Insights Automation Is Splitting Research in Two
Dovetail, the customer-intelligence platform used by 40% of the Fortune 500, shipped autonomous AI agents on July 14, 2026, that watch customer signals without anyone asking them to. AlphaSense, a data-synthesis platform one rung up the enterprise research stack, now carries a $7.5 billion valuation on $600 million of annual recurring revenue. Insights automation isn't creeping toward market research anymore — it's already load-bearing inside the platforms that feed it.
The short version
- Dovetail (40% of Fortune 500 companies) launched AI Agents, "digital twins," and Channels 2.0 on July 14, 2026 — agents that monitor customer signals unprompted and route them to teams. (PR Newswire)
- AlphaSense, an enterprise data-synthesis platform used across market and competitive research, sits at a $7.5B valuation on $600M ARR. (Greenbook, The Exchange)
- Six independent UK research and insight firms folded into SHARCC Group in 2026, pooling into one holding company worth roughly $10M combined — consolidation under margin pressure, not growth.
- Anthropic's 2026 State of AI Agents report puts data analysis and report generation as the top current agent use case at 60%, with 56% of enterprises planning to add agents specifically for research and reporting within the year. (Anthropic)
- Research-industry veteran Leonard Murphy, on what agents change first: "Do you need five people to do the job now? Maybe not."
What did Dovetail actually ship?
Dovetail's July 14 release bundles three things. AI Agents run continuously in the background, monitoring customer signal streams and surfacing what matters without a human prompting a query first. Digital twins are synthesized personas — built from real calls, tickets, and past research — that a team can "talk to" directly instead of re-reading a stack of transcripts. Channels 2.0 turns support tickets, surveys, and sales calls into revenue-weighted opportunity scores.
CEO Benjamin Humphrey framed the problem it's chasing bluntly: "AI didn't solve the problem of understanding your customers, it made it more painful." The release also carries ISO 42001 certification — the AI-management standard — and 30-plus integrations, including first-party MCP connectors into Slack, Linear, Claude, and Microsoft Copilot.
None of this is market research in the agency sense. It's customer intelligence for product and CX teams. But the mechanics — always-on monitoring, automatic synthesis, a persona you can query instead of a report you have to read — are exactly the mechanics that used to justify a research agency's invoice.
Is AI actually replacing market-research jobs?
Not according to the people running research shops today. Leonard Murphy calls it a headcount problem, not an extinction event — fewer analysts needed per project, not zero. Karen Lynch, on the same Greenbook podcast, put it plainly: "AI has disrupted... but I don't think AI has displaced workers in our industry."
The evidence backs a squeeze, not a wipeout. SHARCC Group's six-firm merger is what you do when execution-layer margins compress faster than any single small shop can absorb — you pool fixed costs instead of competing on volume you can't win. That's consolidation math, not a company closing.
Where does the automation actually stop right now?
Lay Dovetail and AlphaSense next to what a research agency still does by hand, and the boundary is clear:
| Layer | Status in 2026 |
|---|---|
| Signal collection (tickets, calls, surveys) | Automated — Dovetail, Channels 2.0 |
| Continuous monitoring & routing | Automated — Dovetail AI Agents |
| Cross-source synthesis | Automated — AlphaSense ($600M ARR at scale) |
| Persona simulation / "talk to the data" | Automated — digital twins |
| Defining the research brief | Manual |
| Designing the study, sample, questionnaire | Manual |
| Turning findings into a client narrative and deck | Manual |
The top half of that table is bought, not built, by most enterprises now — that's what a $7.5B valuation and a 40%-of-Fortune-500 footprint mean. The bottom half — deciding what's worth asking, designing how to ask it, and shaping the answer into something a client can act on — is still a person's job, in most shops, done manually. It's also the exact gap we're building ARIA to close: client brief in, full research deliverable set out, through an agent pipeline rather than a human doing the synthesis by hand. We've written before about how research automation reached CPG intelligence before it reached agencies — this is the same pattern moving up the stack.
Where this leaves agencies
Two forces are running at once. Platforms with venture money keep buying up the signal-and-synthesis layer — Dovetail's Fortune 500 footprint and AlphaSense's ARR are both evidence of how fast that layer scaled. Smaller shops are responding by merging, the way SHARCC's six firms did, because you can't out-hire an agent that runs at 3 a.m.
Neither force has touched the layer above both yet: deciding what's worth researching in the first place. That's the part still priced like expertise. For now.
More on where AI in research is actually headed — and what agencies should build instead of buy — at Dekrypt Labs research.
Frequently Asked Questions
Will AI agents replace market-research agencies? Not according to current practitioner accounts. Leonard Murphy and Karen Lynch both describe AI cutting headcount needed per project rather than eliminating the work, per Greenbook's 2026 "The Exchange" podcast. The squeeze is hitting execution-layer analyst work first, not the strategic design of what to research.
What did Dovetail launch in July 2026? AI Agents that monitor customer signals autonomously, "digital twins" built from real calls, tickets, and research for querying a synthesized customer persona directly, and Channels 2.0 for scoring support tickets and sales calls as revenue-weighted opportunities — alongside ISO 42001 certification.
How big is AlphaSense's business? AlphaSense, an enterprise data-synthesis platform used in market and competitive research, carries a $7.5 billion valuation on $600 million of annual recurring revenue as of 2026, per Greenbook's The Exchange podcast — evidence that enterprise research-signal tooling is already at serious scale.
Why are small UK research firms merging? Six independent UK data-and-insight firms combined into SHARCC Group in 2026, pooling roughly $10 million in combined revenue under one holding structure — a direct response to margin pressure as platforms automate the execution layer of research work.
Abhishek Gupta is Co-Founder at Dekrypt Labs, building ARIA — an AI research pipeline from brief to deck. dekryptlabs.com