# Waii, plain-language SQL, and the Data 360 rename

Waii turns a sentence into SQL. The sentence is not a grant to read every table in the warehouse.

**Published:** 2026-08-12
**Canonical:** https://synconai.com/waii-data-360
**Author:** SynconAI Architecture Team

## In short

Salesforce completed the Waii acquisition on 15 August 2025. Waii turns plain language into SQL, and Salesforce planned to integrate that work into Data Cloud. On 14 October 2025 Data Cloud was renamed Data 360. A sentence is not permission to read every warehouse table. Scope the tables, the grants, and who may ask, before anyone prices an implementation.

## Sources

- **Salesforce Signs Definitive Agreement to Acquire Waii** (retrieved 2025-08-15)
  https://www.salesforce.com/news/stories/salesforce-signs-definitive-agreement-to-acquire-waii/
  > Salesforce completed its acquisition of Waii on August 15, 2025.
- **About Salesforce Data 360** (retrieved 2025-10-14)
  https://help.salesforce.com/s/articleView?id=data.c360_a_data_cloud.htm&type=5
  > As of October 14, 2025, Data Cloud has been rebranded to Data 360.

## What to take away

- The close on 15 August 2025 covers plain language into SQL, planned for Data Cloud. It is not a grant on your warehouse.
- Data Cloud was renamed Data 360 on 14 October 2025. Update the name on the diagram. The rename does not change who may query a table.
- A natural-language question still runs as an identity. If that identity can read a table, the SQL can read it.
- Price the catalogue, the grants, and the refusals. Do not price a promise that any colleague may ask anything.

## The guide

A finance analyst types a sentence into a chat box: show me every customer and what they spent, including the tables the data team said were not for general use. The reply comes back with a SQL statement and a result. The analyst forwards both to a shared channel. The data team finds out on Monday, from the channel, that a table they had closed to the business is now in a screenshot.

The sentence felt like a question. The warehouse treated it as a query. Those are not the same object, and the acquisition does not collapse them.

## What closed on 15 August 2025

The close notice is one sentence: "Salesforce completed its acquisition of Waii on August 15, 2025." Waii turns plain language into SQL. Salesforce planned to integrate that work into Data Cloud. That is the direction. It is not a description of your schemas, your roles, or which tables a seller may touch.

A planned integration is not the same as a switch that is already on in every org. If someone asks whether their Data Cloud tenant gained a natural-language box on 15 August 2025, this source does not say that. Confirm what your contract includes. Do not treat a completion sentence as an entitlement.

The useful reading is narrow. A person may, in time, ask in ordinary language and receive SQL, or the result of SQL, aimed at data Salesforce planned to reach through Data Cloud. Your job is to decide which tables that path may see, under which identity, before anyone treats the box as a convenience feature.

## Data Cloud was renamed Data 360

On 14 October 2025, Data Cloud was renamed Data 360. Salesforce's help text says: "As of October 14, 2025, Data Cloud has been rebranded to Data 360." Put the current name on the diagram, in the statement of work, and in the briefing you give a new analyst. Old labels may still appear in an application or a document during the transition. The rename does not add a table, remove a grant, or decide that a question is allowed.

If your project vocabulary still says Data Cloud, you will confuse the quote and the people reading it. You will not, by updating the word, have solved access. A slide that says Data 360 and then shows every schema in the warehouse has used the new name to hide an old mistake. Fix the name. Then fix the list of tables.

Do not let the two dates blur. Waii closed on 15 August 2025, while the product was still being described as Data Cloud. The rename landed on 14 October 2025. A design written between those dates can be right about the intent and wrong about the label. Update the label. Re-read the grants. They do not update themselves because the product changed its name.

## A sentence is not a permission

Plain language is a way to ask. SQL is a way to read. The right to read sits on the identity that runs the SQL: a user, a role, a service account. If that identity can select from a table, a well formed question can return those rows. If you wanted the table closed, the question is not the control that closes it. The grant is.

This is the line people skip because the interface is polite. The analyst did not ask to "bypass security". They asked a reasonable business question. The tool did what it was built to do: turn the question into SQL and run it with the rights it had. The failure was upstream. Someone gave the path a broad role so the demonstration would return rows. The demonstration then became the way the business reads the warehouse.

Write the identity down before the first question is tried. For an analyst, that may be their own warehouse role, so they see what they already see in the BI tool. For a wider audience, a seller or a support lead, it should be a narrower role that can read the tables you have named and nothing beside them. "The Waii user" is not a design. A role called All Tables Read is a design, and it is the wrong one if the audience is the whole company.

Natural language also hides which tables were used. A SQL statement can be reviewed. A sentence cannot, unless you keep the SQL beside the answer. Require that. If the path returns a number and withholds the statement, you cannot tell whether it read the closed table, the approved mart, or both. Store the statement with the question, under the same retention rule as the result.

## What the migration keyword gets wrong

"Waii data migration to salesforce quote" sounds like a load of Waii data into Salesforce. This source does not describe a Waii database of business records you extract. It describes a capability: language in, SQL out, planned for Data Cloud, which is now Data 360. Migrating that capability is not a row-level project unless you separately have logs, saved questions, or semantic models you intend to keep.

Inventory those artefacts if they exist. Saved questions. Semantic models or metric definitions the team maintained. Connection strings and the roles they used. Logs of who asked what. If the list is empty, say so. An empty list means the quote is about access design and a new path, not about a data load. Do not invent a staging object so the migration services line on the proposal looks familiar.

If you do have saved questions, treat them as queries, not as content. Each one needs a reviewer who can read SQL and a decision: still valid, needs a new grant, or must be retired because it reached a table you will not expose. Bulk-importing saved questions into a new tool, with the old role attached, copies the Monday failure into the new name.

## What to scope before a price

The catalogue. A named list of tables, views, or marts the path may use. Each line has an owner. Tables that are absent are out of scope, even if a question would be easier to answer with them. Adding a table is a change, with the owner's yes, not a prompt tweak.

The grants. The role the path uses, the tables it may read, and the columns it may not. Column rules matter for pay, health, account numbers, and free-text notes that have nowhere else to live. A table can be in the catalogue and still have columns closed.

The audience. Who may ask. A pilot of five analysts is a different risk from a Slack channel of two hundred. Write the audience down. Growing it is a new decision, because the same SQL in a larger room is a publication.

The refusals. Five questions that must not return rows. A closed schema. Another region's payroll. A raw staging table. A question that would need a table you have not listed. A request to show "everything we have on this customer" when the catalogue only holds a mart. Test them with the role you will keep.

The review. A person who can read SQL looks at a sample of statements each week during the pilot. They are checking for tables outside the catalogue, not for elegance. If they cannot keep up, the audience is too wide or the log is too thin.

The stop. When the path cannot answer from the catalogue, it says so. It does not reach for an unlisted table because the question was clear. Clarity is not authorisation.

## Cost is the list you can point at

There is no figure on this page for a Waii implementation, and there should not be. The announcement has no price, and your warehouse is not the announcement's warehouse. Ask a partner to price these lines.

- The catalogue, with an owner per table or mart.
- The role, including columns that stay closed.
- The audience, and the room where answers may appear.
- Five refusals, tested, with the SQL or the empty result kept.
- A decision to store the generated statement with the answer.
- Retention for questions, statements, and results, agreed with privacy.
- A review cadence for the pilot, and the name of the person who reads the statements.
- Time from warehouse owners to confirm each connection after the close.

A single number in a proposal, with the catalogue still blank, is a number for a different project. The one you have is an access project that happens to speak SQL. Pay for the list. Do not pay for a slogan about asking anything in plain language.

## The test, not a ranking

"Best waii salesforce integration agency" is the wrong question. Nobody is the best agency in the abstract. The firm that passes can do the following on your warehouse before you sign.

They ask for the catalogue and stop when you do not have one. They refuse a role that can read every schema so the demo looks complete. They can show a question that returns nothing, and they can explain the grant that made the nothing correct. They keep the SQL. They use the name Data 360, and they do not pretend the rename widened access. They will not tell a buyer that a natural-language question is permission to read every warehouse table.

"Hire waii salesforce consultant" does not change the test. A lone consultant who connects the warehouse with an admin role, so you can see a result in the first hour, has failed on purpose. An implementation partner who talks about migration services before they have asked whether you even have saved questions has brought the wrong template.

If the room cannot pass, you are earlier than a statement of work. Finish the catalogue. Then hire.

## What has to be true on Monday

| Check | What you write | A pass looks like |
| --- | --- | --- |
| Name | Data 360 on the diagram | No live design still titled only Data Cloud |
| Catalogue | Tables or marts in scope, with owners | Unlisted tables are unreachable |
| Role | The identity that runs the SQL | It matches the audience, not the demo admin |
| Statement | SQL stored with the answer | You can see which tables were read |
| Refusals | Five questions that must fail | They fail under the role you will keep |
| Audience | Who may ask, and where the answer lands | A wider room is a new decision |

:::warn
A natural-language question is not permission to read every warehouse table. The question becomes SQL. The SQL runs with a grant. If the grant is broad, the politeness of the sentence will not make the result narrow.
:::

Between the August close and the October rename, teams wrote "Data Cloud" on purpose. Those documents are not wrong about history. They are wrong as a label for work you start now. Change the label, then re-check the role. A renamed box with the old admin connection is the Monday screenshot waiting to happen.

Bring the catalogue to the first workshop on paper, not as a hope that the partner will infer it from a connection string. A connection string proves someone once had a password. It does not prove the table is still the one finance trusts, or that the column with salary data was meant to travel. If the paper is not ready, the workshop is a working session to draft it, and it ends when the blanks are visible. It does not end with a role created so the sentence returns rows.

An earlier machine-learning acquisition is a different shape, and it is easy to confuse with this one if you are scanning a list of old product names. Read [PredictionIO after the 2016 acquisition](/predictionio-salesforce) before you treat every acquired data name as a current Salesforce feature.

## The other searches

The project quote is on [waii implementation cost](/waii-salesforce-implementation-cost). That page does not print a figure. Who bought the company is on [what happened to waii](/what-happened-to-waii). The connector is on [integrating waii with Salesforce](/integrating-waii-with-salesforce). Whether to move is on [should I migrate from waii](/should-i-migrate-from-waii). The customer fee is on [waii licence cost](/waii-salesforce-licence-cost). Who to hire is on [hiring for waii](/hire-waii-salesforce-partner). This page does not rank an agency.


## Questions

### Does a plain-language question let someone read every warehouse table?

No. Waii turns plain language into SQL. The SQL runs as an identity, and that identity is the limit. Salesforce completed the acquisition on 15 August 2025 and planned to integrate the work into Data Cloud. None of that is a grant on your warehouse. If a demo role can read every schema, the demo is the defect. Narrow the role and test a question that must return nothing.

### What did the Data Cloud rename change?

On 14 October 2025, Data Cloud was renamed Data 360. Use that name on the diagram and in the statement of work. The rename does not decide which tables a question may read, and it does not by itself finish the integration Salesforce planned. Re-read the grants after you update the label. A new name on a broad role is still a broad role.

### What should a Waii implementation cost cover?

Not a single invented figure. Price a catalogue of tables, the role that may read them, the audience that may ask, five refusals you have tested, storage of the generated SQL beside the answer, and a named reviewer for the pilot. A quote for a data migration, before you know whether any saved questions exist, is a quote for an object this source does not describe.

### How do we test a Waii Salesforce consultant?

Ignore the claim to be the best agency. Ask them to refuse an all-tables role, to show an empty result for a closed schema, and to keep the SQL. They should say Data 360 and still talk about grants. A consultant who connects an admin role so the first hour produces rows has failed the only test that matters.

### Should we migrate saved Waii questions into Salesforce?

Only after a person who can read SQL has marked each one: still valid, needs a new grant, or retired. Saved questions are queries, not content to load in bulk. If you have none, the project is access design for a new path, not a migration. Say that in the statement of work so a migration line item cannot appear by habit.

### Is waii owned by Salesforce?

Yes. Salesforce completed the Waii acquisition on 15 August 2025. The full account is at https://synconai.com/what-happened-to-waii.

### When did Salesforce acquire waii?

15 August 2025. Salesforce planned to integrate it into Data Cloud, renamed Data 360 on 14 October 2025.

### What happened to waii after the Salesforce acquisition?

Waii turns plain language into SQL. A question is not permission to read every warehouse table.

### Should I migrate from waii to Salesforce?

Do not copy the warehouse into Salesforce so a sentence can see it. Name the tables the question may use.

### What is the waii licence cost, or the price Salesforce paid?

The page this guide cites does not publish a purchase price or a licence fee. The licence read is at https://synconai.com/waii-salesforce-licence-cost.

---

SynconAI is a Salesforce Select Partner and an OpenAI Select Partner, delivering architect-led Salesforce and AI programmes across the United States and Australia. Related service: Data 360 (https://synconai.com/salesforce-data-360-services.html).