People Memory.
A living profile — Speaker Insights — for every person you've spoken with. What they care about, how they communicate, what they've promised, and what's changed.
The "wait, what's their kid's name again?" problem.
You meet with someone every six weeks. Last time, they mentioned their daughter was applying to college. Two months earlier, they mentioned a project they were nervous about. You'd like to remember to ask. You don't.
You meet with twenty people on rotation. They each have context — preferences, opinions, ongoing threads, things they've said that you'd lose track of within a week if it weren't written down somewhere. So it's mostly not written down.
People Memory is the feature that builds that profile — we call it Speaker Insights — from the recordings you're already making, and surfaces it when you need it.
Speaker Insights
Each profile is a Speaker Insights view: their topics, their commitments, their open threads, their communication style — all sourced from things they actually said in your recordings.
What's in a Speaker Insights profile
Topics they care about
Recurring themes across recordings — work concerns, family mentions, hobbies, things they bring up unprompted. Surfaced as a tag cloud and timeline.
Communication style
How they speak — direct, hedged, technical, story-driven. Average words per turn. Words they use disproportionately. Useful for sales reps adapting tone, for managers giving feedback in a register that lands.
Open threads
Topics raised but not resolved across past meetings. The "we should circle back on..." pile that nobody actually circles back on.
Promises (theirs and yours)
Pulled from Promise Tracker. Status of each commitment, timestamped quotes, links back to the recordings.
Mentions of others
Who this person talks about most — colleagues, family members, third parties — and the typical context. Useful for journalists building a network map of a source.
Recent vs. baseline
"They've been talking about their workload more in the last month than in the prior six." Useful for managers noticing burnout signals or sales reps catching account momentum changes.
Personality frameworks
Seven instruments, triangulated.
The deep "Full breakdown" inside Speaker Insights runs each person through seven peer-reviewed personality frameworks. Each is grounded in their own quotes from your recordings — no questionnaire, no enrichment from outside data. Reading the same person across multiple instruments is how you get past any single framework's blind spots.
Trait baseline
Big Five (OCEAN)
Openness, Conscientiousness, Extraversion, Agreeableness, Neuroticism. The most evidence-supported personality model in academic psychology; the substrate every other read on this page derives from. Per Costa & McCrae (1992).
Communication style
DISC
Dominance, Influence, Steadiness, Conscientiousness. Practical communication framework that translates well into "how to ask this person for things." Per Marston (1928).
Team role
Belbin Team Roles
Where this person naturally fits in a working team — Plant, Shaper, Coordinator, Resource Investigator, and the rest. Useful for staffing, useful for understanding why two people keep colliding. Per Belbin (1981).
Conflict mode
Thomas-Kilmann
How they handle disagreement — Competing, Accommodating, Avoiding, Compromising, Collaborating. The frame that explains why some negotiations stall and others move. Per Thomas & Kilmann (1974).
Decision tempo
Cognitive Style
Analytic vs intuitive, deliberate vs decisive. Useful for matching your communication tempo to theirs so the conversation lands.
Work archetype
Predictive Index
Behavioural drivers behind how a person works — pace, formality, autonomy, decisiveness. Maps to the same four-quadrant lexicon the enterprise world already uses.
New · v3.0
Mach-IV (Machiavellianism)
A read on whether someone treats relationships as strategic resources or intrinsically valuable. Three factors: Tactics (willingness to manipulate), Views (cynical worldview), Morality (ends-justify-means thinking). Combined with low Big Five Agreeableness and a Belbin Shaper profile, it's the third leg of the organizational-psychology Dark Triad read. Per Christie & Geis (1970).
Bonfiyah's Mach-IV is conservatively scored — the card defaults to "Population-typical" rather than nudging toward the charged "Machiavellian" label whenever the signal is moderate. Confidence is capped at medium regardless of sample size; the methodology and limitations are surfaced on the card itself, not buried. Read the framework + safeguards →
Every framework on this page surfaces its own methodology, limitations, and source quotes inside the Full Breakdown view. Nothing is inferred from anything other than what the person actually said in your recordings.
Privacy boundaries
People Memory profiles are built only on people who appear in your recordings. Each profile is bound to a voice fingerprint plus the name you've assigned to that speaker. Bonfiyah does not enrich profiles with web data, social media data, public records, or third-party sources. Everything in a profile comes from things that person literally said in your recordings.
Profiles live in your account on our backend — they need to, because they synthesize across recordings to surface patterns, and that synthesis happens server-side. They are isolated to your account, never shared across users, never searched against any global index. Delete a profile in bulk or per-person from Settings → Privacy → People Memory, or delete your whole account and they go with it.
If a person you've recorded with asks to see what their profile contains, you can hand them a one-page export. We think that's the right standard.
FAQ
Which plan includes People Memory?
People Memory is a Pro AI feature. It's part of the cross-recording AI suite, because a profile only becomes useful once it can synthesize across every recording a person appears in.
How does Bonfiyah know which profile a speaker belongs to?
Each profile is bound to a voice fingerprint plus the name you've assigned to that speaker — Bonfiyah identifies people by voice biometrics, never by transcript text. Because the voice fingerprint carries across recordings and devices, a person's profile keeps building wherever they show up.
Does People Memory pull in data from the web or social media?
No. Profiles are built only on people who appear in your recordings, and only from things that person literally said in those recordings. Bonfiyah does not enrich profiles with web data, social media, public records, or any third-party source.
Can I delete a profile, and who else can see it?
Profiles live in your account, isolated to you — never shared across users and never searched against any global index. You can delete a profile in bulk or per-person from Settings → Privacy → People Memory, or delete your whole account and they go with it. If someone you've recorded asks what their profile contains, you can hand them a one-page export.
Can I use People Memory to make hiring or performance decisions?
No. People Memory is a reflective tool, not an assessment instrument. It is not for hiring, promotion, performance evaluation, termination, or any other consequential decision about a person. The EU AI Act and the Colorado AI Act both apply to using personality- or conversation-inference AI as an input to consequential decisions — don't use Bonfiyah's reads that way without your own legal review and your own ground-truth instruments.
Used in these workflows
Where this surface earns its weight.
See the full eighteen-workflow catalog →
What People Memory is — and isn't — for.
It is for: remembering what matters about the people you talk with regularly — what's open, what they care about, how they communicate — so you show up prepared for your own conversations.
It is not for: hiring, promotion, performance evaluation, termination, or any other consequential decision about a person. People Memory is a reflective tool, not an assessment instrument. The EU AI Act — which restricts inferring traits and emotions about people and classifies personal-trait profiling as high-risk — and the Colorado AI Act's parallel employment framework both apply to anyone using conversation- or personality-inference AI as an input to consequential decisions about other people. Don't use Bonfiyah's reads that way without your own legal review and your own ground-truth instruments.
See a sample profile
We'll send an example People Memory profile built from a real (anonymized) sequence of meetings. You'll see what shows up, what doesn't, and where the line is.
No spam. We use ConvertKit. See our privacy policy.