In most SMEs, a sales rep picks up the phone, glances at the display, and tries to remember who it is. They open the CRM, start typing part of a name, half a minute goes by, the customer is already getting impatient. The call never lands in the CRM, because there is no time. That is not a CRM problem and not a phone problem — it is a problem of the two not talking to each other.
Integrating telephony with the CRM — in English this is CTI, computer telephony integration — has been around for years and is not technically difficult. And yet in small and mid-sized companies it is still rare. The reason is simple: vendors sell it as an "integration project", the owner sees a quote for the equivalent of ten thousand euros, and says "maybe later". In most cases it is two or three days of configuration, and the payback is one of the fastest you can point to in an SME at all.
This article is about what CTI actually does, where it breaks, and what has to be in order before the deployment makes any sense.
What telephony-to-CRM integration actually delivers
Forget the slide deck for a moment. Three things really happen; ignore the rest.
The first is the customer card popping up the moment the phone rings. The system catches the caller's number, matches it against the customer record, and before the rep picks up the handset the screen is already showing the company name, the outstanding balance, the last order and the note that says "promised to send a sample on Friday". The 20–30 seconds of hunting after every call disappear, and — more importantly — so does the moment where the customer hears "and what was your name again?".
The second is click-to-dial from inside the CRM. The phone number in the record is a link — one click and the PBX places the call. Nobody retypes digits, nobody confuses a 6 with a 9, and outgoing calls log themselves in the contact's history.
The third and most important is the automatic call entry in the CRM. Date, time, direction, duration, who answered — no manual clicking. The rep only adds what was actually agreed in the call. This is the moment where team activity statistics, follow-up reminders and a "customers nobody has called in 60 days" report start to make any sense.
Further features — transfers based on history, VIP-aware queueing, recording with transcription — are useful but they are the second layer. Without those three basics, none of the rest matters.
A real example — an 18-person electrical wholesaler
At one of our clients — an 18-person electrical wholesaler with six sales reps — the team handled around 140 inbound calls a day. Before the integration, the average handling of a new call (identifying the caller, opening the record, jotting a note) took about a minute, and — once we measured it — only 32% of calls made it into the CRM. The rest stayed in the reps' heads until evening and then evaporated.
We rolled out an integration between the 3CX PBX and Comarch Optima CRM: screen pop on ring, click-to-dial, automatic call entry with metadata. Configuration plus tidying up the customer records together took eight working days, including a pilot with two reps.
Three months of collected metrics showed three things. The share of calls actually logged went from 32% to 95%. The average handling time per call dropped by roughly 25 seconds — across the team that is about an hour a day, roughly 22 hours a month of recovered time. But the most interesting thing happened on the other side: because calls now logged themselves, we could switch on a "customer with no contact for 60 days" report, and in the first month the team brought 14 dormant customers back into active contact. That report hadn't worked before, because it had no data.
It's worth noting what did not happen at the same time. Nobody was fired. Nobody built "big brother for the sales team". The manager saw that two people were doing 45% of the phone work and one was barely on the phone at all — and that fed into a one-on-one conversation, not a termination. CTI gives visibility. What you do with that visibility is a management decision, not a technology one.
Where it breaks
Four classic places where telephony-to-CRM integration falls over. All of them predictable.
Numbers in a mess. In the customer records, one number is stored as +48 22 123 45 67, another as 221234567, a third as 22-123-45-67. The algorithm matching an incoming call against a record either finds nothing or finds the wrong one. Standardising to E.164 (+48221234567) is a one-off day of Excel work — but without it the rest is pointless.
Duplicates in the record base. The same customer as three separate records, each with its own history. The pop-up shows a random one, call history scatters across three cards, activity reports lie. Clean up the records before the integration, not after.
Sales reps calling from private mobiles. "It's more convenient", "the customer already has this number". The whole integration misses the point — those calls will never reach the system. This is a management decision, not a technical one: either the company provides a decent tool (a softphone app on the phone, with the same business number), or it accepts that X% of calls stay off the radar.
A PBX with no integration surface. Older phone-operator packages give you a "management portal" and no API. Nothing there is machine-readable. Before you plan CTI, check whether your telephony has an open interface at all — 3CX, Slican, Focus Telecom, Cludo, LiveCall, Ringostat do; many "free" telco-bundled PBXs do not.
What has to be in place before you start
Three things, each to be ticked off before day one of configuration.
First, a CRM with an open API or a ready-made CTI plug-in. Comarch CRM, Firmao, Livespace, Bitrix24, HubSpot have plug-ins or webhooks. If you use your own CRM or a line-of-business app with a customer module, you need a REST API and an endpoint that accepts an "incoming call from number X" event and triggers the pop-up.
Second, VoIP telephony capable of pushing events out. Not every PBX can do this — the question for your telephony vendor is: "do you expose an API or webhooks for call events (ring, answer, hangup)?". If the answer is "we have a portal", that means "no".
Third, a clean customer record base. One customer, one record. Phone numbers in a single format. This is not an IT project, it is a one-off admin job that someone in the company has to actually do. Without it, the pop-up will show either the wrong record or nothing at all.
Where to start
The order in which we usually do this with clients:
- Measure the starting point. How many calls a day, what share of them makes it into the CRM today. Without that number, you will never prove anything changed.
- Check both systems for an API. Both the PBX and the CRM. If either doesn't have one, the project starts with replacing that piece, not with integration.
- Standardise phone numbers and clean up duplicates. One day of spreadsheet work, one person responsible. The next step won't work without this.
- Run a pilot with two or three reps for four weeks. Measure the share of logged calls, handling time, number of mismatched pop-ups.
- Roll out to the whole team only when the pilot shows the difference. If it shows 20% improvement instead of 60%, that usually means: go back to step 3.
This is not a quarter-long deployment or a six-figure investment. For most SMEs it is on the order of a few thousand euros and two to three working weeks — provided the record base and the systems are in a state that lets you start. If they are not, you clean up first and integrate second.
If your reps are on the phone all day and the activity reports are still a lottery of "how much anyone remembers", have a chat with somebody who does these integrations without unnecessary drama. The first call is free and takes half an hour.
A sales call that isn't logged in the CRM effectively didn't happen. No manager will schedule the follow-up, no colleague will pick the case up, and the team's activity report is guesswork.