One customer record, not five.
Coming soon: IBIOS AI customer deduplication will use fuzzy name matching to flag likely duplicate customer records for your team to review. It is not live yet. Variant spellings, maiden names and transcription differences are what it is designed to catch.
Planned: phonetic and fuzzy name matching, with every merge reviewed by a person.
How do duplicate customer records break a CRM?
A customer visits your showroom in March and gives her name as “Sandra Morrison.” A record is created. In June she calls to follow up on a quote. The call note records her name as “Sondra Morison.” A second record is created. In August she emails about installation, using her married name: “Sandra McAllister.” A third record is created.
Three records. Three quote histories that are each incomplete. When a staff member looks her up, they see one quote. When she calls asking about the “original quote from March,” your team cannot find it. The relationship looks broken from both sides.
Coming soon: IBIOS deduplication is designed to catch the second and third record at the point of creation and show the likely match for staff to confirm.
How does AI customer deduplication work?
Coming soon, not live yet. Three layers designed to keep your customer database accurate as it grows.
Phonetic and fuzzy name matching
Coming soon: phonetic matching for names that sound alike but are spelled differently, and fuzzy matching for typos and transpositions, at record creation and in scans of existing records.
Human-reviewed merges
Coming soon: likely duplicates shown for staff review with a side-by-side comparison of both records. Staff confirm or dismiss, so the decision stays with the person who knows the customer.
Critical for multi-channel businesses
Website forms, phone calls and showroom visits all create records from different starting points. Without dedup, the same customer appears three times with three different histories. Coming soon: dedup to keep the record clean from the start.
What will the deduplication review look like?
Coming soon, not live yet. When IBIOS detects a likely duplicate, it is designed to present a side-by-side comparison rather than making any changes silently. An example:
Likely duplicate flagged — review panel
Record A: “Sandra Morrison” — 07890 123456, 14 Elm Drive. 2 jobs, 3 quotes on file. Record B: “Sondra Morison” — 07890 123456, 14 Elm Dr. 0 jobs, 1 quote. Match confidence: high. Action: Merge into Record A / Keep separate / Not a duplicate.
Outcome after staff confirmation
Staff confirms merge. Record B is merged into Record A. All quotes, jobs, and notes consolidated into a single record. Merge logged in audit trail with confirming staff member and timestamp.
Who this matters most for
Any business where customer history is spread across multiple channels.
Businesses with web and phone enquiries
A customer who enquires through your website and then calls a week later will often give their name slightly differently. Without dedup, that becomes two records with split history.
Businesses using AI receptionists (IBIOS AI receptionist coming soon)
AI receptionists transcribe names from spoken audio. Phonetic variation between what was said and what was transcribed is a common source of duplicates, which dedup (coming soon) is designed to catch.
Businesses with long customer relationships
A customer who bought a bathroom suite three years ago and is now enquiring about a kitchen refit may use a different email address or phone number. Coming soon: dedup matching on name and address to surface the existing relationship.
Any business with a growing CRM
Duplicate records compound over time. The earlier dedup is applied, the less cleanup is needed. Starting clean is easier than cleaning up a large backlog of duplicates later.
What is planned, technically?
Matching algorithms
Phonetic matching for name sounds. Fuzzy distance matching for spelling variation. Combined scoring with configurable threshold.
Trigger points
Runs at new record creation. Can also be triggered as a background scan across the existing customer database.
Match presentation
Side-by-side comparison in the IBIOS dedup review panel. Shows name, address, phone, job history, and quote history for both records.
Merge behaviour
Staff-confirmed merges only. No automatic merging. Confirmed merges are irreversible and logged in the audit trail.
Business name handling
Token-order-independent matching. Handles Ltd/Limited, &/and, abbreviations. Separate logic from personal name matching.
Current status
Coming soon. The possible-duplicates view is built but switched off, so none of this is live yet.
Questions about AI customer deduplication
How does customer deduplication work?
By comparing names that sound alike or are spelled slightly differently; the possible-duplicates view is coming soon and switched off today. It is designed to catch “McNeil” and “MacNeill”, typos and transpositions, and to show a likely match for review rather than creating a silent duplicate.
What causes duplicate records?
Duplicates arise most commonly when a customer contacts the business via a different channel: a web chat enquiry, a call handled by the AI receptionist (coming soon), or a staff member taking a phone message. The customer gives their name slightly differently each time, and without dedup, each contact creates a new record. Over time this fragments the customer history and makes pipeline reporting unreliable.
Will dedup merge records automatically?
No. When it launches (coming soon), IBIOS will present likely matches for human review rather than merging automatically, because two people at the same address can be separate customers. Staff will confirm or dismiss each proposed merge.
Can I review before merging?
Yes, that is the design; the review view is coming soon, not live yet. Each proposed merge will show the two records side by side (name, address, phone numbers, jobs and quotes) so staff can merge, keep separate, or mark as not a duplicate.
Does dedup affect existing records?
Only with staff confirmation, once the feature launches; it is coming soon. Matching is planned both for new records as they are created and as a scan across existing customers, and neither will change a record without staff confirming it.
How does dedup handle business names vs personal names?
Differently, when it launches (coming soon). Business names are planned to match regardless of word order and common abbreviations (Ltd, Limited, &, and), while personal names will allow for spelling, titles, hyphens and middle names.
Which IBIOS product line includes AI customer deduplication?
AI customer deduplication (coming soon) will be part of the IBIOS CRM product line. It is built for UK tradespeople and wholesale and trade suppliers. See also: merging a duplicate customer.
IBIOS (Integrated Business Intelligence & Operations System) is made by IBIOS Limited (NI743167), Northern Ireland. Access is by waiting list; there is no free trial.
AI customer deduplication is coming soon.
Fuzzy name matching and human-reviewed merges are not live yet. Join the waiting list to hear when they launch.