Every service provider runs on an OSS/BSS. It holds the account: what the subscriber pays, which tickets are open, what services their address can take, what happened the last three times they called in. For most operators it is both the system of record and the system of work – the screen a rep has open when the phone rings.
What it does not hold is the conversation. [Text Wrapping Break]The competitor that a subscriber names. The cancellation that threatened on Tuesday and forgotten by Thursday. The promise a rep made that never made it into a ticket. All of the above is information that exists. It is sitting in the call recordings, email threads and support tickets, but most contact centers only have capacity to review fewer than 5% of their customer interactions, which means the other 95% is just stored rather than read.
This is not a new problem either, but operators’ stubbornness to adopt new workflows isn’t the reason it persists. QA has always been sampled because listening to calls costs an hour of a supervisor’s time for every hour of conversation. Dashboards what show at-risk subscribers get built from tickets and billing events because those are the fields that exist. The result is a familiar asymmetry: an operator can tell you precisely what a subscriber pays, but struggle to guess if they are likely to leave soon because of bad experience.
Today, the cost of data analysis, in any form, has dropped significantly. Call evaluation en masse, AI transcription and data analysis have reached a point where scoring every conversation costs less than sampling a fraction of them, and is accurate enough that operators are willing to act on the output. Once a conversation becomes structured data – a churn risk score, a named competitor, a broken commitment – someone has to decide where that data lives. Put it in a separate analytics tool and it becomes another dashboard, which might add complexity to workflows. But, put it in the OSS/BSS and it lands in front of the person who can act on it in the tool they’ve been using the entire time.
A working example
One recent implementation shows the shape this takes. In July, Communications Data Group (CDG), whose cloud-based Elements platform delivers OSS/BSS – billing, invoicing, ticketing, provisioning, subscriber care and interconnect – for more than 100 service providers nationwide, announced an integration with QueSee AI, a customer experience and retention intelligence platform.
The mechanism is worth describing, because the design choices generalize well beyond these two vendors.
Before a call is analyzed, the analysis layer pulls the subscriber’s record out of the OSS/BSS: open tickets, billing history, network data, and the products available at that address. The conversation is then scored against the operator’s own standard operating procedures with that context attached, rather than as a standalone transcript. When the call ends, the score, the analysis, the summary and the call notes are written back to the customer profile in Elements. Churn risk scores and churn drivers populate Elements fields. Retention tickets are created inside Elements.
The context matters more than it sounds. A rep’s correct answer to “can I get a faster plan?” depends entirely on what the subscriber already has, what the address supports, and how their last three tickets went. A transcript-only analysis cannot tell a good answer from a bad one, because it does not know what a good answer would have been.
“Elements knows all of that. QueSee hears the conversation. Connected, they score every single call in full context of the customer relationship – and the operator’s team never has to leave the tools they already run.” – Andrew Konovalenko, CEO, QueSee AI
CDG frames the integration as a function of how Elements is built.
“We operate broadband networks ourselves, so we built Elements around the realities service providers face every day. Our approach has always been to give operators flexibility. Open architecture allows them to use the solutions that work best for their business while keeping everything connected inside Elements. QueSee adds another layer of intelligence, in this case powered by GenAI, by turning customer interactions into actionable insights directly within the platform our clients already rely on to run their operations.” – Tony Stout, CTO, CDG and PRTC
The architectural point above is worth paying attention to: OSS/BSS stays the record of truth, and the intelligence layer feeds it rather than competing with it. That is a different bet than the one most customer experience analytics vendors made over the past decade, when the default was a standalone platform with its own login and its own version of the customer.
What operators are reporting
Two operators have published results from this approach.
360 Broadband, a rural ISP serving parts of Oklahoma and Texas, reports scoring 57,270 calls in its first year, retaining 103 subscribers it had flagged as at risk, and now protecting $9,352 a month in recurring revenue. Over the same period it reports agent SOP scores rising from 76.6% to 84.2%. Amplex Internet, an Ohio ISP, reports retaining six subscribers in its first six weeks and cutting call review time from 45 minutes to five.
These are single-operator figures, reported by the operators themselves, and they are not a forecast for anyone else’s network. Their value is in what they suggest about where the return shows up. It is not in the QA score. It is in the subscribers who did not leave – which is also the hardest number for an operator to see, because a retained subscriber generates no event, no ticket and no line in a report.
The questions this raises
The CDG integration is available now, with the first joint deployments underway. The technology is moving quickly and proving its value. Integration, though, is the easy half.
What an operator does with a signal it has never had before is a set of workflows, not a feature, and those workflows are unique to each provider. What does a regional ISP actually change once every call is visible, and which of those changes is the one that lets it beat a national carrier at the thing it can win – knowing the subscriber? Who owns each issue the data surfaces, when a dispatch policy, a plan agents never offer and a competitor named in forty calls all have different owners, and none of them is the QA team? And how do you design for correlation – a node degrading, subscribers on that node naming competitors eleven days later, an overbuilder filing in the county – when any one of those is a report and only the three together are a decision?
Automated and accurate analysis of customer experience is close to being solved. Deciding who acts on the answer, and building the view that shows the network, the subscriber and the market in one frame, however, is still an open question.
That is the work, and it is where independents have a shot at capturing the market today.
Inside Telecom provides you with an extensive list of content covering all aspects of the tech industry. Keep an eye on our Telecom sections to stay informed and up-to-date with our daily articles.
