Why Faster Internet Still Leaves ISP Customers Unhappy
A fast plan does not guarantee a happy subscriber. Recorded at Broadband Communities, Preseem CEO Dan Siemon joins Georgette Lopez-Aguado and Shelby Smith to talk about alert overload, giving support the same network picture as engineering, why latency matters more than speed, and AI support with a clear path to a human.
With Dan Siemon (CEO, Preseem), Shelby Smith (Customer Success Manager, Sonar Software) and Georgette Lopez-Aguado
November 10, 2026 · 26:03 listen
Watch this episode
Watch it here, or listen in your app of choice.
Show notes
I went to Taco Bell recently and my order was taken by an AI agent. It tried to upsell me. It even clarified something I hadn't said clearly. When I got up to the register, the person I'd normally see juggling orders and filling cups looked a lot less stressed.
I still wanted to know that a real person had understood my order.
That tension between efficiency and trust ran through this whole episode of the Bandwidth Podcast. Shelby Smith, one of our Customer Success Managers who spent years in the operator's seat herself, joined me to talk with Dan Siemon, CEO of Preseem and a longtime Sonar partner. We covered alert overload, what support teams need from network data, and the metric our industry keeps obsessing over.
Too much data, not enough time
Monitoring is supposed to help. For a lot of lean teams, it turns into one more stream of alerts nobody has time to read.
Dan's answer is prioritization. Preseem scores access points by the experience subscribers are actually getting, folding things like contention, capacity and link rates into one number a person can sort on. Sort the table, start at the top, and you know where to go first.
I'd add one thing. People take in data differently. Some want a dashboard. Some want the percentages and capacities. Some just want the five things to do today. Operators should feel free to raise their hand and ask for the version that fits how their team works.
Give support the same picture as the network team
Shelby asked the question I hear most from customers. How do you connect what the network team sees with the person answering the phone?
Dan described tying each subscriber's experience back to the network. If a whole access point is struggling, the rep knows quickly and doesn't spend an hour debugging a call that was never going to be solved from the subscriber's living room. If the access network is clean, the rep knows the problem is likely in the home, and it's worth the time to work through it together.
Consistency matters here. Even your best rep has bad days. A repeatable process lets everyone start at step one every time. Dan suggested it may even change who you hire, toward people who are great with people and backed by good tools.
Proactive is hard, and still worth it
Can proactive work become too much? Dan admitted the industry exposes a lot of data. He was just as clear that a team that never gets proactive never catches up.
Most operators we work with are short staffed. Some are stuck in a loop where customers report problems before the team finds them. Getting out of that loop is hard, and the time you carve out for proactive work tends to come back as fewer tickets.
There's a retention story too. Three subscribers leaving the same access point is rarely a coincidence. People don't like switching internet providers. When the experience is good, a competitor's flier on the doorstep goes straight into the recycling.
Speed sells. Latency is what people feel.
Speed is the number our industry markets. Dan pointed to Preseem's annual network report, which shows usage leveling off once a connection reaches roughly 25 to 30 Mbps. People still buy the big number, and that's the marketing reality.
Latency is the metric he called underrated. A subscriber can have a gig plan with poor latency and still be unhappy. Pages load slowly, games lag, and the spinning wheel shows up anyway. When that happens, the cause is usually somewhere else, in the home router, the Wi-Fi setup, or the network. More bandwidth rarely fixes it.
AI support with a clear way out
Back to my Taco Bell order. Dan mentioned two operators who gave subscribers the choice between an AI agent and a person, and many chose the AI.
Shelby made the point I'd underline. People will trust AI support more when they know they can get out of it and reach a human. Most people calling their ISP aren't calling to say how happy they are. Be transparent about what the agent is, and give people an easy exit.
Where this is heading
Dan sees a shift coming. Right now many operators are so focused on building fiber that one of our customers looked me in the eye and told me he feels like he runs a construction company. Eventually the build is done, and the focus swings back to operational efficiency and the quality of the experience.
We're seeing a version of this inside Sonar too. On recent site visits, I realized some customers didn't know how to use features that have been in the platform for years. Before chasing the next new thing, it's worth making sure your team gets full value from what you already have.
The operators who come out ahead will be the ones who turn all that data into a short list of what to do today, and who make it easy for a subscriber to get a real answer when they call.
Listen to the full episode
Catch the full conversation with Dan Siemon on the Bandwidth Podcast, available now on Spotify, Apple Podcasts and the Bandwidth YouTube channel.
If you enjoyed this episode, follow and subscribe everywhere you get your podcasts so you never miss a conversation.
Questions, answered.
Why do ISP customers with fast internet plans still feel unhappy?
+
Headline speed is not what subscribers feel day to day. Poor latency makes pages load slowly and games lag even on a gig plan, and the cause is usually in the home router, the Wi-Fi setup, or the network. More bandwidth rarely fixes it.
How can ISPs deal with alert overload from network monitoring?
+
Prioritize. Preseem scores access points by subscriber experience, folding contention, capacity, and link rates into one number a team can sort on, so they start at the top of the list and know where to go first.
How can ISP support teams use network data?
+
By tying each subscriber's experience back to the network. If a whole access point is struggling, the rep knows quickly and avoids a long in-home troubleshooting call. If the access network is clean, the issue is likely in the home and worth working through with the subscriber.
Is proactive network monitoring worth it for small ISP teams?
+
Yes, even though it is hard for short-staffed teams. Time spent finding problems before customers report them tends to come back as fewer tickets, and several subscribers leaving the same access point is rarely a coincidence.
Should ISPs use AI agents for customer support?
+
They can, as long as subscribers know what they are talking to and can reach a human easily. Dan Siemon mentioned two operators that offered a choice between an AI agent and a person, and many subscribers chose the AI.
