AI CRM  ·  Research Memo

Lightfield — the AI CRM that builds itself from your inbox

2,500 companies, 5 months, zero manual data entry — but zero disclosed revenue, and the founding team’s last product showed the same traction pattern before failing on retention.

June 2026

1

AI-native CRMs target the incumbent CRM market’s worst end-user experience, but $444M+ deployed across 7+ startups may fragment the prize

Incumbents

Structured-data-first CRM with AI bolt-ons

Salesforce ($38B), HubSpot ($2.6B)

AI-Native CRM ← Lightfield

Built from scratch for AI; unstructured or structured-first

Attio, Lightfield, Zero, Day.ai

Revenue OS

All-in-one GTM platform above CRM

Reevo (KP), Monaco (FF)

Enrichment / Outbound

Data enrichment + orchestration, not SoR

Clay ($30–60M ARR)

Horizontal LLMs

Commoditization threat from below

ChatGPT, Claude, Gemini

The structural divide

Structured-data-first (Attio, HubSpot, Salesforce): AI analyzes records humans create. Unstructured-data-first (Lightfield): AI derives structured records from emails, calls, and meetings. SaaStr confirmed Attio and Clay are “still fundamentally structured-data-first.”[4] The question: is this a 3–5 year moat or an 18-month feature lead?

2

The founding team walked away from 25 million users — is Lightfield a corrective or a repeat?

Bull — Architecture break + 5-minute product-market resonance
  • Unstructured-data-first is a genuine architectural break: SaaStr confirms Attio and Clay are “still fundamentally structured-data-first.”[4] Rebuilding a CRM data model around unstructured ingestion is multi-year work.
  • 5-minute product-market resonance is a distribution weapon: Two independent evaluators cited initial Gmail scan as the decisive moment. ~700 contacts in 5 minutes, no CSV, no config.
  • HubSpot migration agent processes 90K records/hour: “Hundreds” of completed migrations in 5 months.[1] HubSpot stock down 58.7%.[10]
  • All key investors stayed through 70→7 team reduction: Coatue, Greylock, Lightspeed explicitly re-underwrote.[3]
  • No cross-customer model training: Compliance advantage most AI CRMs cannot claim.[7]
Bear — No ARR, thin corpus, Tome replay risk
  • No ARR disclosed at 5 months: Founders typically lead with ARR when it validates. Illustrative ceiling: 2,500 sign-ups × ~10 seats × ~$20/seat/mo (blended below the $36 list price for free/discounted tiers) × 12 mo ≈ $6M ARR if every sign-up fully converts to paid — real but not venture-scale without expansion.
  • Startup ICP has 50%+ structural churn: No product improvement eliminates the failure rate of the customer base. CEO: “If you’re an enterprise with 500 reps… this isn’t for you… yet.”[7]
  • Independent evaluators are lukewarm: GTM consultants who assessed Lightfield call it “not there yet” beyond small startups; one operator who trialed it chose a rival GTM platform on outbound and meeting-intelligence depth.
  • Tome at equivalent stage showed identical traction profile: Viral adoption, consumer-grade engagement, strong community — then retention collapsed.
  • No patents on core architecture: All underlying components available via public APIs and open-source tooling.

“I was gobsmacked by how almost ready to go the CRM was.”

— Associate | Venture Capital Firm

“That’s my worst nightmare. I do not want to use Salesforce.”

— Head of Operations | Hardware AI Company
3

Contents

01
Company
The Tome pivot, founding team, product architecture, and the unstructured-data-first thesis
02
Competitive
Market landscape and the competitive forces shaping the AI-native CRM category
03
Risks & Signals
The five forward-looking triggers that will resolve the Lightfield thesis within 12 months
4
01 · Company

Lightfield was born from Tome’s failure — the founders diagnosed why LLMs couldn’t finish work and built the CRM that supplies the missing context

Founders & Team

  • Keith Peiris (CEO) — Facebook 2011 (Search, ML, voice). Instagram (Direct to 500M MAU). Founded Tome 2020; 25M users. Pivoted to Lightfield 2025.[7]
  • Henri Liriani (President) — 6 years at Meta. Led Messenger Human Interfaces; reduced core code by 84%. CPO at Tome.[3]
  • ~35 employees (up from 28 at Nov 2025 launch); 14 open roles[3]
  • HQ: 600 Townsend St, San Francisco

Funding stack (raised as Tome)

RoundDateAmtLead
Seed2020$6.3MGreylock (Reid Hoffman)
Series ASep 2021$26M—
Series BFeb 2023$43M@ $300M post (as Tome)

Three disclosed priced rounds sum to ~$75M. Contrary Research reports $81M cumulative[3] and CBInsights records $107.5M[6] — the gap above the itemized rounds is undisclosed bridge/extension capital. No Lightfield-specific raise announced.

Traction (Mar 2026, self-reported)

2,500+
Companies signed up
100+
YC startups active
20K+
Waitlist
?
ARR (not disclosed)

Pivot Timeline

  • 2020–2023: Tome → 25M users, ~$75M+ raised, $300M val
  • 2023–2024: Retention problem diagnosed; team 70 → 7
  • Nov 2025: Lightfield launches publicly
  • Mar 2026: Migration Agent + 2,500 companies[1]
5
01 · Company

Connect your inbox, get a working CRM in 5 minutes — then ask it questions in plain English and it writes back

Four core capabilities

  • Instant CRM from email scan: Connect Gmail, get 700+ contacts + 400+ companies in 5 minutes. No CSV, no configuration. 95%+ recall accuracy.[2]
  • AI chat that reads AND writes: Query your pipeline in natural language (“What accounts have been stale for 2 weeks?”) and write back (paste LinkedIn contacts, auto-update databases). Zero had read-only at evaluation time.[7]
  • Proactive action items: Extracts follow-ups from calls and emails with timing. “Hey, you’re overdue for a follow-up 4 days from now, 2 weeks from now.”
  • One-Hour HubSpot Migration Agent: Automated transfer of contacts, deals, fields, relationships. 90K records/hour.[1]

Data privacy

No cross-customer model training. “We don’t take any insights from one customer’s CRM and move it to another.” Easy export to Salesforce. Portability-first positioning accelerates adoption.[7]

“Within 5 minutes, Lightfield was able to find significantly more contacts and companies than Zero could. The initial build-out for Zero was just a little bit slower and a little bit glitchy.”

— Associate | Venture Capital Firm

“I like that write capability — at least I didn’t see at Zero yet. If I want to import a list of contacts from LinkedIn, I copy and paste it there, it will automatically write it in my CRM.”

— Head of Operations | Hardware AI Company

“The contact enrichment — oh my God, it saves me so much time. So much time. I can’t even tell you.”

— Head of Operations | Hardware AI Company
6
01 · Company

Unstructured-data-first means the CRM writes itself — but the feature gap list is real and growing

Pricing

PlanPriceKey Feature
Free$0Core CRM
Startup$36/user/moStandard features
Pro$79–99/user/moAdvanced AI, API/MCP

ICP

Founder-led sales teams, 5–100 people, $0–$10M ARR. YC alumni + SF founders. CEO: “a five-person company building tech for farmers, or a ten-person company building tech for compliance officers.”[7]

Go-to-market

Product-led with high-touch founder events. Weekly YC boot camps at SF office. “We have a truck parked outside accelerators.”[7] Copilot positioning — explicitly rejects AI SDR replacement thesis.

Known product gaps (expert-cited)

GapStatus
Slack read/writeOn roadmap
Email sequencingNot available
Mobile appNot available
LinkedIn integrationNot available
Call recorder visibilityShows on call
Granola integrationCustom workflow
Marketing / CS featuresNot available
7
02 · Competitive

Lightfield is the only unstructured-data-first entrant — but 7+ AI-native CRMs compete for the same startup buyer

CompanyFoundedRaisedCustomersPositioning
Salesforce1999Public ($38B rev)150K+Enterprise lock-in; Agentforce AI bolt-on
HubSpot2006Public ($2.6B rev)200K+SMB CRM + marketing bundle; Breeze AI
Lightfield2020/2025$81M2,500 sign-upsUnstructured-data-first; email-scan onboarding
Attio2019$116M~7,000 payingCustomizable relational model; AI attributes
Zero2024Pre-seedUnknownZero-click CRM; Helsinki
Day.ai2023$24M (Sequoia)~120Context graph from unstructured data
Monaco2025$85M (Founders Fund)Public beta (6 named)Revenue OS; Sam Blond
Reevo2024$90M (Khosla/KP)Undisclosed (5 named)All-in-one Revenue OS

$444M+ deployed across 7+ AI-native CRM startups, all targeting the same ICP. Attio has 7,000 paying customers and a 3-year head start. Lightfield has architectural novelty but no confirmed paying customer count. The market likely supports 1–2 winners.

8
02 · Competitive

Incumbent AI is marketing, AI-native fragmentation is real, vibe-coded CRMs threaten the low end — and model deceleration may freeze positions

  • Incumbent commoditization is rising: Salesforce Agentforce hit ~$800M ARR at 200% YoY growth. But architecturally constrained — “No one believes Salesforce is an AI CRM.”[7] HubSpot stock down 58.7%.[10]
  • AI-native fragmentation is intensifying: $444M+ across 7+ startups chasing the same startup buyer. “They all do a lot of the same stuff… it’s really hard to see who’s going to win.”
  • Vibe-coded CRM threat is overstated: Altis’s N=70 GTM survey found only 9% have built internal CRM tooling and 59% never considered it — the startup ICP keeps buying off-the-shelf. Not a near-term displacement risk.
  • Model deceleration could freeze positions: CEO: “We’re starting to plateau a little with reasoning — the difference between Claude 4.1 and Claude 4 is marginal.”[7] If models plateau, application features converge faster.
  • Switching costs cut both ways: The migration agent that pulls customers off HubSpot also makes it easy to leave Lightfield. Data-portability positioning accelerates adoption but limits lock-in.

“No one believes Salesforce is an AI CRM. It is an archaic database from the early 2000s, and they’ve bought an Anthropic account and connected some features to it.”

— Keith Peiris, CEO | Lightfield — AI Summit [7]

“I don’t think any large-scale enterprise organization would use Lightfield as its standalone right now — I don’t think it’s there yet. It’s good, but it’s more for startups and small teams.”

— GTM Consultant | Independent (landscape evaluator)
9
03 · Risks & Signals

Five triggers that will resolve the Lightfield thesis within 12 months

  1. November 2026 cohort retention. The full November 2025 launch cohort hits 12 months. If logo retention is below 60% (net of natural startup failure), the Tome-repeat thesis is confirmed. If above 75%, CRM daily-usage claim holds and the parallel breaks favorably.
  2. Lightfield-specific funding round. No post-pivot raise announced. At estimated $700K–$1.2M/month burn, runway is time-limited. A step-up valuation validates the ARR trajectory. A flat or down round signals the ARR is not materializing.
  3. Attio or Clay ships unstructured-data ingestion. Any announcement of automated email-scan onboarding or unstructured-data CRM records from these competitors compresses Lightfield’s architectural lead. The single most important external signal.
  4. First enterprise logo (>100 seats). Validates upmarket pull before formal product readiness. Any confirmed account at $10K+ ACV meaningfully shifts the ICP-ceiling debate.
  5. Paid conversion rate disclosure. Above 20% from sign-ups with average ACV tracking toward $3,000+/year makes unit economics defensible. Below 10% confirms sign-ups are not converting to revenue at venture scale.
10

Sources

Expert Calls

  • 3 deep Lightfield product calls (1 VC internal evaluator, 1 paying customer / 2 transcripts)
  • Independent AI-CRM landscape evaluators (GTM consultants + operators who assessed Lightfield against peers, no Lightfield referral)
  • Reconciled against Altis flagship AI-CRM research cohort (Attio, Day AI, Monaco, Reevo) + N=70 GTM decision-maker survey

Public Interviews

  • AI Summit: Keith Peiris, CEO (Boardroom Club, Apr 2026)

Web Research

  • 10+ public sources (press releases, analyst reports, product reviews)

All expert call participants anonymized per Altis policy. No management access. No internal financials. ARR not disclosed.

Public References

      11

      Legal Notices

      1. Informational Purposes Only: This report is provided for informational purposes only and does not constitute investment, legal, tax, accounting, or other professional advice. Altis Ventures, Inc. (“Altis”) is not an investment advisor and does not provide financial advice or financial services.

      2. No Recommendation or Solicitation: Nothing in this report should be interpreted as a recommendation, offer, or solicitation by Altis to buy, sell, or hold any security or to pursue any investment strategy.

      3. No Reliance: Recipients may not rely on this report as the basis for any investment decision. Any decisions taken in reliance on any portion of this report are taken at the recipient’s sole risk.

      4. No Guarantee of Accuracy: This report may include information obtained from third party sources, company provided materials, expert interviews, or other external inputs, as well as analyses generated with the assistance of artificial intelligence. Altis makes no representation or warranty as to the accuracy, completeness, or reliability of any information contained herein.

      5. AI-Assisted Analysis: Portions of this report may have been produced using AI. All AI-assisted content is subject to human review, but no assurance is given that such review will identify all issues.

      6. Opinions and Forward-Looking Statements: Any views expressed are those of Altis as of the date indicated and are subject to change without notice. Forward-looking statements are inherently speculative.

      7. No Duty to Update: Altis undertakes no obligation to update this report.

      8. No Redistribution: This report is intended solely for the internal use of the recipient. It may not be reproduced or distributed without Altis’s prior written consent.

      9. No Fiduciary Relationship: This report does not create any fiduciary, advisory, or agency relationship between Altis and the recipient.

      10. Conflicts and Interests: Altis, its affiliates, or their personnel may have relationships with entities referenced in this report.

      11. Warranty Disclaimer: ALTIS DISCLAIMS ALL WARRANTIES, WHETHER EXPRESS OR IMPLIED, WITH RESPECT TO THE REPORT.

      12. Limitation of Liability: IN NO EVENT SHALL ALTIS’S AGGREGATE LIABILITY ARISING FROM THIS REPORT EXCEED $100.

      12

      Thank you

      13