By Mitch Rice
Business intelligence software has won. Most companies past the seed stage run some combination of Tableau, Power BI, Looker, or Metabase; dashboards greet every Monday meeting; “data-driven” stopped being a differentiator and became table stakes. The category did exactly what it promised: it made the what visible. Pipeline coverage is at 2.1x against a 3x target. Churn risk is concentrated in mid-market accounts onboarded last spring. The DACH region is outperforming forecast by 40%.
Then the meeting ends, and someone has to do something. And here is where the stack goes quiet — because every action a BI insight demands begins with a person, and BI software does not know any people.
What BI is actually good at
To be precise about the gap, be precise about the strength. BI software aggregates your internal data — CRM, product analytics, billing, support tickets — and makes patterns legible. It answers questions like:
- Which segments are converting above or below baseline?
- Where is revenue concentrating or leaking?
- What changed this quarter versus last?
These are aggregate, internal, retrospective questions, and modern BI answers them brilliantly. The dashboards aren’t wrong. They’re just structurally silent on the next question, the one every operator asks within thirty seconds of seeing the chart: okay — so who do we call?
The dead-end pattern
Watch how each insight dies without a name attached:
“DACH is outperforming — we should double down.” Doubling down means finding buyers: which heads of operations at mid-size German manufacturers are evaluating tools like yours right now? Your BI stack contains zero rows about them. They’re not in your CRM because you’ve never talked to them. The insight is real; the actionable population is invisible.
“Churn risk is up in accounts onboarded last spring.” The fix is human: find the champion at each at-risk account, check if they still work there (a surprising fraction won’t), find who inherited the relationship if they left. BI flagged the accounts; the people inside them are stale CRM contacts from fourteen months ago.
“Competitor X’s customers are posting complaints about their price increase.” A displacement campaign needs the actual people complaining — names, roles, verified emails — not the sentiment trend line.
“Our best-performing partners are regional system integrators.” Great: who are the managing partners at the 50 similar integrators you haven’t signed yet?
Four insights, four dead ends, one shared root cause: the moment action requires someone outside your existing data, internal analytics has nothing to offer. The boundary of your BI stack is the boundary of your CRM — and the people who matter most to growth are, by definition, not in your CRM yet.
The missing layer: people intelligence
What fills the gap is a layer that does for external people what BI does for internal data. A people intelligence platform takes the question BI hands off — “who are the buyers / champions / partners implied by this insight?” — and answers it by searching live external sources: professional networks, company sites, funding databases, hiring pages, podcasts, social activity. You describe the population in natural language (“operations directors at German manufacturers, 200–2,000 employees, currently hiring for digitalization roles”), and it returns ranked, verified, evidence-backed people.
The structural difference from buying a static contact database matters here. The populations BI insights point at are defined by current conditions — currently hiring, recently complained, just inherited the account. Static databases can’t see those conditions; they were crawled before the conditions existed. A query-time search over live sources is the only architecture that matches the freshness of the question.
Two layers, one workflow
The point isn’t that people intelligence replaces BI — they answer disjoint questions and the value is in the handoff:
| Business intelligence | People intelligence | |
| Primary question | What is happening? | Who do we act through? |
| Data source | Internal (CRM, product, billing) | External, live (web, networks, filings) |
| Unit of output | Metric, trend, segment | Person, with role + verified contact |
| Time orientation | Retrospective | Current-state |
| Typical consumer | Leadership, analysts | Sales, recruiting, partnerships, CS |
| Failure mode alone | Insight without action | Action without direction |
The last row is the one to sit with. BI alone produces meetings that end with “we should look into that.” People intelligence alone produces outreach with no strategy behind it. Wired together, the loop closes: the dashboard finds the pattern, the people layer finds the humans, and the outcome data flows back into the dashboard.
A concrete version of the loop, using the DACH example: the BI dashboard flags the region → you run a people search for the buyer profile the region’s wins suggest → before outreach, you pull a company intelligence snapshot on each target account to check headcount trend, hiring focus, and tech signals → outreach references what’s actually true about each company this month → replies and conversions land back in the CRM, and next quarter’s dashboard tells you whether the play worked.
Why this layer is emerging now
Three things had to become true:
- LLMs made intent queries possible. “Find people matching this described situation” wasn’t a computable query until models could decompose natural language into checkable conditions.
- The public professional web got rich enough. Funding announcements, hiring pages, podcast appearances, conference agendas, open-source activity — enough live signal exists to verify conditions that no database indexes.
- BI saturated. When everyone has the same dashboards, the edge moves to execution speed — and execution speed is gated on how fast you can get from insight to a verified human.
That third point is the strategic one. Analytics parity means your competitors see the same patterns you do, at roughly the same time. The differentiating interval is the gap between seeing and talking to the right person — and that interval is precisely what the people layer compresses, from a week of manual research to minutes.
A 30-day pilot for wiring the two layers together
The good news about the people layer is that it doesn’t require a platform migration, a data warehouse project, or anyone’s budget cycle. A credible pilot fits in a month:
Week 1 — pick one insight that dead-ends. Go through the current quarter’s dashboards and choose a single finding whose obvious next step is “talk to people we don’t know yet.” Expansion regions and competitor-displacement signals work best; pure retention plays involve people already in your CRM and won’t test the layer.
Week 2 — turn the insight into populations. Write the buyer or champion profile the insight implies as two or three natural-language queries, run them, and audit the evidence on the top fifty results. This is also the week you learn whether the insight was actionable at all — sometimes the dashboard pattern describes a population too small or too diffuse to pursue, which is itself worth knowing before committing a quarter to it.
Week 3 — outreach with the signals in the message. Contact the verified list, referencing the live conditions that surfaced each person — the funding event, the hiring push, the public post. Hold volume modest; the test is reply quality, not blast scale.
Week 4 — score it like a channel. Compare replies, meetings, and qualified opportunities per hour invested against your current outbound baseline. Pilots of this shape typically show their result clearly in one cycle — the common outcome is reply rates at a multiple of cold-list baseline, because the outreach is anchored to facts that are true this month.
Two failure modes to avoid: choosing a census-shaped question (where a static database is genuinely the better tool, and the pilot proves nothing), and skipping the evidence audit in week 2 — the discipline of checking sources is what keeps the loop honest when it scales.
The takeaway
Audit your last five strategy meetings. Count how many decisions reduced, in the end, to “we need to find and talk to a specific set of people we don’t currently know.” In most go-to-market organizations it’s nearly all of them — expansion, churn defense, competitive displacement, partnerships, hiring. Then count how much of your tooling budget addresses finding those people versus charting the data you already own.
BI software earned its place by making the what legible. The next layer of the stack makes the who reachable. Companies that wire the two together don’t have better dashboards than their competitors — they have shorter distances between a chart and a conversation. In markets where everyone reads the same signals, that distance is the strategy.
Data and information are provided for informational purposes only, and are not intended for investment or other purposes.