Philip Lacor is the CRO of Personio, a $3B+ HR and payroll platform with 1,500 employees, 15,000 customers, and a 400-person sales team. He shared their AI transformation journey at SaaStr AI London — and the learnings are a masterclass for any revenue leader trying to figure out how to actually deploy AI in GTM.
We’re all hearing about AI-native companies crushing it. Replit, Gamma, Harvey.
But what if you’re running a real B2B company? One with 400 salespeople, 15,000 customers, and years of accumulated process debt?
That’s exactly where Personio was in May 2024 when their CEO kicked off an “AI Surge Week” — and what happened next is one of the most practical AI transformation stories I’ve heard.
In just six months, they went from “90% of our team uses LLMs weekly” (which sounds good but isn’t transformation) to building 400+ AI assistants, cutting research time from 2 hours to 15 minutes per rep, and booking 140 meetings in 7 days through their AI SDR.
Here’s what Philip learned — the stuff that actually worked, and the mistakes you should avoid.
The 5 Lessons: What Actually Works
Lesson #1: You Need Both Top-Down AND Bottom-Up Motion
Here’s the trap most companies fall into: They give everyone access to ChatGPT, run some training, and call it an AI initiative.
Personio did that too. Their AI Surge Week was a huge success — speakers from OpenAI, Mistral, AWS. Project teams building agents. Company buzzing with excitement.
But then Philip noticed something: High usage isn’t the same as transformation.
“After the AI Surge Week, we felt that although usage was high, this is maybe not enough to reach true transformation and to really fundamentally change the way we go to market.”
The problem? Bottom-up motion alone can’t make the hard decisions:
- Resource allocation — Who’s going to spend 40% of their time on AI initiatives?
- Permission — Can people actually stop doing their old workflows?
- Budget — Which tools do you actually buy vs. just test?
- Prioritization — Of the 50 possible use cases, which 3 do you build first?
This is why Philip started the “AI Powered Go-To-Market” working group in June — a top-down initiative to complement the bottom-up energy.
The takeaway: Bottoms-up gets you experimentation. Top-down gets you scale. You need both.
Lesson #2: Cross-Functional is Non-Negotiable
This one seems obvious but almost nobody does it right.
Personio built a working group with three distinct capabilities:
- Data & Systems Team — Owns infrastructure, Snowflake, the technical backbone
- Revenue Operations + GTM Engineers — The bridge between tech and business (they have 2 dedicated GTM engineers now)
- The Business — Marketing, Sales, Customer Success, the actual users
Why does this matter? Philip saw both failure modes:
“We have seen cases where our data systems team built things with LLMs but it was lacking the business context and therefore the models didn’t work very well.”
And the reverse:
“We had sales people who wanted to do something but originally they did not have the support from either data or systems or RevOps.”
They deliberately made the working group large — 15 people — to get broad coverage across functions and build cultural buy-in.
The takeaway: AI in GTM isn’t a sales project or a data project. It’s a cross-functional transformation. Build the team accordingly.
Lesson #3: Use Jobs-To-Be-Done to Prioritize Ruthlessly
Here’s what happened after they launched the Slack channel and started working on use cases:
“People started to share opportunities, raising their hand, and the problem was that as people started to work on these new ideas, we hadn’t finished the first one. At one point it started to spiral a little bit out of control.”
Sound familiar? Everyone gets excited, ideas flow, and suddenly you have 20 half-built things.
Their solution: Jobs-to-be-done mapping.
One of their GTM engineers literally shadowed account managers for two weeks. What she found:
- AMs were working in 7-8 different systems to perform simple tasks
- Constantly switching contexts, pulling information together
- Losing 2.5 hours per day on one activity, 3 hours per week on another
They mapped every role’s jobs-to-be-done:
- SDRs
- AEs
- Customer Success
- Solution Engineers
Then they overlaid these jobs onto the customer journey to see how they fit together — and where the biggest pain points were.
The takeaway: Don’t just chase shiny AI use cases. Map your roles’ actual jobs, quantify the time waste, then prioritize based on where you have the biggest P&L challenge or customer experience gap.
Lesson #4: Building an AI Culture Requires Leading, Sharing, and Celebrating
Philip has a formula he uses for transformation:
Effect = Quality of Plan × Acceptance
5 × 5 is way bigger than 10 × 1.
So how do you build acceptance? Three things:
Lead It: Philip does deal reviews where AEs used to show up with big PowerPoints. Now:
“I would always go like, okay, please go to Gong, open up your account. There’s this little AI sign. Go in there. Now you look for the account brief and everything is there. And in real time, we would do away with PowerPoint.”
The next time? Reps already know to use Gong’s AI. Leaders have to model the behavior.
Share It: They put their own teams on stage to share what they’ve built:
- An assistant to personalize customer decks
- An assistant to answer RFPs
- The expansion SDR assistant
Internal success stories inspire more adoption than any training program.
Celebrate It: This one’s clever: They announced that President’s Club will have 2-3 seats reserved for best AI contributions.
Not sales performance. AI contribution. And next year? Even more seats.
The takeaway: The #1 trait for an AI-powered GTM org? Curiosity. Hire for it, reward it, model it.
Lesson #5: Great AI Comes From Your Stack + Your Context
Here’s an insight most people miss:
“Let’s not go out and buy all these tools because usually the tools are not the panacea. There’s usually a lot of work that you need to do in your workflows, in your data.”
Personio’s approach: Start with what you have, add LLMs on top, iterate from there.
Their core stack:
- Salesforce (or HubSpot)
- Gong — They made a big bet here because “for a go-to-market organization, the customer conversation is obviously a very important source of data”
- Qualified — Started with fast meeting booking, then layered on AI
- Snowflake — Both structured and unstructured data
- Amazon Bedrock — LLM layer
But here’s the work nobody wants to do:
- One-third of their Salesforce data were duplicates — They installed automatic de-duping
- Months cleaning the prospect database — Buying external data, connecting sources
- Loading 5,000 Gong calls into Snowflake
- Adding emails, connecting everything together
Then — and this is critical — they added Personio-specific context:
- ICP definitions
- Pitch decks
- Onboarding processes
- Product training materials
“This is really critical to really train the LLMs and to make it specific for your go-to-market, for your customers, and for your products.”
The takeaway: The AI is only as good as the data and context you feed it. Clean your data. Connect your systems. Add your tribal knowledge. This is the work that makes AI actually useful.
The 4 Use Cases That Actually Worked
Use Case #1: Win/Loss Intelligence
The problem: Reps fill out Salesforce after winning or losing deals, but 30% of reasons were “Other” and even good data wasn’t deep enough.
The solution: They loaded all conversation data, emails, and Salesforce data into Snowflake and built a GPT for go-to-market.
The results:
- Added 10-15% new insights to competitive battle cards
- Created dynamic, continuously-updating battle cards (instead of static docs that go stale)
- Data-driven product feedback: “Based on 10,000 calls, this is where we have weaknesses”
This is evolving into a “go-to-market brain” — rep coaching, marketing campaigns, product prioritization, all from the same foundation.
Use Case #2: Expansion SDR Assistant (2 Hours → 15 Minutes)
The problem: Expansion SDRs were spending 2 hours per day researching customer information before making cross-sell calls. They’d check account health in Amplitude, contract details in another system, usage data somewhere else…
The solution: A GTM engineer built an assistant embedded directly in Salesforce. Type in an account name, and it pulls data from 10-20 systems, formats it for cross-sell, and provides a recommendation (green/yellow/red).
The results:
- Research time: 2 hours/day → 15 minutes
- Pipeline per FTE: ~2x increase
- SDRs love it — “they’re using it every day and it makes the job better”
Use Case #3: Intent Scoring for Outbound
The problem: Finding the right accounts, right people, right message is solved. But right time — knowing which prospect is actually in a buying cycle — is incredibly hard.
The solution: Their data science team built a dynamic intent score based on multiple signals:
- Website visits
- Former users who moved to new companies
- G2/Trustpilot activity
- And other signals they continue to enrich
The score shows up directly in Salesforce with flame icons (🔥🔥🔥 = start here) and refreshes daily.
Key learning: “We saw that initially the model was not great. It was picking up some signals that we didn’t think were good. We changed it and then it got better and better.”
Use Case #4: AI Chat/SDR (“Nia”) — 140 Meetings in 7 Days
The problem: Demo requests are your best leads. Waiting a week to book a meeting is crazy in a real-time world.
The solution: They deployed Qualified’s AI chat (“Nia”) on their website. When prospects request demos, Nia books meetings immediately — 24/7.
The results:
- 140 meetings booked in 7 days
- 200,000 website sessions processed
- Deep insights into what customers actually ask about (pricing, product questions, etc.)
But here’s what Philip found most valuable:
“When you start reading the chats, you see all of a sudden customers have questions about your product. They want to know what your minimum price is. You’re getting very, very rich insights in what is top of mind for these customers. I got totally hooked on it.”
The 24/7 reality: “People at 11 PM on a Friday evening, they’re thinking about requesting a demo. Why? But they do it.”
The 5 Mistakes: What NOT To Do
Now for the part that matters most — what went wrong and what to avoid.
Mistake #1: Endless Tool Testing Without Going Deep
This came up multiple times:
“I would not endlessly test tools. You got to dig in and go deep, learn from it.”
Everyone wants to try Clay, then Artisan, then the next hot thing. But surface-level testing teaches you nothing. Pick a few tools and go really deep with them.
“The point is not to test every single one of them. You got to pick a couple and go really deep with them. It’s all about the training. It’s all about the data.”
Mistake #2: Learning AI vs. Doing AI
“If you try to read all the papers and not do anything, I don’t think you’ll move fast.”
The insights come from deploying, breaking things, and iterating — not from another podcast episode or Twitter thread.
Mistake #3: Not Having Dedicated People Monitoring Agents Daily
When they rolled out Nia, the AI chat:
“There were definitely like maybe four weeks where we didn’t do enough and I don’t know how many demos we wasted by not training Nia really.”
Now they have a dedicated person (“Ami”) who looks at output every day, applies feedback, tests in real time.
Things they discovered only through daily monitoring:
- Nia started giving legal advice (“Better not”)
- Nia started bashing competitors (“That’s not us”)
- When customers ask multiple things at once, Nia would answer the product question but forget to book the demo
“You only learn that when you really start doing it and when you see where the AI stops.”
Mistake #4: Building Without Business Context
Their data systems team built LLM-powered tools, but:
“It was lacking the business context and therefore the models didn’t work very well.”
You can have perfect data infrastructure, but if the people building don’t understand the sales motion, ICP, or customer journey, the output won’t be useful.
This is why the cross-functional working group matters — and why GTM engineers need both business and technical backgrounds.
Mistake #5: Expecting AI Tools to Be Plug-and-Play
“Usually the tools are not the panacea. There’s usually a lot of work that you need to do in your workflows, in your data.”
Personio spent months:
- De-duping Salesforce (1/3 of records were duplicates!)
- Cleaning prospect databases
- Loading conversation data into Snowflake
- Adding company-specific context
The tool vendors won’t tell you this. The AI is maybe 30% of the work. The other 70% is your data, your context, and your workflows.
The Big Question: Can You Double Revenue Without Doubling Headcount?
When asked about next year’s planning:
“Managers say, ‘Hey, I need like 30 more people.’ That default should be, ‘Can I solve this with AI?’ The big question is: can we double the business with the same headcount?”
That’s the real question every CRO should be asking.
Philip’s honest take on where things stand:
- Spending multiple six figures on AI tooling
- Each SDR agent costs about $100K
- Some teams will get smaller, others bigger (channel/partner teams can use more people)
- “We will reallocate people” — it’s not about cutting heads, it’s about growing faster
A 6 Month Surge Is All It Took (To Really Get Going)
Six months. That’s all it took for Personio to go from “AI Surge Week” to 400+ assistants, 2x pipeline per SDR, and AI booking 140 meetings a week.
But here’s what actually made it work:
- Top-down support for the hard decisions
- Cross-functional team bridging data, systems, and business
- Jobs-to-be-done to prioritize ruthlessly
- Culture of AI — leading it, sharing it, celebrating it
- Stack + context — not just tools, but your data and tribal knowledge
And what to avoid:
- Testing tools endlessly without going deep
- Learning AI instead of doing AI
- Not monitoring agents daily with dedicated people
- Building without business context
- Expecting plug-and-play magic
Philip’s final advice:
“The best career advice I can give you is lean into AI. What we do know is that everybody’s jobs will evolve, including mine.”
The AI-native companies are moving fast. Your job, if you’re running a real SaaS company, is to move faster than they expect — and now there’s a playbook.
Philip Laheurte is the CRO of Personio. He flew from New York to London specifically to share this at SaaStr AI London, submitted through our AI speaker form (scored yellow the first time, had to resubmit to get green 🔥), and yes — the irony of an AI-powered GTM leader being evaluated by our AI speaker scorer was not lost on anyone.
