Most B2B CEOs I speak with have the same story: They've bought the tools. Tried the pilots. Given their team time to experiment. And still, the AI output is generic, inconsistent, or worse, risky enough that someone has to double-check everything before it goes near a prospect.

That's not an AI problem. It's a foundation problem.

I'm Hannah, founder of TrustLeader and author of Lead with Trust. With 25+ years in B2B marketing and sales — and a background in computer science and marketing — I've spent the last three years doing one thing: helping growth-stage B2B CEOs turn their internal knowledge into a scalable, trustworthy AI engine.

The results speak for themselves. I've helped multiple B2B companies grow double digits year-over-year — including one 10x over 10 years. I built a content repurposing workflow that produces 20+ on-brand content assets per week in just 7–8 hours, reducing a full-time content function to a single day of focused work. My clients have won HubSpot's Impact Award for best client growth story and recognition as NJ Emerging Business of the Year.

The TrustLeader Method — Extract, Codify, Structure, Implement, Amplify — is a five-step system for turning what your best people know into a foundation that AI can actually use. Not another prompt library. Not a tool recommendation. A governed, quality-controlled system that lets your team ship AI-assisted work with confidence.

If your teams are using AI without shared standards, or you're watching costs rise while quality remains inconsistent, you don't need more experimentation. You need the foundation.

Start with the Safe to Scale AI Readiness Scorecard — a free assessment that shows you exactly where your AI-readiness gaps are and what to fix first.

Hannah Eisenberg

Peter Caputa

CEO at Databox

I help the leaders of professional services firms leverage data to get the respect, power and money they deserve -- in that order. (Sorry Lil' Kim, you had it backwards.) 

I have been building systems that help professional services firms get to the next level for nearly 20 years now. As the founder and 9-year leader of the HubSpot Solutions partner program, I helped thousands of marketing agencies increase recurring revenue and margins by teaching them how to sell and deliver a measurable and compounding ROI -- helping many go from their dining room table to multi-million dollar exits. 

As CEO of Databox, my team & I have established a system (software, training, support) that actively helps 1200+ professional services firms use data to acquire, retain and expand their client engagements.

I believe in giving value first. 

So, if you're a leader at a professional services firm, you can start by 

1️⃣ Leveraging us to get the RESPECT you deserve by taking 5 minutes to contribute a quote for some of our upcoming research articles (https://databox.com/become-a-contributor) or inquire with my team about our other free co-marketing programs, 
2️⃣ Leveraging us to get the POWER you deserve by taking our free course on how to leverage benchmark data in your marketing, sales and client services processes, or
3️⃣ Leveraging us to get the MONEY you deserve by taking our free course on how to leverage data to predictably improve your (and your clients') performance. 

Both free courses are here --> https://academy.databox.com/

Of course, if you're ready to explore how we can help you the most, start a trial of our analytics software. More than 1200 professional services firms use us to efficiently monitor, manage, report, benchmark and forecast performance for their clients -->  https://databox.com/analytics-software/agencies

I also write and network a lot on Linkedin. While my inbox is usually overflowing, don't be afraid to ask me for help.

At Databox, I am currently 100% focused on helping marketing agencies automate their reporting. But, thousands of companies have figured out how to use our software on their own. They use it to set numeric marketing and sales goals, auto-track progress by pulling performance data from 100s of other software systems and visualizing it on any device.  The mobile app is slick and we even have a free version that allows you to track performance from up to 3 different apps. www.databox.com

Stay tuned...

The word "recession" is all over the news. I searched it on LinkedIn today. Nothing real came up. An info marketer selling a course on making your business recession proof. Someone talking about a "trust recession" on TikTok.
 
In private, though? Completely different story. I've had 50+ conversations in the last 90 days with business owners and senior execs. Almost every one of them is absorbing some version of the same externally imposed chaos: losing clients in certain industries, budgets frozen, hiring paused, deals stalling, tariff confusion, higher material and energy costs, labor shortages in some places, job cuts in others.
 
This week, Jerome Powell used the phrase "we don't know" 14 times in one press conference. The U.S. shed 92,000 jobs in February. Economists are already calling it a "hiring recession." Inflation won't budge. Consumer confidence has collapsed to levels not seen since COVID.
 
Tariffs and the administration's trade policy created much of this uncertainty — and the legal battles around them mean nobody even knows what the rules are. The Iran conflict is sending oil prices through the roof. Consumer debt is at all time highs. The deficit keeps widening. Businesses aren't expanding because they can't see six months ahead. That's a compounding set of policy decisions causing direct harm to everyday businesses and the people in them.
 
And yet, scroll LinkedIn. Thought leadership. AI hot takes. "Excited to announce."
 
I think I understand why. Anything that touches policy feels like politics. So we go quiet.
 
I'll say it plainly, even if it earns me a "Trump Derangement Syndrome" diagnosis in the comments: when the decisions of our elected leaders are the direct cause of economic damage, staying quiet isn't apolitical. It's just avoidance.
 
And the economy isn't even the whole story. We're watching the free press get attacked, the DOJ get weaponized, corruption operate out in the open, death being celebrated, and the erosion of institutions most of us assumed were untouchable. These things don't fix themselves while we're busy projecting confidence on LinkedIn.
 
Collective silence doesn't protect us from any of it. It just means we face it alone, without the benefit of each other's thinking or support.
 
The most powerful people in business have already made their choice. Some are silent because they can afford to ride this out or fear being targeted. And some aren't silent at all — they're funding and shaping the very policies causing this chaos, optimizing for their own enrichment.
 
Meanwhile, the rest of us are left feeling helpless. In normal times, we spend our energy building things worth building: good jobs, happy customers, strong partnerships. We didn't sign up to absorb chaos we didn't create. But maybe staying in our lane isn't enough right now. Maybe we need to do more than just build. Maybe we need to start speaking up collectively and demanding better. 

Over to you...

Please help me with this. We debuted our AI Analyst today and are stuck in 3rd place among ProductHunt launches with 241 votes. 

It sure would be cool if we beat out Claude and Lightfield to take the top spot. 

I don't like to ask for favors and as a general rule, I don't. 

But, if I've ever done anything to help you out along my 49 years on earth, here's something you can do to return the favor. 

My team has worked their bums off to get this done. So, it would mean a lot to me if you spent 60 seconds to do two things. 

1. Leave a comment below 
2. Go over to product hunt and vote.  https://lnkd.in/evcpptqp

And if you want to learn more before you vote (You should-voting has consequences), check out this interactive demo: https://lnkd.in/emxuuisn

After today, in-depth data analysis can be done by anyone. It's no longer something that only developers, analysts and tech-savvy executives are qualified to do. 

For the last 10+ years Databox has helped companies monitor and report their performance data with the help of our many robust integrations with other tools, dashboards and reports. 

We designed our system to make the numbers visible and shareable.

However, performance management still required smart people looking at the data, interpreting the trends, comparing against goals, drilling down by different dimensions to investigate why changes were happening, etc.

Databox couldn't directly answer questions like: 

• Why did leads drop this week?
• Are we likely to hit our quarterly goals? 
• Which marketing channel's results actually improved or worsened?
• Which sales reps are crushing it and why? 
• And so on. 
  
The numbers were there, but the answers still required some digging, comparing periods, building new metrics, creating one more dashboard, or asking someone for more context.

Today, we’re launching something that changes that.

Genie —  an AI analyst inside Databox trained on your business.

Instead of digging through dashboards, you can just type into Genie's chat interface (like Claude or ChatGPT's interface):

• Why are leads down this week?
•  How is Q2 pacing against our goals?
• What changed in paid performance month over month?
• How are each of my marketing channels trending?
• Compare my rep's activity and performance to tell me which ones are crushing it and why.
  
Genie analyzes metrics, goals, and trends... then explains what’s happening in plain language. 

It can also build new metrics from raw data, and build or edit dashboards and reports from a simple prompt, like: 

• Create a metric for x from z software tool
• Add metric to x dashboard
• Find me a dashboard that shows y metric so I can add this new metric to it. 

It’s like having an analyst at your disposal 24/7. 

Genie changes how anyone can interact with data. It makes it simpler. Simple enough even for scatter-brained CEOs like me. :-) 

For years, analytics meant navigating dashboards & piecing together context. It usually required an analyst who understood your business & exactly what you were measuring how. 

With Genie, all you need to know how to do is type. 

Faster answers, clearer context. 

This means more time spent improving performance instead of assembling reports. It means more clarity for everyone. It means quicker decisions, easier delegation, more accountability. 

We’ve been using Genie internally for a few months, and it’s already changed how we run the business. It increases the speed at which we can act. 

If you want to see an interactive demo, here you go: https://lnkd.in/emxuuisn

My Linkedin posts have gotten: 5M Impressions, 45k reactions. 22k comments. Would anyone want my Claude project instructions for writing them? 
 
I write most of my posts myself. (Špela Jurič drafts some too.) They always start with primary,  original research.  But I use Claude to draft them. The process...
 
My team and I record calls with customers and partners. We create other long-form content too: research reports from surveys, podcasts, long-form articles. That's the raw material I feed into Claude. 
 
When I'm ready to write a post, I open a Claude Project I've set up specifically for this. I paste in a transcript or an article. Then I describe what I want: the angle, the audience, the one thing I want someone to walk away thinking. 
 
The most counterintuitive part: some of my best-performing posts in the last six months were directly about Databox.
 
Most people assume talking about your own product tanks your reach. In my experience, it doesn't, as long as the story is real & specific.
 
Three recent examples:
 
I posted about how HubSpot's MCP server doesn't include marketing analytics data, and that anyone can pull HubSpot's marketing analytics for free through the Databox's MCP instead. 87 reactions. 38 comments. 12k+ impressions. https://lnkd.in/euq4_8Gt
 
I shared a call with Grant Covault, a fractional CMO and Databox Partner, about how he sets clients up on Databox and then trains their internal team to own it. He told me: "𝘐'𝘮 𝘯𝘰𝘵 𝘨𝘰𝘯𝘯𝘢 𝘵𝘢𝘬𝘦 𝘮𝘺 𝘴𝘺𝘴𝘵𝘦𝘮𝘴 𝘸𝘪𝘵𝘩 𝘮𝘦. 𝘠𝘰𝘶 𝘰𝘸𝘯 𝘵𝘩𝘦 𝘴𝘺𝘴𝘵𝘦𝘮." 89 reactions. 136 comments. https://lnkd.in/e92bzHye
 
I wrote about how I miss opinionated software and why Databox still has a point of view while most platforms have abandoned theirs. It talks about the importance of our Predictable Scale methodology and how it aligns with our product features, like OKRs and forecast modeling. 70 reactions. 46 comments. https://lnkd.in/eGQ9tqUs
 
None of those were viral. But they consistently drive real conversations and real awareness. The trick isn't avoiding the product mention. It's making sure the story earns it.
 
I'd love to see more people write like this on Linkedin. 

My team built the first version of my Claude Project instructions. (Actually a custom GPT, but I moved to Claude recently.) They fed my top posts into ChatGPT, extracted the patterns & handed it to me. I've been using a version of it ever since.
 
I'm rebuilding it now, from scratch, myself. I have noticed that Claude is really good at asking lots of questions before doing anything. I love the multiple choice ones. My instructions verify angles, quotes, stats, etc before drafting. So, the new version asks me 20 clarifying questions before writing a single word. It's so much better. 
 
Anyone want me to share the instructions? If so, tell me what's most interesting to you below. If 50 people raise their hands, I'll clean up the instructions and share next week.

I talked to 6 more agencies this past week who are building systems that automatically analyze their clients data, spitting out insights and recommendations. 

Devin Littlefield is working with Rick Kranz & Max Traylor to bake in client communications, account management best practices and live performance data (via Databox's MCP server) into one system, enabling Devin to do quarterly business reviews on demand. 

Claudio Lira, PhD built a Claude Skill that uses the Databox MCP (https://databox.com/mcp) to pull ads, web, and CRM data for any two periods he chooses. Then it applies a fixed set of rules that returns exactly one answer: where the growth system is breaking and what to focus on next. Learn more here: https://lnkd.in/eU-EeFmy

Rick Kranz at the AI Marketing Lab built a Claude skill that uses information about his client's ideal client profile to do a deep dive on a website's SEO using Google Analytics, Search Console and CRM data (from Databox's MCP) to identify very specific changes to make to a website's content, in order to attract and convert more of the right fit customers.  He actually made it a free download --> https://lnkd.in/eGSAVjzM

Benny Stapel walked me through his new system for sending a weekly client report with exactly what the agency did and the impact that work had on results. They're using n8n to pull completed tasks from ClickUp and performance data via Databox's MCP server. They used to manually do this report every other week and it only included the tasks they did. Their clients now better understand the impact of Unison's work and are way more engaged. Further, his team is using Genie, Databox's AI analyst (https://lnkd.in/emxuuisn) to ask questions on-demand, as well as build new metrics, dashboards and reports for clients as they execute. There's now much less of an analyst bottleneck when it comes to gathering insights and showing impact. Their weekly standups are more efficient and focused too. 

Gary Magnone, who does fractional ops work for agencies, just launched a new service helping agencies implement Databox and AI to help agencies better connect their work to client outcomes. https://lnkd.in/e7pPR57x

Steve James at Stream Creative has been using Databox's MCP in his day-to-day whenever he needs to prep for a call. He just logs in and asks it to analyze the client's account for improvement opportunities to come up with things he can suggest. 

If you're not building something like this in March, you need to start if you don't want to get left behind. 

These agencies are going to be able to analyze faster, spot issues and opportunities faster, act on them faster and ultimately: increase client impact and retainer sizes. 

A few months ago, a system like this would have been a differentiator. A month from now, it's going to be table stakes.

I have been separately talking to 10+ agencies about using AI to automate client performance analysis. Not just the visualization of the data, but actual analysis of the data.

I've initiated a few of them, but most of them are happening organically. 

Everyone has a bit of a different approach to it, using different tech. 

(Most popular tools: Claude & n8n, but also seeing ChatGPT and Make still.) 

Regardless of tech, here's the gist of what they're building: 

1️⃣ Gather client context into a file or folder. (Proposals, call transcripts.)
2️⃣ Connect to MCP servers to retrieve more context and performance data. Slack MCP, Clickup MCP for more context. Databox MCP server (https://databox.com/mcp) for performance data. 
3️⃣ Write out project instructions or create Claude Skills which describe how to do analysis. 

I have not seen agencies this excited about automation since I started at Databox in 2017. 

There seems to be 3 big motivations: 

1️⃣ Taking what their smartest people know about doing analysis with different types of datasets (ads, CRM, web analytics, social, SEO/AEO, etc) and baking that into a system that anyone at the agency can use to do thorough, quality analysis at any time for any client. 

2️⃣ Time savings: Agencies that are good at analytics spend a lot of time on it. This shortens that time from hours to minutes. 

3️⃣ Management insight: Not every agency owner pays attention to client performance. In fact, something I realized early on at Databox is that most agency owners simply rely on their account managers to make sure things are going in the right direction. But there are some agency owners that strive to monitor client performance closely. These are the early adopters of this tech. However, even for these owners, I don't think they'd do this if this system didn't also save the agency time and better empower team members. 

The ultimate value here is better, longer, bigger client relationships. But, I think most agencies will have to see that happen to believe it.

I caught the Linkedin follower odometer rolling over 37k this morning. 

Not sure what's more exciting to me: catching a screengrab of the "000" or hitting the actual 37k milestone. 

Either way, I'm so genuinely ecstatic that so many people love following along as I post what I'm learning and doing every day. 

I know that many people follow me because you're huge Databox fans, and some just follow me because I inspire you to improve your own use of the Linkedin platform.

Either way, it's nice to know I'm having some sort of positive impact on so many others. 

PS. Last week, 70 of you asked to see my Claude project instructions for writing Linkedin posts. I'm still planning to share that, but will need a bit more time to make them understandable. So, if you missed that opportunity, here is how to request access before I drop it: https://lnkd.in/e8zjwkY4

If you're not asking AI to analyze data, I feel bad for you son. 

I got 99 problems, but analyzing data aint one....

Wanna learn the safest and fastest way to use AI to do analysis? Join this webinar --> https://lnkd.in/eE77JEUz

Databox Genie examples

You're probably not bad at marketing analytics. You're probably just too slow at it and not doing it enough.

Because when it takes a few hours to pull the numbers together and walk through them thoughtfully, it slips to next week once something looks off or once someone asks for it. 

Rick Kranz at The AI Marketing Lab built a free Claude skill that fixes this: makes it possible to run thorough analysis quickly. 

He calls it his Weekly Growth Dashboard.

To use it, he goes into Claude and writes: "Run the weekly dashboard."

It connects to the Databox MCP Server, pulls 8 weeks of GA4, Search Console, and CRM data, compares two rolling 4-week windows, and delivers a full HTML dashboard with charts, sortable tables, and specific recommendations. 

Turns what would have been a few hours of hunting and pecking through tools and dashboards into a few minute exercise that happens while he grabs his coffee. 

And it doesn't just report numbers. Every recommendation gets filtered through his ICP before it reaches him. Content attracting the wrong audience gets flagged. Opportunities that match his ideal buyer get prioritized. He defined his ideal client once at setup. After that, it thinks like a strategist who knows his business.

Thankful he made it free for the rest of us.... 

👉 https://lnkd.in/eGSAVjzM

In theory, MCP servers should enable you to analyze your company's performance in Claude or ChatGPT.
 
Connect GA4, HubSpot, Stripe, Google Ads, Mixpanel, etc— and let the AI figure it out.
 
I tried it. Claude churned through a lot of credits and gave me the wrong numbers.
 
Most of these MCPs weren't built for data analysis. HubSpot's MCP is still in beta and doesn't include marketing analytics data. Meta's was built by the community, not Meta. Ahrefs hits API limits. These aren't edge cases — they're the norm. Shopify's doesn't even provide store analytics.
 
So, when I tried to do this with multiple MCP servers, I could not get it done.
 
Yesterday, I tried Claude with just two MCP servers: Databox and Slack. Databox for the numbers. Slack for the context: every update, decision, and discussion from our teams.
 
It worked brilliantly.
 
I produced a 13-page business performance teardown — covering customer acquisition, upgrades, churn, and team performance, plus leading and lagging indicators across our 3 main objectives and 12 key projects.
 
We were doing it manually for years. It used to take a week — fifteen people cascading data up through every VP. By the time I saw the final version, half the decisions were already made.
 
I asked my team why this doesn't work with all of the MCP servers connected....
 
First, there is a context window problem. Every AI has a fixed working memory: a whiteboard. The conversation, the data, and the instructions for every connected tool all have to fit on it. MCP tool definitions are verbose. Six servers burns 35–50% of that whiteboard — up to 72,000 tokens — just listing available tools, before a single word of the actual question gets read. As a result, accuracy drops. The wrong tools get called. This is why when I was trying this the first time, I actually hit Claude limits & had to upgrade to the $125/mo plan just to keep going.
 
Second, there's a data consistency problem. Meta calls them leads. GA4 calls them users. Stripe calls them customers. Same person, three different labels, three different systems. Individual MCPs won't make sense of this. Further, they can't join across sources or even different objects within sources very reliably. This is why numbers came back wrong: AI was guessing how to turn the raw data into KPIs and understand what I was asking for.
 
Databox solves this with a different architecture. One MCP connection to 130+ data sources: data defined: joined, aggregated, calculated. Cross object and source queries run inside Databox, not inside Claude's context window. Claude gets the time series data, not raw data to wrestle with.
 
If you're trying to do analysis, don't repeat my mistake: use a unified data layer for analysis, not individual MCPs.

More details in comments.

Most of the time, spreadsheets are a terrible way to analyze your data.
 
Bobby Kircher, who runs Papaya Search, a search marketing agency knows this. One of his clients is a multi-location engine service business with nine locations. A lot of their demand comes through phone calls, not website forms.
 
Every month, someone at that company was manually pulling call data into a spreadsheet from three sources: Google Business Profile, Google Ads, and Google Analytics. Nine locations, one number at a time. That's 27 cut and paste jobs for an incomplete picture. Bobby's team didn't know it was happening until they saw it live on a monthly call.
 
They went into Databox and built a single dashboard pulling all three call sources across each location into one view, updating every hour (not every month.) 

Bobby's approach: "Break down sources before combining totals." What that revealed surprised them. Most prospects were never visiting the website at all. They were calling directly from Google Business Profiles or tapping numbers straight from ads, making their decision before they ever landed on a page.
 
Bobby's takeaway: "Aggregate before you analyze."
 
Google Sheets is the 5th most popular integration amongst Databox customers. Spreadsheets make sense when you're tracking something manually or when a tool doesn't have an API. But when they do, automating the data syncing not only saves time, it enables more thorough analysis. And as more and more companies use AI to analyze their performance more frequently and more thoroughly, having data synced into a central place becomes even more critical.
 
If you're still cutting and pasting (or downloading data) into spreadsheets, watch Bobby's walk through the setup in the video.

When the AI is so "good", it's not believable... that a human wrote it.

In-depth performance analysis used to require a very specific kind of person. Someone curious enough to ask the questions and methodical, analytical & technical enough to find the answer.
 
Here's what that actually looked like:
 
• Figure out which metric looks off
• Find or build the right dashboard
• Drill down by a dimension to see why it's off
• Pull a second metric to check for cause and effect
• Form a hypothesis & validate it by talking to people closest to the work. 
• Brainstorm solutions
• Write up the conclusions so they're digestible and hold up to scrutiny by others
 
Most people didn't have the skill, patience, or time for all of that. So most questions just didn't get asked.
 
That's changing.
 
Benny Stapel works at Unison, a digital marketing agency in the Netherlands. They use the Clickup and Databox MCPs to automatically generate weekly client reports that show not just what the agency did, but how it performed and what they're doing next. 
 
They also use Genie to get questions answered on demand. "𝘋𝘳𝘢𝘸𝘪𝘯𝘨 𝘪𝘯-𝘥𝘦𝘱𝘵𝘩 𝘤𝘰𝘯𝘤𝘭𝘶𝘴𝘪𝘰𝘯𝘴 𝘪𝘴 𝘴𝘰 𝘮𝘶𝘤𝘩 𝘦𝘢𝘴𝘪𝘦𝘳 𝘯𝘰𝘸, 𝘮𝘢𝘪𝘯𝘭𝘺 𝘣𝘦𝘤𝘢𝘶𝘴𝘦 𝘺𝘰𝘶 𝘤𝘢𝘯 𝘭𝘪𝘵𝘦𝘳𝘢𝘭𝘭𝘺 𝘫𝘶𝘴𝘵 𝘢𝘴𝘬 𝘵𝘩𝘦 𝘱𝘭𝘢𝘵𝘧𝘰𝘳𝘮 𝘵𝘩𝘦 𝘦𝘹𝘢𝘤𝘵 𝘲𝘶𝘦𝘴𝘵𝘪𝘰𝘯 𝘺𝘰𝘶 𝘩𝘢𝘷𝘦 𝘪𝘯 𝘮𝘪𝘯𝘥. 𝘛𝘩𝘦 𝘧𝘰𝘤𝘶𝘴 𝘪𝘴 𝘯𝘰𝘸 𝘰𝘯 𝘮𝘢𝘬𝘪𝘯𝘨 𝘤𝘳𝘦𝘢𝘵𝘪𝘷𝘦𝘴, 𝘪𝘯𝘴𝘵𝘦𝘢𝘥 𝘰𝘧 𝘭𝘰𝘰𝘬𝘪𝘯𝘨 𝘢𝘵 𝘥𝘢𝘵𝘢."
 
Benny's not the only one:
 
Claudio Lira, PhD built a Claude skill that returns exactly one answer: where the growth system is breaking and what to do next. 

Rick Kranz (The AI Marketing Lab) does deep-dive SEO analysis filtered through his client's ideal customer profile, automatically. 

Kamil Rextin (42 Agency) used voice commands with the Databox MCP & Claude to determine whether a viral LinkedIn post actually drove incremental website traffic. 

Ryan Eger (USA Home Listings) auto-generates a white-labeled investor dashboard weekly.

László Fazakas (Arcanian Consulting) built a vector database of client communications and feeds it alongside Databox performance data into an LLM for deep client analysis. 

Steve James (Stream Creative) pulls up the Databox MCP before every client call to surface improvement opportunities. 
 
I'm experiencing the same shift personally. Last month, I stopped asking my team to pull numbers before leadership meetings. I get the answer myself and walk in asking "what are we doing next?" or "How can we address x problem?" instead of "how are we doing?" The shift isn't just speed.
 
The unlock isn't just time saved. It's the questions that now get asked that never did before.
 
If you want to see how other companies are building these capabilities, we're walking through it on April 8: https://bit.ly/3NKteVr
 
What's a question you're not asking right now because getting the answer takes too long?

Open article: [Launch Event] Introducing Databox AI | Databox by bit.ly

[Launch Event] Introducing Databox AI | Databox

What are your go-to MCP servers right now? My three... Slack, HubSpot, Databox (ofc). 

1️⃣ Slack MCP: At Databox, we use Slack for much of our internal communications. Updates on initiatives, communicating new processes, even sharing performance with context about what we did and how it worked. 

Whenever I'm trying to figure out what we did when, this is my go to. I just ask Claude to use the Slack MCP and search for stuff. 

2️⃣ HubSpot MCP: I take about 2-3 calls per day with prospects and partners. A lot of times, prospects have already spoken to my team about something, have tried our product already, etc. I need that info so I don't look unprepared. For customers and partners, we have an even longer history that I need to look up. HubSpot MCP lets me quickly surface all that communication.  

I could log into HubSpot to get this info, but Claude is quicker. 

3️⃣ Databox MCP: I use our own MCP (https://databox.com/mcp) to do quick performance checks as well as deep-dives. All by myself, I actually wrote a 13 page teardown of our business the other day pulling and plotting our performance from 16 different datasources (HubSpot, Google Analytics, Mixpanel, Profitwell, etc, etc) as well as context from Slack (and my own head) to put this together. It took me four hours, but before AI, it would have taken 15 people about 30 minutes each and someone spending a bunch of time nagging them to update a report. 

As a CEO, context switching is probably the hardest part of the job. I have to be able to jump on a call with a prospect, prep for a 1:1 or do analysis all in one day. Sometimes, all in one hour. 

Having Claude as my one interface connected to these 3 MCP servers has been a major boost to my ability to do so. 

What's your go-to MCP servers? I'm curious what else I should be using and why.

taplio

The Battle

Hannah Eisenberg

Peter Caputa

Followers

Posts

Engagements

Likes

Performance ScorePeter Caputa's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Hannah Eisenberg

Peter Caputa

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Peter Caputa's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScorePeter Caputa's LinkedIn Performance