Editor’s note: This new Relay by Chargebee series features field notes from AI-native SaaS founders as they tinker, build, and monetize their paths towards a new kind of product-market fit playbook. This post captures the story of Pocus, in Alexa’s words. Dig in, drop a question in the comments, and if you have a suggestion, write in as well.
Here’s an index documenting the top takeaways, to help you find the right sections quickly:
- Why 90% of AI tools are built wrong for enterprise
- How Pocus is confidently differentiating from the incumbents with 2.5 strong levers
- Product discovery for AI features, what has changed, and why jumping to “using AI” is counterproductive
- Navigating AI pricing for a buyer persona that is too used to the predictability of the seat-based model
- The 4 win-win elements Pocus relies on to inform their usage-based pricing model
- Hot take: “I don’t believe in AI SDRs right now”
- State of the enterprise buying market for AI and why 2025 is the “get ROI” stage of the adoption journey
Why 90% of AI tools are built wrong for enterprise
“There’s a reason B2B SaaS exploded in the 2010s. Companies realized that individual productivity tools (think: personal Excel sheets, local databases) were creating chaos at scale. The winning B2B products?
They built for teams first:
- Salesforce replaced individual contact lists
- Slack replaced person-to-person emails
- Figma replaced solo design files
- GitHub replaced individual code repositories
So why are AI companies ignoring this lesson?”
Source: Alexa’s original LinkedIn post here
We’re different from a lot of startups in the sense that we started with enterprise versus starting with SMB.
Enterprises evaluate things differently. They want to know how they can use AI to roll out capabilities to their entire user base, supercharge their sales team, and hit revenue and pipeline goals.
On the other hand, a lot of SMB products or point solutions are good at targeting one specific workflow for one rep/user. At scale, though, it becomes complicated to replicate the success they see with individual workflows.
Our focus is on being more than just a point solution and actually help teams deliver AI at scale. This means a few different things.
We focus on making sure that every single rep has the right prescriptive guidance/playbook from AI versus enabling a few siloed reps with a playbook that’s totally disconnected from other reps within the same org.
Every rep should be running their playbooks on top of unified data that is tailored to your business. This is foundational data intelligence that is consistent for each rep helps guide everyone the same way.
We think this is important. There needs to be org-wide consistency when using AI so that you can actually learn what doesn’t work, what works, and iterate over time while being able to consider all the relevant context from all possible sources and teammates.
There’s a lot of tactical features that we prioritized as a result:
- Building our AI on top of your data and learning from it (think CRM notes, call recordings, product data, and enablement materials)
- tools for admins to roll out features properly,
- reporting to help them monitor results,
- dashboards for frontline managers to understand how reps are using AI,
- and so on.
You could call these the “unsexy” parts of rolling out AI, but these are the features that actually help us deploy to larger organizations.
Going back to why I think you need org-wide consistency with AI platforms, a really important and fundamental benefit of AI is that it learns and gets smarter over time as you use it more.
So if you have just one rep using an AI point solution, it’s getting smarter only for that one person and their context; the company doesn’t realize the full value.
When the entire organization is using and operating off of the same platform, it learns from ALL the reps and their context, and you get to realise compounded benefits.
How Pocus is confidently differentiating from the incumbents with 2.5 strong levers
In terms of the incumbents, I’d say there’s 2.5 main differentiators.
The 0.5 is the easier one. Pocus is more intuitive and user-friendly and sales reps love it. You can’t win on that alone, but it is an easy way to beat an incumbent.
One big differentiator is that we have built something called a Relevance Agent, which is an AI agent in the backend that is constantly learning everything from internal and external data about your company.
We are building a backend and an AI agent that is so much smarter and can give so much more contextualized, actionable recommendations to reps than any incumbent will. In order for incumbents to beat us on this, they would have to rebuild their entire backend.
The second thing is we’re very prescriptive in our approach.
I’ve heard a lot of traditional sales leaders say that they just give their junior sales reps a “phone book”, plop it on their desk, and expect them to know what to do.
It doesn’t work.
With Pocus and with AI, you can be way more prescriptive and tap reps on the shoulder to say, “here’s the account you need to focus on”, “these are the people you need to target within the account”, “here’s the messaging and here’s the value hypothesis”, and so on.
The third problem we tackle is data overload.
On any day, reps have to go to Salesforce, Google Sheets, internal admin, Tableau, ZoomInfo, sales nav, company website, and a million other tabs just to figure out who to go after and how.
With Pocus AI, you can not just consolidate this information but also give the “so what” of this data. Why it matters and how to use it well.
All of these become our right to differentiate and win.
Product discovery for AI features, what has changed, and why jumping to “using AI” is counterproductive
The product discovery process for our AI features has been exactly the same as before and I’m pretty confident it should be.
I think a lot of people today start building with a sentiment that they want to “use AI”. They start with the solution and then try to find a problem. We never try to do that.
We always try to start with the problem, find the solution, and see if the solution can be done better, faster, cheaper with AI.
For instance, our AI strategy feature that consolidates internal and external data and then puts together an account plan for a rep, was new as of a couple months ago.
It didn’t start with “we need AI account plans”.
It started with the problem of reps spending a lot of time trying to bring this data together and doing the first level of research and understanding. It also started with us seeing that leaders were spending a lot of time trying to enable their reps.
So once we decided that it was a problem to be solved:
- we started putting together different mocks of what it could look like in Figma,
- running those by customers,
- iterating with them
- and then really understanding where AI could actually make it faster, better, cheaper,
- and, as a result, help our customers generate more pipeline.
Our customer product discovery is exactly the same as it has always been.
The only difference is that with AI it’s much faster to prototype features and get to a working alpha or beta.
Another thing is that the wow moments are a little bit bigger. When you take an otherwise manual workflow and automate it in seconds with AI, users find it fascinating.
It is also tricky in some way. AI has an effect. You have to be aware that when you’re showing customers a product for the first time, there’s always going to be a wow moment.
It’s enabling you to do some really cool things, but you also have to make sure that the wow moments happen on the third, fourth, fifth, sixth time as well, not just the first time.
Navigating AI pricing for a buyer persona that is too used to the predictability of the seat-based model
Figuring out monetization and pricing with AI is really hard and I don’t think anyone has nailed it yet. I also think that we will keep iterating on it over the next couple of years.
Historically, we have been primarily seat-based. We would sell a plan based on the number of reps with a limit on first party data access.
We’re doing the same thing with AI. What has changed is instead of applying a limit on the CRM data ingested or warehouse data ingested, we’ve introduced credits for the additional data we’re providing.
For us, we’re in this interesting balance where we sell to CROs and a lot of them only think in seat-based models. So if you’re trying to sell usage-based or AI-based or work-based to a CRO, they’re not going to like it because it’s hard for them to predict.
But we also know that in order for us to drive really meaningful business and make sure we’re covering our costs while giving our end users value, the model has to be usage-based.
We’re trying to thread the needle between both and it’s hard right now. The way we are managing this right now is that we sell bundles where companies can buy credits in bulk, instead of a pay-as-you-go model which can feel more uncertain and unpredictable.
I could see a future moving away from seat-based. So then you have to think, would it be outcome-based? Usage-based?
I don’t think it will happen this year. And based on what I was saying before, CROs aren’t ready to buy in a new way and it increases a lot of friction in the sales process. We’re a startup, we don’t want friction.
The 4 win-win elements Pocus relies on to inform their usage-based pricing model
Our pricing is changing, and it will continue to change based on how we cost our AI. We have a seat-based model, and then we typically have limits to the usage.
There are four main levers to this.
First is first party data. This pertains to how much data you are ingesting from your CRM or your data warehouse — the actual records that you bring in.
Second is website visitors. This is the number of people who land on your website every single month.
Third is email lookup. This comes into play when you want to see contact information for a person.
And the fourth is what we call Watched Accounts. This is the number of prospects that we’re constantly using AI to scrape data on.
The reason why we have all these limits is that it costs us money to do all these actions. It costs us more money to take more of these actions and it also gives the reps or the end users more value the more they use them.
Currently we offer a set limit for all four with the base plan and then we give you the option to buy more credits if you need more.
Hot take: “I don’t believe in AI SDRs right now”
I don’t believe in AI SDRs right now.
There was a lot of hype around them and then a lot of churn. AI SDRs are fine when there’s a very transactional sale, but if you’re selling to more commercial, mid-market/enterprise accounts, you need a human to build relationships.
AI can help you be more prepared for meetings and prospecting, but it’s not going to completely replace humans. This might change in the future as AI gets better, but I don’t see it happening in the next couple of years.
I also think that there are two extremes with AI for sales. One extreme thinks that AI is going to be able to replace sellers, which I don’t believe in at all. I think it’s just marketing at scale.
And then there are some people who think AI is actually making sellers not as smart because it’s taking away from their critical thinking, which I don’t agree with either.
I think that AI can do the first level of research and reasoning and then set up the reps to do stronger, better, more deeper critical thinking.
State of the enterprise buying market for AI and why 2025 is the “get ROI” stage of the adoption journey
You get two extremes with enterprise buyers today, in 2025.
On one end, you have people who are skeptical of AI and the second that Pocus or any AI tool hallucinates or gives a bad result, they’re done, they lose trust, they hate it. They’ve moved on.
And then you have the other side of the spectrum where there’s people who are really bought in on the future of AI and they can see the vision and they’re passionate about building with you. There is more leeway than you had before in SaaS because they get it.
They would like to give you feedback and almost build with you.
Last year there were also a lot of boards and CEOs telling their teams that they needed to “adopt AI”. So there was a lot of experimental AI and a lot of people in organizations feeling pressured to test AI.
As we get into 2025 people are realizing that their approach was probably not the right way to go about things. It was just experimental AI.
Now they actually want to feel real ROI on what they are doing with AI.