Propose AI isn't one model. It's a stack: a data science foundation that ranks opportunities, an intelligence engine that turns each one into the research a partner can actually use, and an agent that surfaces the daily moves worth making across every account.
Generic AI gives you a paragraph about a client. That's not enough to change a partner's behavior. Propense AI takes the same client, ranks the next service they need, generates the research that makes the case, drafts the email, and surfaces it on the day it matters.
Each layer of our AI builds on the one below it. The data science finds the opportunity. The intelligence engine explains it. The agent makes sure it never gets missed.
Six propensity models trained on your firm's data and the public market. The math that decides which service every client and prospect needs next.
Turns every recommendation into a Why Now, a BD Angle, a draft email, a research report, a BD plan, and a full account plan with a 30-60-90 day playbook.
The daily layer that helps every professional see where they can make a difference today. News, thought leadership pairings, cross-serve alerts, anniversaries.
Six purpose-built propensity models trained on the patterns we see across your firm's clients. They run continuously and update with every change to your time and billing data.
For every client, the services they are most likely to need next - ranked by propensity score, potential fees, and peak sale month.
Funding rounds, leadership changes, acquisitions, regulatory events, and news signals tied to clients and prospects in real time.
Lookalike scoring across your client base. Every client linked to the firms that share their pattern of services, growth, and characteristics.
The path clients typically take through services. The first engagement, the natural next service, and the long-tail expansion services.
Relationship intelligence across your partners' contact graphs. Surfaces the warmest path into a target prospect or executive.
The same engine applied outside the client base. Score every prospect on service-level propensity and prioritize by total expected lifetime value.
A ranked service recommendation is useful. A research-backed brief that explains why now, who to call, what to say, and an outreach draft ready to send is what actually moves the deal. Layer 2 generates four interconnected outputs on demand, on any client or prospect.
Sits inside every recommendation in the platform. Open any client, pick any service, and the intelligence engine surfaces the three things a partner needs before they pick up the phone.
A multi-page research dossier on any client or prospect, generated on demand. The document a partner would build over three days of research, delivered in minutes with cited sources.
A targeted plan focused on a single service recommendation. Delivered directly to a partner's inbox with the angle, the talking points, and a pre-drafted email ready to send.
Lyra raised $485M across three 2025 rounds specifically to accelerate Pulse, Forge, and Atlas - that's $80M+ in qualifying engineering payroll generating uncaptured R&D credits each year. The Dublin and Singapore engineering hubs trigger Section 174's 15-year foreign R&D amortization vs 5-year US treatment - the trap most pre-IPO AI companies miss. Every credit identified now compounds into EBITDA ahead of S-1.
CFO Marcus Wei was hired this year with one explicit mandate: public market readiness. R&D Tax Credits are the cleanest EBITDA lever he has left before the S-1. Federal R&D credit at Lyra's engineering scale typically runs $5-12M annually, plus state credits - and Section 174 mandatory capitalization (now in its fourth tax year) demands a defensible methodology before diligence opens the books.
With Pulse, Forge, and Atlas all launched in the last 18 months by engineering teams in New York, Dublin, and Singapore, your qualifying R&D expenditures likely run $80M+ annually. Federal credit at that base is typically a $5-12M annual benefit, before state credits stack on top.
Section 174 mandatory capitalization is the trap most pre-IPO AI companies miss - particularly the 15-year amortization on foreign R&D from Dublin and Singapore versus 5-year on US spend. Clean treatment is required for the S-1, and the methodology affects your effective tax rate every quarter you're public.
Every dollar of R&D credit flows straight to EBITDA. At your IPO valuation multiple, those credits compound directly into the public-market story you're building - this is the cleanest lever you have left before the S-1.
The complete strategic plan for an account. Built when a partner needs to walk into a meeting, board review, or strategy session with the entire picture in one place.
Discovery call with CFO + VP Tax. Preliminary R&D credit and Section 174 scoping across NY, Dublin, Singapore.
Develop credit methodology + 174 capitalization model. Deliver vCISO engagement proposal.
Launch R&D credit study and 174 documentation. Launch vCISO engagement.
Hatfield is the agentic AI that watches every account, every market signal, and every recommendation for every professional in the firm. It surfaces the six moves worth making each day so partners stop spending hours keeping up with a growing book of business.
Every morning, Hatfield delivers a brief to every partner showing exactly where they can make a difference with their clients today.
Book a working session. We will run a sample of recommendations and generate a real research report and BD plan on a real client before we meet, so you can see the intelligence engine on your data before the longer conversation.