Data First
For years, the problem was that data lived in silos. Different systems for sales, support, and calls. But the worst offenders were email and Slack. Email isn’t one silo; it’s as many silos as there are people on your team. Every rep, every CSM, every exec running their own inbox, none of it visible to anyone else. Slack is no different. Conversations buried in channels and DMs that nobody ever sees again.
What Changes
"Your LLM now has a single, usable data layer any user can query to inspect the full context of every prospect and customer."
“Every team now works from a single view of the relationship, not fragments of it. Sturdy gets everyone on the same page, no matter what screen they use.”
MCPs were a material step forward. They give LLMs a standardized way to reach outside their context window and pull live data from external systems without a human copying it in manually. An account record, an open ticket, a call summary, all accessible at query time without a custom integration.
Today, teams are dealing with a different version of the same problem. Every MCP server exposes a slice of the picture. The LLM can pull structured records, read a ticket, or fetch a call summary. What it cannot do is answer a question that requires all of them at once, because the data across those systems was never resolved against each other.
The entities don’t match. The timeline is fragmented. The thread that started the conversation often isn’t there at all.
The question every revenue team actually needs answered isn’t “what does this system say about the account?” It’s the question that requires the full picture: what has every person at our company said to every person at this company, across every channel, and what does that tell us about where this relationship actually stands right now.
No single MCP server can answer that. Most LLMs, handed raw data, will approximate an answer and present it with false confidence. That’s not intelligence. It’s a good guess.
That answer doesn’t live in any single system. It lives in the relationship between all of them. And if the LLM has to call multiple MCP servers to piece it together, resolve duplicate records, and reassemble a coherent account state on every query, the fragmentation problem hasn’t been solved. It’s just been moved into the inference layer.
What Sturdy’s MCP Does
Sturdy ingests from all of it. Email, call transcripts, support tickets, Slack, CRM, and meeting tools. Every channel where communication happens.
Before any of that reaches an LLM, Sturdy does the work that makes it usable. Entities are deduplicated and matched to canonical records. Interactions are classified. Signals are enriched, permission‑scoped, and source‑referenced. The relationship between interactions across systems is established once upstream.
Not inferred at query time. Resolved in advance, maintained continuously, and auditable.
That last part matters more than it sounds. LLMs are getting better at fuzzy matching, but revenue decisions cannot rely on it. “Probably the same account” is not good enough when you’re making retention calls, forecast commits, or expansion bets.
Then Sturdy exposes all of it through a single MCP server. One call. Pre‑resolved context with citations. The LLM starts from the signal, not the raw material.
The Token Cost Nobody is Talking About
There’s a practical consequence to raw MCP that most teams haven’t priced in yet. When an LLM has to reconstruct account context from scratch on every query, it burns tokens doing work that shouldn’t need to happen at query time.
Pulling from multiple sources. Resolving conflicts. Traversing relationships. Figuring out what it’s looking at.
At low volumes, this is invisible. At scale, it isn’t. The rediscovery tax on a raw MCP call runs roughly 60 to 80 percent of total token consumption per query. That’s the LLM figuring out context, not reasoning over it.
Sturdy removes most of that overhead. The context arrives already structured. The LLM starts from a position of knowing. The inference budget goes toward answering the question, not reconstructing the data.
What This Means for Teams Building on it
Sturdy’s MCP is designed for teams that have already provisioned an LLM and are now trying to make it useful. CTOs deploying models across their organization. Heads of Data and AI are trying to get real answers out of them. Operations teams are building agents that need reliable account intelligence.
The properties that matter:
Canonically resolved
Entity deduplication and matching happen upstream. The same account appears as one account regardless of how many systems it lives in.
Permission‑aware
Access controls are baked into the data layer. What a user can see reflects what they’re authorized to see in the source systems.
Source‑referenceable
Every signal comes with a citation. When something surfaces, the underlying interaction is linked.
Model‑agnostic
The data layer doesn’t change based on which model you use.
Nobody wants to spend 12 to 18 months normalizing data before they can build something useful. Resolving that data upstream changes what your LLM can do on day one.
Talk to us about connecting Sturdy to your existing AI deployment.
Another Step to Unlocking AI Outcomes: Resolve the Data First
The bottleneck is not your AI model. It’s the data it has access to. Sturdy’s MCP server delivers pre‑resolved, canonically organized context so your LLM can reason over it instead of guessing around it.
Another Step to Unlocking LLM Outputs: Resolve the Data First
For years, the problem was that data lived in silos. Different systems for sales, support, and calls. But the worst offenders were email and Slack. Email isn’t one silo; it’s as many silos as there are people on your team. Every rep, every CSM, every exec running their own inbox, none of it visible to anyone else. Slack is no different. Conversations buried in channels and DMs that nobody ever sees again.
What Changes
"Your LLM now has a single, usable data layer any user can query to inspect the full context of every prospect and customer."
“Every team now works from a single view of the relationship, not fragments of it. Sturdy gets everyone on the same page, no matter what screen they use.”
MCPs were a material step forward. They give LLMs a standardized way to reach outside their context window and pull live data from external systems without a human copying it in manually. An account record, an open ticket, a call summary, all accessible at query time without a custom integration.
Today, teams are dealing with a different version of the same problem. Every MCP server exposes a slice of the picture. The LLM can pull structured records, read a ticket, or fetch a call summary. What it cannot do is answer a question that requires all of them at once, because the data across those systems was never resolved against each other.
The entities don’t match. The timeline is fragmented. The thread that started the conversation often isn’t there at all.
The question every revenue team actually needs answered isn’t “what does this system say about the account?” It’s the question that requires the full picture: what has every person at our company said to every person at this company, across every channel, and what does that tell us about where this relationship actually stands right now.
No single MCP server can answer that. Most LLMs, handed raw data, will approximate an answer and present it with false confidence. That’s not intelligence. It’s a good guess.
That answer doesn’t live in any single system. It lives in the relationship between all of them. And if the LLM has to call multiple MCP servers to piece it together, resolve duplicate records, and reassemble a coherent account state on every query, the fragmentation problem hasn’t been solved. It’s just been moved into the inference layer.
What Sturdy’s MCP Does
Sturdy ingests from all of it. Email, call transcripts, support tickets, Slack, CRM, and meeting tools. Every channel where communication happens.
Before any of that reaches an LLM, Sturdy does the work that makes it usable. Entities are deduplicated and matched to canonical records. Interactions are classified. Signals are enriched, permission‑scoped, and source‑referenced. The relationship between interactions across systems is established once upstream.
Not inferred at query time. Resolved in advance, maintained continuously, and auditable.
That last part matters more than it sounds. LLMs are getting better at fuzzy matching, but revenue decisions cannot rely on it. “Probably the same account” is not good enough when you’re making retention calls, forecast commits, or expansion bets.
Then Sturdy exposes all of it through a single MCP server. One call. Pre‑resolved context with citations. The LLM starts from the signal, not the raw material.
The Token Cost Nobody is Talking About
There’s a practical consequence to raw MCP that most teams haven’t priced in yet. When an LLM has to reconstruct account context from scratch on every query, it burns tokens doing work that shouldn’t need to happen at query time.
Pulling from multiple sources. Resolving conflicts. Traversing relationships. Figuring out what it’s looking at.
At low volumes, this is invisible. At scale, it isn’t. The rediscovery tax on a raw MCP call runs roughly 60 to 80 percent of total token consumption per query. That’s the LLM figuring out context, not reasoning over it.
Sturdy removes most of that overhead. The context arrives already structured. The LLM starts from a position of knowing. The inference budget goes toward answering the question, not reconstructing the data.
What This Means for Teams Building on it
Sturdy’s MCP is designed for teams that have already provisioned an LLM and are now trying to make it useful. CTOs deploying models across their organization. Heads of Data and AI are trying to get real answers out of them. Operations teams are building agents that need reliable account intelligence.
The properties that matter:
Canonically resolved
Entity deduplication and matching happen upstream. The same account appears as one account regardless of how many systems it lives in.
Permission‑aware
Access controls are baked into the data layer. What a user can see reflects what they’re authorized to see in the source systems.
Source‑referenceable
Every signal comes with a citation. When something surfaces, the underlying interaction is linked.
Model‑agnostic
The data layer doesn’t change based on which model you use.
Nobody wants to spend 12 to 18 months normalizing data before they can build something useful. Resolving that data upstream changes what your LLM can do on day one.
Talk to us about connecting Sturdy to your existing AI deployment.
Our articles
Create a self-sustaining customer reference funnel
It’s impossible to overstate the value of customer references. Whether you have $1m or $100m in ARR, when your customers demonstrate how they’re using your solutions, future customers see themselves in these examples bringing to life the value of your offerings. A strong reference from a current customer is so powerful that it can even transcend fierce competition, the norm for most of us in the rapidly maturing B2B SaaS industry.
Unfortunately, the reality is that even the most enamored customers aren't likely letting anyone else know about you. In studies from two industries, only 10% of the “promoters” in NPS surveys actually referred profitable business. Plus, executing a customer reference program takes lots of discipline and resources. For these reasons, we decided that this is a problem worth solving.
It has taken us over a year but we’ve cracked the code and productized a scalable way to harvest more customer references. Fortunately, we didn’t need to look very hard to find the signals that lead us to believe a customer is referenceable. The answers were right under our noses all along in the day to day interactions that our teams have with customers. That’s right, customers are signaling their willingness to provide references, testimonials, positive reviews and the like every day.
Here’s how it works. We’ve built technology that detects items of importance like customer references (among other things) in customer-to-business communications (email, support tickets, video conferences, etc.). For example, when a customer responds to an email or support ticket with, “I can’t thank you enough --- you just saved me so much time! You’re the best!”, Sturdy will instantly recognize this as a potential reference signal, flag it, and alert the appropriate person or team. We’ve even set up integrations with Slack and Teams so when a potential customer reference is detected, Sturdy chirps the notification right into a #customer-reference channel. Say hello to a self-sustaining customer reference funnel.
Getting started is easy.
Set up takes less than an hour for most operations / IT teams. We installed the Sturdy Salesforce.com app from the AppExchange (this is in private beta at the moment and we are accepting new users weekly here). Next, we synced our customer success and support teams’ email accounts (we use Gmail but we have an Outlook integration as well) with Sturdy’s email ingestion API. Once connected to the data sources where customers communicate with us, the rest is really easy. Time to value is a matter of days and weeks. And, there is no significant change management required. Turn it on and let the machine run. The cherry on top is that the AI gets smarter with every customer message.
1. Log into Sturdy (if you have a Slack or Teams integration you can skip step 1) and select the “Reference” signal. Sturdy will immediately surface any customer communications that contain potential references. Screenshot below.
Below is an example of a customer reference signal that I found this morning. Note that I have a privacy feature enabled here that anonymizes the data for the purposes of privacy and compliance.. In this message, Henry Goldberg is effusive in his praise for the product and the level of service provided to him.
2. Next, Sturdy alerts our customer marketing team of a new potential reference. Upon receiving this signal, our team will gather some information about the account and the user and determine what type of reference we want to ask for (peer-to-peer, review, case study, referrals, testimonial, etc) and who will make the request. Here’s another example of a customer all but volunteering to be a reference. Based on the anonymized aggregate data of our B2B SaaS customers, we've found that >1.5% of all customer communications include a customer reference signal.
3. The final step is to reach out to the customer with your “ask”. Every team is going to have some nuance here. We use a couple of different “plays”. Our favorite is the progressive / multi-step “swag+” play. When our team receives a reference signal, our CS and support teammates are empowered to ask our customers for an address where we can send a care package of swag. A few days later, we let the customer know that their Sturdy swag is on the way and then we ask if the customer would consider serving as a reference. Our success rate when using this play is nearly 100%.
Creating a self-sustaining customer reference funnel starts with consistently detecting the right signals and getting your customers’ voice to the right people at the right time. These signals are pure gold to every customer marketing team. The best part is that unlike traditional, resource -intensive customer reference strategies, Sturdy makes it virtually automatic. Turn it on and let it rip. Get a steady stream of potential customer references delivered in Slack, email or via dashboard - however you choose.
Want to get more customer references?
If this sounds interesting, our enterprise beta program is in full swing. If you are interested in creating an automated customer reference funnel, here is a link to register for our public beta program.
Net dollar retention - a SaaS metric juggernaut
The SaaS industry is still roaring towards ubiquity. Blissfully’s 2020 SaaS Trend Report notes that overall spend per organization on SaaS-based products is up 50%. However, the report also notes that this is down from previous years, and the growth rate seems to be slowing. This gradual slide has the industry turning its attention to optimizing for customer retention and leveraging existing customers for substantive growth.
Anymore, churn is just SaaS slang. Churn as a metric is confusing and ambiguous. There are too many ways to calculate customer and revenue churn. Analysts and investors have been increasingly skeptical of churn rate calculations for years. Anymore they just want a raw data dump from companies so they can run their own math.
“There are too many darn ways to calculate churn. That makes it ambiguous.” - Dave Kellogg
The focus is on net dollar retention (NDR)?
NDR has emerged as one of the top SaaS metrics that matter. NDR takes into account upgrades, downgrades, and churn to quantify how much recurring revenue from current customers you retained across a defined period of time. There are two hugely important questions that NDR can answer.
- Is your product delivering the value promised during the sale?
- Are your customers growing with you or without you?
What makes net retention so powerful is that for most companies, it’s cheaper to sell to existing customers than to sell to new logos. This makes net retention the most cost efficient way to accelerate revenue growth.
Calculating net dollar retention.
If your NDR is over 100%, this means that an increase in revenue is attributable to your existing customers. Here’s how to calculate NDR.
(Starting MRR + expansion — downgrades — churn) / Starting MRR = NDR
Let’s say you start the month at $100,000 in recurring revenue (MRR). Over the month it added $25,000 in expansion revenue, has $10,000 in downgrades and another $5000 in churn. ($100,000 + $25,000 — $10,000 — $5000)/$100,000 = 110% NDR. Your MRR is $110,000 with an NDR of 110% This is good. Essentially, your upgrades / upsells lifted your revenue despite losses.
What good looks like.
At least 100% is considered a good NDR rate for SaaS businesses selling to the SMB market. Selling to smaller accounts naturally yields a lower NDR. SMB clients are less financially stable, ripe for acquisition, and have smaller budgets. A good enterprise NDR is 130%. As with many SaaS metrics there are other things to consider. For example, Workday’s NDR is 100% but gross retention is 95%. Either Workday is very good at selling the “whole” deal or their product footprint presents limitations.
Here are some examples of net dollar retention rates for some interesting SaaS and SaaS-enabled companies.
Caring about net dollar retention.
NDR provides a revenue-based view of customer retention. NDR is increasingly important as you scale from a small to medium-size business and beyond. For example, a $5m business that churns 20% can replace that $1m with net new business when it’s growing +50% a year. But when a $30m business needs to replace $6m this becomes insurmountable especially if the growth rate is slowing.
The effects of NDR compound with time. It’s either additive or punitive with every customer that you acquire. This means that small upticks in NDR can add up to very large differences in total revenue over multiple years. For example, assume a business had $10 in revenue last year and consistently generates 20% of revenue from new customers. Improving the NDR from 95% to 105% may sound meager, but over five years the business will gain another $5m in revenue.
Lifting NDR and a plug for Sturdy as a solution to help.
How can you start identifying more opportunities to grow and deliver value? Here are two ideas that sound great in articles and when delivered by panelists at conferences. First, hire a great team of CSMs who are well enabled and know your customers intimately. Second, develop more premium services to sell your customer base. Frankly, these are right answers but they take a lot of time, resources and change management to create an enduring impact.
Now consider this. What if you had a “tool” that could analyze customer emails, tickets and conversations for important signals that are typically related to predicting churn? Maybe something that can listen for suggestions about features and products that might accelerate value capture and lift revenue? What if you could get started with such initiatives without major upfront investments in data infrastructure or change the way your teams work? We might know of such a tool. Hit us up. We’d be just as happy to talk about NDR and our experiences over the years tracking this SaaS metric juggernaut.
The deck we used to raise money for Sturdy
The idea for Sturdy was born from asking this all too common question far too many times, “What is going on with Customer X?” And many times over the years we have griped, “How is it the 21st Century and we need to get 5 different people in a room to login to 5 different apps in order to know whether a customer is happy or not?”
This is why every SaaS company has a “Top Customer List”. At Newton, our previous company that was acquired by Paycor in late 2015, we had a rule, “Whenever someone on this list contacts us for any reason, let So-and-So know.” If you think about it, such lists admit a fundamental failure of running a modern business...you only have the time and resources to listen to your most valuable customers (which means you most often ignore the rest).
This was our first slide...
Our earliest decks talked about, “getting your data in one spot”. But that wasn’t the problem we were trying to solve (wanting to see all the data in one spot is a symptom, not a solution). The problem wasn’t really a communication problem, it was a mining and refining problem. When a customer requests a copy of her contract, that message must get forwarded to the Saves Team - immediately.
Our “Aha” moment was when we realized that our customers are telling us what they want and need everyday. They give us information to run our businesses better, to predict churn, to capture references, to get in front of renewals, to prioritize features, yet this data is trapped and decaying in dozens, if not hundreds of data silos.
A big problem is that our customers are giving us this information in Slack, Email, Salesforce, Webinars, Training Sessions, Zoom calls, etc.. And the only way we utilize this information is if someone manually identifies, records and escalates it.
Remember when we said it was the 21st century? We still manually identify, capture and route feature requests. And bug reports. And cancellation requests. And sometimes this means that we don’t always see the signal, or we forget to log it, or when we route it, no one pays attention.
But these signals are immensely valuable. For example, reducing churn from 10% to 9% in a $10 million ARR business means that every customer is worth $17k more in lifetime value (500 customers, $20k annual contract value). And reducing churn in this example is just saving 5 customers.
Obviously we should do everything possible to mine our customer communications, and yet many companies know more about their anonymous website visitors than their own paying customers. Almost every company has a way to track and monitor its website visitors, and almost zero have any way to monitor and monetize the happiness of their actual customers.
Here’s a challenge...Answer this: If your company was about to lose a customer, who would be the best person to save that customer? What metrics would you use to support your answer? Most companies have no data to answer this question.
Or, how many times did a customer say, “You guys are great!” last month? How many times were those happy customers converted to references? And how many of those references are delivered to your sales team to help them close new business?
Again, it's the 21st century. Yet we have no analytic capacity or automation as it relates to customer feedback or happiness. But don’t despair. You're not alone.
We realize the challenges are great. But in this area, failure is truly unacceptable. To have a truly operationalized customer focused company, you need to mine these communications, without bias and without manual data entry. You need something that never gets tired, that doesn’t need training, and that gets better the more you grow and the more you throw at it. And most importantly, you can’t wait until the quarterly business review is complete to triage a churning customer.
And that’s why we started an AI company. But not just any AI company and not just for the sake of using AI.
We aren’t here to reinvent and change the way teams or companies work. And that is what is so exciting about what we do. Sturdy is the force multiplier for your business. If you already have a cutting edge BI tool, we just give it better data. If you have a killer CX app, we make it more insightful. If you have a great Customer Success, Account Management, Operations, Marketing, and Product teams, we make them more efficient and provide them with better data.
Sturdy is open
Sturdy has developed a BI product that analyzes customer communications, detects important signals, and empowers teams with real-time data to act on situations with speed and intelligence.
We’re thrilled to announce the launch of Sturdy, a ground-breaking business intelligence platform that leverages advanced data science in order to detect items of importance in customer-to-business communications.
In simple terms, Sturdy helps people at B2B SaaS businesses leverage a data set that is hiding in plain sight - data that your customers want you to use.
Trapped in communication layers, and across teams, are critical signals like, point of contact changes, potential references, churn likelihood, and competitor mentions. These signals gather digital dust in email accounts, ticketing systems, transcriptions, chat software, and CRMs - until Sturdy.
Customer-to-business communication data is an untapped data frontier. Massive value is realized when the data is aggregated, analyzed, refined, and redeployed. Sturdy was created to empower teams to act on mission-critical situations with speed and intelligence.
If you wanted your team to capture 10 new referenceable customers, what would need to happen? Or, how many of your customers got a new Point of Contact last month? Which customers asked for their Renewal Data this week?
As a leader you want to manage risks and capitalize on opportunities (we call them “signals”). Signals are sitting in email accounts, videoconferencing transcripts, chat logs, and buried in ticketing systems. They are manually captured, if at all, and then data-entered into spreadsheets and other systems. And you have to create, enforce and constantly train people on rules that change the way your teams work.
Not to mention, there is no analytical capacity.
The idea for Sturdy came from building, bootstrapping, and scaling successful SaaS businesses. We founded SturdyAI to empower businesses to solve problems that we faced as entrepreneurs and executives. Before SturdyAI, the capture of these signals has been inconsistent, fragile and inefficient.
We’re experienced executives and engineers. We believe that every business has revenue and earnings potential trapped inside of its communications systems.
In mid-2020, Sturdy received an investment from Super{set}, a team that has created $1.2b in exits. This accelerated our product development and commercial efforts. Partnering with Super{set} was natural. We share the belief that “data is the new oil” and that refining data defines the new basis of competition across sectors and problem spaces.
Many of us worked together at Newton Software. This is a company that we bootstrapped, scaled and sold to Paycor, one of the largest independent HCM companies in the world. At Newton, we lived by some simple rules. We live by these rules at Sturdy.
- If you make a mistake, tell someone right away. We’ll fix it.
- We design technology that we want to use.
- We sell software how we’d want to buy it.
- We support our software the way we would want to be supported.
- We do things the right way, not the easy way.
- We don’t take shortcuts.
We’re energized and ready to roll. Let’s talk.
We’re encouraged by the feedback and results from early customers using Sturdy. And, we’re fired up to help businesses preempt customer issues before they spiral and seize revenue opportunities in time to improve this quarter’s results.
What will you find in your data?
Click here to get access and see for yourself.
