DiscoverStill Early
Still Early
Claim Ownership

Still Early

Author: Niklas Stephenson & Peter Suhm

Subscribed: 0Played: 0
Share

Description

Every other week, Niklas Stephenson and Peter Suhm hang out and talk about what's been going on in their businesses.
17 Episodes
Reverse
In this episode:Why a new feature isn’t enough on its own anymore (Phone.inc is launching Whisper together with a mini documentary about building it)Plans for a studio where AI does most of the editing, which other builders in Copenhagen can use tooIs SaaS dead? Niklas thinks people forget the “as a service” part, but small tools that are basically one feature are in troublePhone.inc’s AI receptionist, built to plug into the phone system rather than replace the humansOver 20% of new signups upgrade to Pro, and two customers built Phone.inc integrations without being askedPeter finally figured out how to use LinearPhone.inc is about to ship Whisper. When you pick up a call, an AI voice says a few words about who’s calling and why. Peter is filming the next vlog about how the team built it, and it clicked for him that the two should launch together. There’s so much noise now that a feature on its own won’t get people who aren’t already customers to pay attention. Someone said it recently and Peter really wants it to be true: you kind of have to be a media company now.The contract for the new office is ready. Office rentals are one of the few things that need investor approval, so it’s waiting on that. The plan includes a studio with a few scenes set up for product demos and podcasts. You press record and AI makes the first draft of the edit. Gear is cheap now, so anyone (the new sales team included) should be able to walk in and record a five minute demo for LinkedIn without Peter setting anything up. It’ll sit empty most of the time, so they want other people building cool stuff in Copenhagen to use it too.Peter already has a first version of the AI editing working. Niklas built Cable, a little Mac app that shows as many plugged-in iPhones and Androids as you want on one screen (super useful for Phone.inc demos). Claude Code turned a one shot demo of it into a pretty decent edit in two or three prompts. Peter has a PR that moves the whole thing into Amp: raw recordings go in a Dropbox folder, an Amp thread does the edit, and a video comes back. For the latest day in the life vlog, Amp sorted around a hundred clips into screenshot grids, indexes and transcripts. ChatGPT turned that into a paper edit, and the editor had Codex lay out the Final Cut timeline from it.Peter and Katrine also built a changelog that lives on the website and in the app. Every day an Amp workflow looks at what got merged, drafts entries and posts them in Slack for review. Peter used to pay $50 a month for a tool that did this, which got them onto whether SaaS is dead. Niklas thinks people forget the “as a service” part. Tools that are basically one feature are in trouble, but he hasn’t met a customer who wants to build what Phone.inc does themselves (if anything, they want to get away from building on Twilio). Keeping the website next to the code has its own problems though. They often put up a page for a feature before building it, to see if anyone cares, and Peter’s new writing guide keeps pointing out that those features don’t exist yet. The AI receptionist page even mentioned which feature flag it’s hidden behind.With Peter full time and Brian on sales, Niklas wrote himself a new job description for running a company with a go to market side. Shopify stores are the beachhead. He’s also finishing the AI receptionist. A year ago almost nobody asked for AI calls, and customers still want to talk human to human. But voice agents keep getting better, and receptionist is the obvious first job for one. It transfers calls to a department or a specific person, takes messages, and answers questions from a crawl of your website. It also hooks into the rest of the phone system (warm transfers with a short brief for whoever picks up, availability checks, the call queue). It needs Pro and is billed per minute on top.Over 20% of new signups have upgraded to Pro in the last two weeks without the team doing anything. Niklas reads that as a sign that not enough people are coming in yet. Partnerships are pulling harder than anything he’s seen (he usually tells founders to stay away from them early), and a free product is easy for partners to promote. Two customers built Phone.inc integrations into their own SaaS products without telling anyone, which is part of why partner apps launched last week. Twilio is the developer platform for telephony, but it doesn’t own the screen the call happens on. Phone.inc does, so other products could live right there during a call.And Peter finally figured out Linear. A weekly plan sets out what a good week looks like, so he knows when he’s done. A private team holds his own notes, and he can reach it from Claude and ChatGPT over MCP. Everything is organized around five pillars from his job description (right now the main one is getting ready to sell to Shopify stores). Niklas wants that as the first demo in the new studio.Still Early is sponsored by Phone.inc: a free business phone number, welcome greeting, call routing, and opening hours, all from an app on your personal phone.
In this episode:Phone.inc plans monthly, and a month already feels too longNiklas is a solo founder for the first time, and with growth and sales hires coming in, he has to step back from building product (his safe space)The “human CPU” test: if something new needs a person to keep it alive every week, is it worth it?Peter wants their editor working on the content system instead of every single videoPhone.inc needs a bigger office with a studio, and Niklas wants to let other entrepreneurs use both for freeOG Kit went down during a PHP security issue, and Peter went to the bar anywayPeter is back after two weeks off. He started at Phone.inc on October 1, worked for a week, then took the break he’d made a condition of joining (a move, the Danish fall break, a friend visiting from the States). He half expected to spend it coming up with startup ideas, but Phone.inc was the only work thing running in the back of his head.OG Kit did go down while he was away. A PHP security issue meant waiting for a patched build for the Linux distro it runs on, on the same evening as a bar night with his local entrepreneur friends. When he built OG Kit he promised himself it would never be mission critical, so he took the app offline and went out. Visti (who works at Laravel) asked for SSH access and had it back up before Peter got home. Existing images kept working the whole time, since they’re cached behind Cloudflare.Niklas’s job is changing again. Phone.inc is still behind a waitlist, but hundreds of companies use it every day, Peter is leading growth, and Brian is joining as VP of sales. Niklas has always had a commercial co-founder before, so this is the first time he’s both. His old rule was not to hire product people before 100 employees, but that assumed a co-founder wearing the product hat, so he’s thinking about a product manager sooner. Building product from the couch is his safe space, and he knows that can’t be his job anymore.Phone.inc plans monthly right now, and even that feels too long. The planned work could be done in a week, the rest of the month fills up with unplanned stuff, and planning a month ahead feels like planning a year. Next up is a goal for the month, with projects planned week by week.They’re both becoming more manager than IC. Peter is mapping his role as engines and weekly rhythms (when someone ships a feature, Amp should at least draft a post for the internal releases channel), and putting his energy into the AI writing guide instead of nitpicking hundreds of pages of copy. Niklas’s test for anything new is whether it needs a human to keep it alive every week, because you run out of human CPU. He thinks one-person billion-dollar companies will be pretty niche, and pointed to Amp running with a team of 21.Peter also has a call with Joseph (their editor, who also makes the music for the videos) about building a content machine, so everyday videos get faster and there’s more time for launch videos and customer stories. Niklas wants a 15-minute way to announce small features, which right now just ship quietly.Phone.inc is outgrowing its nine-person room in a co-working space (sales needs a boiler room, and that doesn’t work in the same room as the engineers), and any new office will be about 20 seats too big. Niklas got his start in 2009 in an Aarhus office that a consultancy didn’t need anymore and gave to young startups for free. He’d like to do the same: free desks for young entrepreneurs, and a studio anyone can use when Phone.inc isn’t. The gear is cheap now (maybe $5-6k for a basic setup), the room is the expensive part.Niklas built an Amp plugin for their support tool, so a bug report goes into Linear, Amp fixes it and drafts the reply to the customer. A human still presses ship, and Lea is working on a support voice because the drafts don’t sound like Phone.inc yet. He’s also spent two weeks trying to reach Shopify about partner terms (a 15% cut feels fair for customers found in the App Store, not for everyone else). And he got the iPhone Duo on Friday, where the killer app turns out to be Amp.Still Early is sponsored by Phone.inc: a business phone number, welcome greeting, call routing, and opening hours, all from an app on your personal phone.
In this episode:Why Phone.inc gives away phone numbers for freeWhat’s getting built in October (call whispers, voice-changed greetings, maybe SMS)A bug report button that hands every report to a coding agentChatGPT did almost all the work of submitting Phone.inc’s ChatGPT pluginWhy they’re making their videos in-house instead of hiring an agencyFirst episode recorded in person, in the Phone.inc office.A lot of people just need a phone number for something (the App Store, Stripe, registering a company), and Phone.inc’s free plan gives them one. Peter thinks it’s a huge use case that’s still up for grabs. With freemium you own the lead, as Patrick Campbell used to say, and Phone.inc literally has their number. Niklas cares less about why they sign up. If Phone.inc is where someone got their first business number, it’s probably where they’ll go the next time they need one.October is the first month that isn’t only foundational work. Call whispers give you a few words when you pick up (who’s calling, and probably why) based on caller context and previous calls. Welcome greetings get a voice changer, so you record the greeting yourself and it comes out in a different voice, with your pronunciation of the company name intact. The stretch goal is texting on business numbers, the most requested feature. It’s pretty easy in the US and UK, and a lot more work in Denmark.Niklas vibe coded a bug report button. Customers click the part of the page that’s broken, Sentry captures the session, and an Amp thread starts on a fix right away, all the way to a PR that’s ready to merge.In his first week as growth lead, Peter submitted the Phone.inc MCP server as a ChatGPT plugin. ChatGPT’s plugin creator did nearly all of it (test cases, checking his demo video, telling him to set up a test account with a call and a voicemail). He thinks that’s roughly how onboarding should feel.Peter wants Phone.inc to eventually make the kind of launch videos later-stage startups put out. They could hire an agency for that, but they’d rather get good at it themselves, which means shipping something every week and improving as they go. Companies that buy one great video often struggle to make the next one. And with so much AI-generated video around now, they want theirs to clearly be made by people.Still Early is sponsored by Phone.inc: a business phone number, welcome greeting, call routing, and opening hours, all from an app on your personal phone.
In this episode:Peter's first day at Phone.inc (office breakfast, a design sync, and half an hour picking Lea's brain about the customer journey)He didn't have to set up a local dev environment, and made his first PR through AmpJesse Hanley from Bento tweeted that he didn't want to be on the waitlist, which turned into a bunch of new signupsPhone.inc hit the September goal of 100 companies (124 now), with about 15% of free users converting to paidNiklas hired a VP of sales, starting November 1What does an activated user look like for a phone number that might never ring?Peter started full time at Phone.inc on Thursday. He filmed the whole day for a vlog (the coffee machine, the car door, the drive to Copenhagen in his right-hand drive manual from 2012). Lea had ordered breakfast, then came standup, the Thursday design sync that Katrine runs, and a priorities call with Niklas (who was home with his kids after three days of travelling). After lunch he spent half an hour asking Lea where customers are right now.The nicest surprise was that he probably doesn't need a local development environment at all. Everything was already set up in Amp, so he typed into a text box, got a tour of the app and the website, and had staging sites created for him. His first PR was adding the Bento logo to the website.On Friday Peter woke up early and saw that Jesse Hanley from Bento had tweeted that he needed a number and didn't want to be on the waitlist. Peter had gotten admin access the night before, so he invited Jesse and a few others from Twitter. That led to a lot of tweets back and forth, new followers and a bunch of signups. Jesse asked how this can be free. His use case is simple, Phone.inc runs its own voice infrastructure so the cost is low, and him tweeting about it is worth more than what he would've paid.Phone.inc hit the September goal of 100 companies and is at 124 now. Around 15% of free users convert to paid, and there are no gates yet (just an upgrade button). There are paying customers in the US, the UK and Finland, and the US has passed Denmark in the last 50 waitlist signups.Niklas spent two days in Stockholm at a go-to-market meetup one of their investors organized. A revenue leader from a fast-growing AI company talked about running self-serve and enterprise sales at the same time (most of their early enterprise deals started on the self-serve side). Niklas also hired a VP of sales, someone he's worked with before who builds direct sales teams from zero. He starts November 1, and there'll probably be a sales team from January.Peter is in the Copenhagen office two or three days a week for meetings, feedback and lunch, with remote days for deep work. He's still sorting out the logistics (a new electric car, a school carpool, the right train, and moving house in two weeks).Peter is responsible for activation now, and it's hard to define for a phone number. Jesse might never get a call and still be an active user. Niklas wants one simple metric (weekly active companies that made or received a call), plus softer signals like whether the number is on the company's website. They talked about a test call button, asking people at signup why they need a number, and what to do with free numbers nobody has used for six months. Neither of them thinks there's a right answer yet.Niklas keeps both his notes and his tasks in Linear, in a private team, and adds to it by talking to Claude in the car.Still Early is sponsored by Phone.inc: a business phone number, welcome greeting, call routing, and opening hours, all from an app on your personal phone.
In this episode:Tailwind is joining Shopify, and Peter isn't going with themWhy he said no to pursuing Shopify (big company, no side projects, night calls from Europe)He's joining Phone.inc as product growth lead on October 1Why Niklas wanted a founder type in the role instead of someone with a finished playbookThe 2012 job Peter quit after four months, and the case for founders taking a job for a seasonPhone.inc shipped do not disturb and call transfer, and invites are open againTailwind is joining Shopify. Adam Wathan announced it last week, and Peter's last day at Tailwind Labs was the day before. He'd been there four years, never full time (a couple of days a week as a contractor), but it was his main stable thing the whole time. He originally joined to build a job board, which he'd completely forgotten until recently. His own startup was near the end of its runway, and Adam asked if he wanted to start a business inside Tailwind Labs part time. It later turned into running partnerships. He's known about the acquisition for a long time, which made Laravel Live Denmark a month ago pretty strange (the perfect place to talk to everyone about it, and he couldn't).Following the team to Shopify was something Peter could have pursued, and there were a ton of good reasons to do it. But he's never really wanted to work at a big company, and if he ever did, he wanted it to be a decision he made on purpose. He also had too many things going on the side that he wasn't ready to let go of, and he'd have been the only person in Europe on a North American team. Adam told him he didn't have to come, and that it would still be a good outcome for him if he went and did his own thing.Instead, Peter is joining Phone.inc as product growth lead on October 1. He was already an investor and helping out as a contractor. The role covers marketing and the part of product that overlaps with it (the founder-led growth videos, onboarding, changes in the app itself). Niklas thinks of it as founder in residence, and Peter says it looks a lot like his job when he was starting Reform. A friend had been asking him for months why he was only there one day a week, since he already sounded like he was running growth.Peter isn't very employable (the last time he didn't work for his own company was 2012), and his real worry was joining when he knows he'll want to do his own thing again someday. Niklas said that's exactly who he's looking for. An experienced growth hire usually brings the one playbook that worked somewhere else, and at this stage Phone.inc needs someone who can go find the one that works for them. Peter also tells the story of his first job as a junior Rails developer. He quit after four months, and his boss offered to fund his startup if he stayed. Niklas makes the case (from his blog post "Seasons in life") that founders can take a job for a few years without it being an identity crisis. He did it at Trustpilot after Firmafon and says it was the best thing he's done career-wise.On the Phone.inc side, invites were paused in early September until two core features shipped: do not disturb and call transfer. Both went live yesterday. At Firmafon, transferring a call meant remembering a star code, so people just didn't (some customers literally threw the phone across the office instead). Now there's a transfer button next to end call, a live list of colleagues and their status, and you can brief the person before handing over while the caller listens to hold music Mikkel made in a day of jamming (Niklas says it's the best he's ever heard). Invites are back on, with ten new companies in a day and a goal of 100 live companies in two weeks. Phone.inc is also hiring someone to lead direct sales.The podcast will follow Phone.inc even more closely from here, hopefully recorded in person at the office.Still Early is sponsored by Phone.inc: a business phone number, welcome greeting, call routing, and opening hours, all from an app on your personal phone.
loading
Comments