AI Receptionist: A Year in Review (2025)
February 1, 2026
It's the new year, and before we dive into the exciting features and integrations we have planned, it feels like the right moment to reflect on how far we've come.
The Beginning
Tim and I started working on ai-receptionist.com exactly a year ago. Our starting criteria was simple: we wanted to work together on a project that leveraged AI, and we wanted to build something that would help everyday people.
The idea of creating an AI-powered phone receptionist stemmed from that desire - but also from pure curiosity. I wanted to experiment with the latest advances in AI voice models and see what was actually possible, and just hooking a modern voice AI up to a phone number was a real fun project. It was clear the potential was enormous and that most or all business calls would be handled by a front desk AI receptionist in the future.
It took us three solid months of coding to reach MVP. When we started, our AI tooling was little more than a better autocomplete. Today, AI writes close to 100% of the code, under our strict guidance and review. As senior developers with a combined 50 years of coding experience, it's been an amazing journey. Tim and I have accomplished in one year what I would have guessed would require millions in funding and dozens of developers. The world is changing, and modern AI is going to reshape everything. While I can't say what will happen, I can say that AI has unleashed a wellspring of creativity for me. So many interesting ideas or projects that I would only daydream about because I did not have the time I can now realize in a day or two. Rather than Tim and I falling into obscurity as we age, we have a renewed super-power! We have the experience of writing production code for two decades, meaning we can evaluate the AI's output for correctness. We are now both essentially managers of AI agents, providing direction and review.
First Users, First Validation
Starting in April, we reached out to family and friends and were lucky enough to recruit some early believers. Most notably, my aunt signed up with her Mobile Pet Clinic business. The calls rolled in, and we got our first real-world feedback.
The system worked.
It was saving her assistant - who previously answered all the calls - hours on the phone every week. The AI Receptionist handles routine calls ("What are your hours?"), emergencies ("My pet is very sick!"), and even emotionally difficult conversations ("Tell me about euthanasia").
The Valley of Doubt
Shortly after launch, we had a wave of sign-ups - clients who are still with us today. But then... nothing. Just deafening silence. We experimented with ads on Reddit, sending custom generated postcards, and a number of other things but just... nothing.
We kept building. We kept telling people online. But nobody new came. This was discouraging, and Tim and I both wondered if we had just wasted four solid months. The system rolled on without any new clients. Tim took a vacation. I focused on other work. We made tweaks, fixed bugs, and quietly started to resign ourselves to the possibility that this was going to be much harder than we thought, and possibly just one big write-off.
But then something on the internet turned on.
Starting in August, new clients began signing up regularly - and it's been like that ever since. It's now clear to both of us that what we've built is genuinely useful and growing. Our customers are sticking around, satisfied with the service. While the world and national news grows more alarming by the day, we're excited to have found an application of modern AI that is unambiguously good for everybody.
As my aunt said: "The callers actually say thank you!"
The Numbers: 2025 By the Stats
As part of our service, we collect general statistics on each call (available in your dashboard's analytics section). I've gathered all this data across all accounts - anonymized, of course - and computed some interesting insights.
High-Level Statistics
| Metric | Value |
|---|---|
| Average Call Duration | 64 seconds |
| Resolution Rate | 76.1% |
| Sentiment Improvement Rate | 20% |
| Spam Calls Filtered | 12.5% |
Resolution rate means the caller got what they needed without our clients ever having to get involved.
Sentiment improvement means callers' attitudes improved over the course of the call as they realized they were speaking with actual intelligence - not a dumb machine.
AI Performance Scores
While the AI Receptionist performed well across most dimensions, here's how it scored:
- Best: Brand Representation - 4.75/5 ⭐
- Needs Work: Emotional Handling - 4.2/5
The Core Problem We're Solving
This next insight captures the core value proposition of AI Receptionist - probably the single biggest reason we created it.
The phone system is broken for small businesses.
So much of the traffic is spam and low-value repetitive calls that it's nearly impossible for a single person or small team to keep up. Our data reflects this reality:
| Call Category | Percentage |
|---|---|
| Low-value calls | 71% |
| Medium-value calls | 25.4% |
| High-value calls | 3.6% |
After a while, some small businesses just stop answering the phone altogether - overwhelmed by the 71% while hoping for the 3.6%. But of course, if you miss that "needle in a haystack," you miss an opportunity that could grow your business.
AI Receptionist handles the 71% so you never miss the 3.6%.
What People Are Calling About
The bulk of all calls are simple inquiries, which makes sense. "What are your hours?"
The next significant category? Hang-ups.
This might seem surprising - until you consider this: our AI Receptionist always includes a clear disclaimer that (1) the call is recorded and (2) it's answered by AI.
You know who hangs up reliably when they hear that? Spammers.
Only the most persistent spammers stick around to talk with the AI, which explains why our spam categorization of ~10% seems low. The rest flee at the sound of "call recorded by AI Receptionist" (learn more about our customizable legal disclaimers).
| Call Type | Percentage |
|---|---|
| General Inquiries | ~50% |
| Hang-ups | ~20% |
| Support | ~12% |
| Spam | ~10% |
| Scheduling | ~5% |
| Sales | 3.1% |
Note: The 3.1% sales calls align closely with our 3.6% "high-value" categorization above.
Call Complexity
The vast majority of calls are low-complexity - simple questions or inquiries that are easily handled.
Breaking Down the 76.1% Resolution Rate
Let's dive deeper into how calls get resolved.
Some highlights:
- Hang-ups: 100% resolved (by definition - nothing to do!)
- Butt Dials: Near 100% resolved. Yes, this is a real category. A significant number of calls are simply accidental dials where the caller's phone connects without them knowing.
- Spam Calls: 90% resolution rate
How Calls Get Resolved
Calls are resolved for various reasons, but mainly because:
- The requested information was provided
- Spammers hung up (problem solved!)
- Messages were taken
- Calls were forwarded to the right person
Caller Emotions
This might be the most interesting insight of all.
| Emotion | Percentage |
|---|---|
| Neutral | 55% |
| Unknown (likely neutral) | 23% |
| Satisfied | 10% |
| Confused | 4.4% |
| Concerned | 3.4% |
| Frustrated | 2% |
Let that 2% sink in.
Think about your history of dealing with automated phone systems. How often do you get frustrated? For most people, the answer is "almost every time."
AI Receptionist represents a clear disruption to what is possibly one of the most frustrating experiences in modern life: talking to a machine that doesn't understand you.
The 4.4% "confused" makes sense - talking with genuinely intelligent AI when calling a small business is still a new experience for most people.
The 3.4% "concerned" reflects that some of our clients deal with emotionally difficult situations, like sick or dying pets.
When Do People Call?
And finally, the least surprising insight: people call during business hours.
Calls peak around noon (shown in GMT) and drop off heavily on Sundays.
Looking Ahead
Now that our work is validated and our business is growing, we're setting our sights on the future. We recently added SMS Agents and will be opening up more communication channels like email and WhatsApp. We'll continue to prioritize our clients, so if you're reading this and you're subscribed - tell us what you want!
In many ways, this last year has been a journey and an education into what the future will look like. Technical barriers are falling every day. While there's a growing pushback against AI - well deserved, in my opinion - it also seems like being mad at AI is like being mad at fire. Yes, it can burn you. But it can also keep you warm in the winter and cook your food. It's something to be mastered, not something to be angry about, and it's definitely not going away.
Tim and I have spent many years in technology - from cutting-edge government-sponsored research, to popular cryptocurrency wallets, to more down-to-earth education technology companies. AI Receptionist has been one of the more interesting ventures simply due to its practicality. We've operated for so long at the bleeding edge that we became used to not having everything we needed to accomplish our goals. We always had to build our own tools and bridge gaps.
But this venture is different. The technology is ready for the application - and in a very real sense, it's starting to build itself. What's possible feels endless.
We hope our work continues to save small businesses the time they need to focus on what matters most: your craft.
Thanks for being part of this journey. Here's to an even bigger 2026. Whether you're a current customer or just learning about us, we're excited to see what we can build together in the year ahead. The future of business communication is here, and it's more intelligent, more accessible, and more affordable than ever before. Start your free trial today and see the difference for yourself.
