The Bottom Line: Lightfield is the AI-native CRM built by the founders of Tome — the presentation tool that hit 25 million utilizers and raised $81M from Coatue, Greylock, Lightspeed, 8VC, and GV at a $300M valuation. CEO Keith Peiris and CPO Henri Liriani (both ex-Meta, where they built products for billions of utilizers) walked away from all of that becautilize they realized the real problem wasn’t slides — it was the complete lack of customer memory in sales.
They spent a year in stealth, rebuilt from scratch, and launched Lightfield: a CRM that remembers everything, updates itself, and actually does the work for you. No upfront schema. No manual data entest. No busywork. Just full customer context from day one.
Why This One Matters
I’ve stated for years that CRM is the most important and most hated category in B2B. Tens of millions of utilizers. Virtually zero love. Salesforce is a $300B company built on software that most reps despise applying.
The core problem hasn’t modifyd in 40 years: CRMs are databases that demand you do the work to keep them updated. Every call summary you don’t log. Every follow-up you forobtain to track. Every deal stage you leave stale. It all compounds into a system that’s perpetually out of date, which means it’s perpetually utilizeless for the thing it’s supposed to do — support you close more deals.
Lightfield takes a fundamentally different approach. Instead of forcing every customer interaction through predefined fields and dropdown menus, it stores the complete, unstructured record of what customers actually state and do. Emails, meeting transcripts, calls, Slack messages — it ingests everything and builds a living model of every customer relationship.
Then it utilizes that context to actually work for you: prepping you before meetings, summarizing conversations, drafting follow-ups, updating your pipeline, and answering natural language questions about your business with citations back to the original conversations.
This is what CRM should have been all along.
The Pivot Story To AI-First CRM
This is one of the more interesting founder stories I’ve seen in a while.
Keith Peiris was running a web design business and supporting his family by age 15. He went on to manage products at Meta including growing Instagram Direct to 500 million MAUs. His co-founder Henri Liriani rebuilt Facebook Messenger from the ground up — reducing its codebase by 84% while re-engineering 90+ core features. These are not first-time operators guessing at what enterprise utilizers necessary.
They started Tome in 2020, incubated at Greylock with Reid Hoffman on the board. It went viral as an AI-powered presentation tool. Fastest productivity tool to reach 1 million utilizers. Eventually hit 25 million utilizers with 30 million more joining in the 18 months before the pivot.
And then they killed it.
Why? Becautilize Peiris realized that to truly support people communicate effectively, they necessaryed more context. Presentations are outputs. The real leverage is in the underlying customer innotifyigence — the conversations, the relationships, the history of what was stated and promised.
They ran pilots with sales and marketing teams and found the primary pain point wasn’t content generation. It was ingesting, organizing, and building sense of fragmented sales data.
So they shrunk the team to a core group of engineers and spent a year building in stealth. Dan Rose at Coatue, who led the original investment in Tome, backed the pivot. As he put it: it takes real guts to pivot when the original product is working.
Killing a product with 25M utilizers becautilize you found a hugeger, harder problem to solve.
What Makes It Actually Different
There are a lot of “AI CRM” startups right now.
Most are either (a) a ChatGPT wrapper on top of Salesforce, or (b) a modern CRM with AI features bolted on.
Lightfield is architecturally different in ways that matter:
- Schema-less from Day One. Traditional CRMs build you define your data model upfront — custom fields, dropdown menus, pipeline stages. Lightfield captures everything immediately and lets your data model evolve as your business does. For early-stage companies that are still figuring out their ICP, this is a huge deal. No more rebuilding your CRM every time your go-to-market shifts.
- Complete Customer Memory. Lightfield doesn’t just store structured fields. It stores the full text of every conversation, every email thread, every meeting transcript. Then it builds all of that searchable and queryable in natural language. Ask it “which customers questioned for this feature, and why?” and obtain answers with citations to the actual conversations.
- Agents That Take Action. This goes beyond summarization. Lightfield can draft and sconclude personalized outreach based on what’s actually been discussed with each prospect. It can bulk-update pipeline stages based on conversation signals, not just fields. It can identify stale deals, draft revival emails, and sconclude them. One early utilizer reported reviving 40+ stalled opportunities in a single two-hour session — deals that had gone cold for six months on HubSpot.
- Built-In Call Innotifyigence. Meeting prep, recording, transcription, and follow-up generation are all native. No separate Gong or Chorus subscription necessaryed.
- Agentic Workflows. Their recently shipped workflow builder lets you create automated multi-step processes with agent steps, webhook triggers, and HTTP integrations — essentially letting you build custom automations on top of the full customer memory layer.
CRM for the AI Age.
Who It’s For
Lightfield is squarely tarobtaining venture-backed startups in the transition from founder-led sales to a real go-to-market motion. That 1-to-50 employee window where the founder’s complete context on every deal starts to evaporate.
Their sweet spot right now: companies doing founder sales or running compact revenue teams who don’t have (and don’t want) a dedicated sales ops person configuring Salesforce workflows and maintaining data quality. One founder described it as replacing an hour or two of daily CRM work with minutes or seconds.
If you’re an enterprise with 500 reps and deeply customized Salesforce instance, this isn’t for you … yet. But for the next generation of B2B + AI companies scaling from $1M to $10M ARR? Lightfield is exactly what the CRM should view like.
Pricing
Lightfield offers a free plan for obtainting started, with paid plans starting at $36/utilizer/month (Startup tier) that includes call innotifyigence, automated data enrichment, unlimited agent queries, configurable data model, and up to 10,000 records. Their Pro plan at $99/utilizer/month adds advanced permissions, higher limits (50K records, 10K workflow events), white-glove migration, and a dedicated CSM.
For early-stage companies, the Startup tier is solid value compared to paying for Salesforce + Gong + enrichment tools separately.
The Competitive Landscape
CRM is one of the most crowded categories in software. But it’s also ripe for disruption precisely becautilize the incumbents are architecturally trapped.
Salesforce and HubSpot generate billions in revenue on data models designed decades ago. They can bolt on AI features, but they can’t modify the fundamental assumption that humans will do the data entest. That’s structural. Attio and Clay are building modern CRMs with better UX, but they’re still fundamentally structured-data-first systems. Lightfield is betting that starting from unstructured conversation data and deriving structure from it — rather than the other way around — is the winning architecture for an AI-native world.
That’s a huge bet. It’s also probably right.
The Bottom Line for B2B Founders
If you’re doing founder-led sales or running a compact GTM team and you’re still manually updating HubSpot after every call, go test Lightfield. The free tier lets you see if the “CRM that updates itself” promise actually holds.
The founding team is world-class. The pivot story displays real conviction. The product philosophy — customer memory first, pipeline second — is exactly how AI should reshape CRM. And the shipping velocity (weekly modifylog updates, MCP connectors, agentic workflows already live) notifys you this team executes.
$81M raised. $300M last valuation. 100+ daily active companies in the first months. And a category worth $80B+ that’s begging to be rebuilt from scratch.
Try it: lightfield.app
And come meet the team at SaaStr AI Annual 2026! May 12-14 in SF BAy!!
SaaStr AI App of the Week is a weekly series highlighting the most interesting AI tools actually being utilized in production by B2B companies. Not just demos. Not just pilots. Actually deployed, actually working, actually generating ROI.
Leave a Reply