What the Lightfield Startup Program Gives You
Lightfield's startup program offers $1,800 in credits toward Lightfield, an AI-native CRM whose agents capture email, meetings and calls into the customer record automatically, then act on that record rather than handing a rep another task.
AI Perks tracks it in the CRM category alongside $7.7M in credits across 194 companies.
$1,800 is modest next to a six-figure cloud grant, and reading it as a dollar amount is the wrong frame. Lightfield meters two things at once: capture seats and agentic work. The credit converts into very different runway depending on the tier you land on, and the gap between best and worst case is roughly twenty to one.
Working out that conversion before you activate is the whole exercise. Current program terms are listed on getaiperks.com.
What an AI-Native CRM Does That a Spreadsheet or a Legacy CRM Does Not
Every CRM is a shared system of record for a sales process that already exists. The AI-native bet is that the record assembles itself from your inbox, calendar and call recordings, so the data stays complete even when nobody has time to log anything.
That matters because manual entry is the reliable failure point of CRM software. Reps log the deals they are proud of, skip the rest, and within two quarters nobody trusts the pipeline view.
Lightfield is built by the team behind Tome, and raised a $47M round led by Andreessen Horowitz in September 2026. What you are actually buying:
- Capture without logging. Agents pull structure out of email threads, meeting transcripts and calls. This is the only durable way to get activity data, because voluntary logging always decays.
- Full conversational memory. The record keeps the underlying text, not just a summary field, so "what did we actually promise this account" stays answerable.
- A schema that moves. You capture first and shape objects later, which suits a company whose definition of a qualified account changes every quarter.
- Agents that do work. Enrichment, follow-up drafting and sequence execution run as jobs, not as suggestions you still have to action.
The adoption test is unchanged by the AI part: do two or more people need to act on the same relationship without asking each other what happened? If not, your inbox plus a spreadsheet is still correct, and the credit is worth more activated later.
How Lightfield Pricing Behaves at Scale
Lightfield bills on a hybrid meter: capture seats, which are the inboxes and calendars being synced, plus a pooled allowance of workspace credits that agentic work consumes. Your bill therefore tracks both headcount and how aggressively you automate.
Published tiers, as listed by Lightfield:
| Plan | List price | Capture seats included | Viewer seats | Workspace credits |
|---|---|---|---|---|
| Starter | $89 per user per month | 3 | Not included | Chat allowance only |
| Pro | $999 per workspace per month | 5 | Unlimited | ~450,000 per year |
| Growth | $1,999 per workspace per month | 10 | Unlimited | ~900,000 per year |
| Enterprise | Custom | Custom | Unlimited | Custom |
Additional capture seats are published at $200 per month. Credits are consumed by automations, workflows, sequences, per-record enrichment and heavy chat, with overages billed at agreed rates or capped.
Lightfield's published tiers have moved more than once during 2026, so treat these figures as a snapshot. The shape of the meter is the durable part, not the numbers.
Now the arithmetic that turns $1,800 into a date:
| Configuration | Monthly list cost | What $1,800 covers |
|---|---|---|
| One founder on Starter | $89 | About 20 months |
| Three-person team on Starter | $267 | About 6 to 7 months |
| Pro workspace | $999 | Under 2 months |
| Growth workspace | $1,999 | About 4 weeks |
A $1,800 CRM credit is an entry-tier instrument. At the founder-led end it buys most of two years while you work out whether you have a repeatable motion. On a Growth workspace it buys a long free trial.
This is what separates Lightfield from a straight per-seat CRM. Per-seat bills rise when you hire, which you approve and forecast. Credit-metered bills rise when the software works, because the automations that prove their value are the ones you run most.
| CRM cost shape | What the meter counts | Whose growth raises the bill |
|---|---|---|
| Capture seats plus agent credits (Lightfield) | Synced inboxes, plus automation and enrichment runs | Hiring, and your own success at automating |
| Per seat (Attio, Pipedrive, Close) | Users with a paid login | Hiring, which you control |
| Per marketing contact (HubSpot) | Contacts you can email or target | Marketing working, which you do not |
| Seats plus modules (Salesforce) | Users, add-ons and API limits | Both, plus procurement |
AI Perks lists these shapes side by side so the comparison is against the category rather than against whichever logo you saw first.
What Lightfield Credits Stack With, and What They Cancel Out
CRM credits stack cleanly with cloud, model, analytics and billing credits, because those are separate vendors on separate invoices. They do not stack with each other: hold two CRM grants at once and one expires unused, because nobody runs two systems of record.
One trap is specific to AI-native tools. Your Anthropic or OpenAI credits do not offset a Lightfield bill. The inference behind Lightfield's agents runs on Lightfield's account, and you pay for it as workspace credits at Lightfield's rates. Model credits only reduce cost for inference you buy directly. Founders routinely assume otherwise and budget the same dollars twice, which is the sort of thing a tracked list at AI Perks exists to prevent.
Decide deliberately: pick the CRM on the fit of its capture model, then take whichever credit that vendor offers, never the reverse. A $14,000 grant for the wrong system of record costs more than $1,800 for the right one, because migrating two years of history burns a quarter of somebody's year.
What Founders Get Wrong About AI-Native CRM Credits
The expensive mistake is reading "AI-native" as "no work." It is different work. Instead of logging calls, someone reviews what the agents inferred, and if nobody does that review the data is confidently wrong rather than merely incomplete.
Five patterns, roughly in order of cost:
Buying a CRM before there is a repeatable motion. Pipeline software does not create a pipeline. If deals close because a founder knows someone, the CRM documents that and changes nothing.
Rationing capture seats. This is sharper for Lightfield than for a conventional CRM. An unsynced inbox is not a missing login, it is a hole in the record the agents are supposed to assemble. Leave a support lead off the sync and you pay for automatic capture while getting partial capture.
Ignoring the credit meter until it bites. An enrichment job across a large imported list, or a workflow firing on every inbound, can consume a pooled annual allowance in weeks. Set a cap on day one and check consumption monthly.
Migrating the mess. Importing a neglected HubSpot instance gives the agents a corrupted base to reason from. Clean the account and contact objects first, or import narrow and let capture rebuild the rest.
Forgetting the credit is denominated in dollars at a tier you may outgrow. At 70% consumed, price the unsubsidised bill at your then-current headcount and automation volume, then check what else the CRM category offers at getaiperks.com.
What to Settle Before You Take a Lightfield Credit
Decide which tier the credit is for. At published rates the same $1,800 is twenty months or four weeks. Pick that deliberately rather than discovering it at checkout.
Connect real inboxes before you judge the fit. An empty workspace tells you nothing here, because the entire product thesis is what the agents do with real history.
Name the monthly bill you can sustain unsubsidised at your planned headcount and automation volume, then size the tier to land there.
Compare the whole category. AI Perks lists CRM, sales and revenue programs together with current amounts, so the decision gets made against the field.
Frequently Asked Questions
How much is the Lightfield startup program worth?
$1,800 in credits toward Lightfield's AI-native CRM. At its published $89 per user per month entry rate that is roughly 20 seat-months for a solo founder, or six to seven months for a three-person team. On a workspace tier it is closer to a month. Current amounts and eligibility are tracked at getaiperks.com.
What makes Lightfield different from Attio or HubSpot?
Capture and metering. Attio prices per seat and syncs email and calendar. HubSpot prices per marketing contact and bundles marketing tooling. Lightfield bills capture seats plus a pool of credits that its agents spend doing enrichment, sequencing and automation, so the bill tracks automation volume as well as headcount.
Do OpenAI or Anthropic credits offset an AI-native CRM bill?
No. The inference behind Lightfield's agents runs on Lightfield's own infrastructure and is billed to you as workspace credits at Lightfield's rates. Model credits only reduce cost for inference you purchase directly from a provider. The two stack rather than substitute, which is exactly why holding credits across several layers beats holding more in one.
Is $1,800 enough to actually evaluate a CRM?
Yes, at the entry tier, and that is the point. Six to seven months with a small team is long enough to load real pipeline, watch the capture quality and see whether the schema holds. A four-week workspace trial is not. Compare the CRM programs at getaiperks.com.
When is a startup too early for an AI-native CRM?
When one person holds every relationship in their head and nothing breaks. The trigger is handoff, not deal count: the first time a second person needs the current state of an account without asking. Take the credit early if approval is slow, then activate it when that handoff starts happening.
What happens when the Lightfield credits run out?
You inherit a list-price bill at your then-current seat count and automation volume, with months of captured history that makes switching expensive. Model that number while the credit is still running and cap credit consumption before it surprises you. AI Perks tracks $7.7M in credits across 194 companies at getaiperks.com.
Pick the capture model on its merits. Let someone else pay for the first year of filling it.