Starch vs ChatGPT. Get RFP responses written in your template and your voice.
Starch keeps watching and drafting in the background.
Starch is set up around your firm's past responses, templates and approval steps, and keeps working between requests. When an RFP posts, your banker sees what posted and that the draft has already started. ChatGPT can stay your firm's assistant for everything else.
Facts about ChatGPT checked
For public finance banks
What Starch is set up to do for your bankers.
A banker and an engineer from Starch set Starch up around your past responses, templates and approval steps, and keep it running.
Set up around your work
Starch reads your past responses and templates, so the draft comes out in your format and your firm's voice.
Working in the background
Starch watches the issuers you choose, from 500 to 35,000, and starts the draft when an RFP posts, then brings the work to the banker who covers that issuer.
Approved by a named banker
Nothing goes out until a named banker approves it, and your compliance steps stay in the route.
Built and maintained for you
A banker and an engineer from Starch set up the first workflow with your team and stay until it runs the way your team works.
How one RFP moves through Starch.
Starch does the first eight steps. Your bankers review the draft in Word and give final approval.
- StarchFinds the RFP
- Scores it
- Alerts your team
- Reads past work
- Maps every question
- Writes the draft
- Writes up your analysis
- Checks it
- Your teamReview in Word
- Final approval
For school districts
What Starch does for a school district.
Your business office and your board get one place to ask about your district's debt, bond measures and tax rate.
One place for your debt and taxes
Starch keeps every bond, bond measure, rating and tax rate current, read from your official statements, audits, ballot measures and county reports.
Plain-English answers
Ask about your finances the way you would ask a colleague. Starch does the analysis and explains each term in the answer.
Board pages your team approves
Starch drafts staff reports and board-ready pages, and your team reviews each one before it goes to the board.
Starch turns a board request into a staff report.
- The board asksOne line in the minutes
- Questions listedWhat the request needs
- AnsweredStarch does the analysis
- Report draftedIn your agenda format
- Your CBO signsThen back to the board
Side by side
Use ChatGPT and Starch together.
Your people keep ChatGPT as the firm's assistant. Starch runs the public finance work that repeats.
The detail
Starch and ChatGPT, row by row.
Facts about ChatGPT come from press reports of OpenAI's launch and OpenAI's published business pricing.
For a public finance bank
| Topic | Starch | ChatGPT |
|---|---|---|
| What it is | Public finance workflows that Starch builds for your bankers and keeps running. | A general assistant, with a financial-services tier for institutions that qualify. |
| Built for | Public finance banks. | Large banks and investment firms. Built with Morgan Stanley and Evercore. |
| How you get it | A banker and an engineer from Starch set it up with your team. | An enterprise account and institution clearance, through direct sales. |
| Where the information comes from | Portals, board materials, filings, ratings, your past responses and your email and calls. | Daloopa, PitchBook, Crunchbase and LSEG News are preloaded, with roughly fifty connectors. |
| Recurring work | Keeps watching issuers and drafting in the background, then brings the work to the right banker. | Scheduled tasks run unattended, including when an event triggers them. |
| Output | RFP responses, pitch books and orals decks in your templates, reviewed in Word, Excel and PowerPoint. | PowerPoint, Excel and dashboards, with branded templates. |
| Deployment | Single tenant, in your own environment or a dedicated cloud. | OpenAI's infrastructure, with SAML SSO, SCIM, role-based access and configurable retention. |
| How it is priced | Enterprise pricing. | Business seats are $25 monthly or $20 billed annually, and premium seats are $125. The finance tier has no published price. |
For a school district
| Topic | Starch | ChatGPT |
|---|---|---|
| What it is | A place for your district's debt, measures, ratings and tax rates, with questions answered in plain English. | A general assistant for the whole firm. |
| How you get it | Tell us which district you would like to see, and we open it with you on a call. | An enterprise account, through direct sales. |
| Where the information comes from | Your official statements, audits, ballot measures, county reports, EMMA and the rating agencies. | Daloopa, PitchBook, Crunchbase and LSEG News are preloaded, with roughly fifty connectors. |
| Output | Staff reports and board pages that your team reviews before anything goes out. | PowerPoint, Excel and dashboards, with branded templates. |
| Deployment | Private to your district. Documents never train a model. | OpenAI's infrastructure, with SAML SSO, SCIM, role-based access and configurable retention. |
| How it is priced | Enterprise pricing. | Business seats are $25 monthly or $20 billed annually, and premium seats are $125. |
An honest answer
When ChatGPT is the better fit.
Choose ChatGPT if you want one assistant across every department, a single vendor and enterprise controls like SAML and SCIM.
One vendor for the whole firm
One vendor and one contract can cover legal, compliance, HR and the business lines.
Preloaded data and connectors
Daloopa, PitchBook, Crunchbase and LSEG News come preloaded, and your existing Bloomberg or FactSet can connect.
Built with two banks
Morgan Stanley and Evercore shaped the finance tier around real institutional work.
Before you book
Asked about ChatGPT.
Can we just use ChatGPT for this?
ChatGPT is a good fit for drafting and analysis across the firm, and its scheduled tasks run unattended, including when an event triggers them. Starch runs the public finance work for your bankers: Starch watches the issuers you choose, writes each RFP response new and brings it to the banker who covers that issuer. A Starch team builds and maintains it.
How does getting started work?
A banker and an engineer from Starch learn how your team works and set up the first workflow with you, usually the RFP response. We meet weekly for the first three months, then monthly or quarterly as needed. For a district, tell us which one to open, and we'll walk through it on a call.
How much does ChatGPT for Financial Services cost?
OpenAI publishes no price for the finance tier, which needs an enterprise account and institution clearance through direct sales. Business seats are public: $25 per seat monthly, $20 billed annually, premium $125. Those prices are for general seats. The finance tier has no published price.
Who reviews what Starch produces?
At a bank, a named banker approves it before anything goes out. At a district, your business office approves what goes out.
Where does our data live, and does it train a model?
A bank's Starch runs as a single tenant, in your own environment or a dedicated cloud, and a district's deployment is private to that district. Starch reads your documents to do the work, and your documents are never used to train a model.
See it run
Bring a real RFP.
Or pick a district to open.
Bring an RFP and a few of your past responses. We'll run it with you on the call, from the posting to the draft your banker approves.
Sources and methodChecked 2026-09-16
Everything stated about ChatGPT on this page comes from the material below, checked on 2026-09-16. Pricing and product detail change; confirm on their site before deciding.
- Fortune: OpenAI courts Wall Streetthe finance tier, its design partners and its data sources
- CIO: premium ChatGPT Business seatspublished seat pricing
- OpenAI: scheduled tasks in ChatGPTscheduled and event-triggered tasks
ChatGPT is a trademark of OpenAI. Starch is not affiliated with, endorsed by or partnered with OpenAI. OpenAI blocks automated access to its public sites, so the product details on this page rest on press reports rather than OpenAI's own pages; confirm with OpenAI before deciding.