Ask Zia and Zia Skills for a sales team: what this guide covers
This guide explains how to use Ask Zia and Zia Skills for a sales team working in Zoho CRM. It covers what a salesperson can ask, what comes back and what you need to configure so the answers are right. It also covers what to tell your team so they use the assistant instead of ignoring it.
It is part 3 of 3 in our series on customising Zoho CRM with Zoho's own AI, and it closes the series. Part 1 covered writing Deluge code with Zia and checking it before you deploy. Part 2 covered setting up Zia lead scoring for a UK sales team.
If you want a wider view of the AI first, our post on what Zia actually does for a small sales team is a shorter read. This post goes further into the daily questions and the setup behind them.
The central point is simple. An AI assistant answers from your data and your field names. A tidy Deals module with clear column names gives useful answers. A module full of blank fields and vague labels gives confident answers to the wrong question.
The terms you need before you start
Zoho uses the name Zia for several related features, which causes confusion. Here is what each term means in this guide.
- Zia is the AI built into Zoho's applications. In Zoho CRM, Zoho says Zia "can create modules, reports, and workflows in seconds, as well as retrieve any data you ask for".
- Ask Zia is the conversational part: you type a question in plain language and get an answer. In Zoho Analytics, Zoho's reporting product, the help centre describes it as a search-driven analytical assistant that creates reports from questions asked in natural language.
- Zia Insights is a Zoho Analytics feature that writes short narratives about important trends in your data, without you asking.
- Zia Skills is the name used when building answers for a Zia chatbot in Zoho SalesIQ, Zoho's website chat product. It faces your website visitors, not your internal sales reports.
- Column means a field in your data, such as the amount on a deal or the name of the deal owner. Zoho Analytics calls them columns because it stores CRM data as tables.
Keep these apart when you brief your team. A salesperson who expects the website chatbot to answer pipeline questions, or expects Ask Zia to talk to website visitors, will be disappointed by both.
What a salesperson can ask Zia and get back
The most common daily use is retrieval: asking for data you would otherwise dig for through filters and list views. Zoho says Zia in the CRM can retrieve any data you ask for and can also build reports and workflows on request. For a salesperson, that means questions about their own pipeline, open deals and recent accounts.
In Zoho Analytics, Ask Zia goes further into numbers. Zoho says it answers questions with KPIs and visualisations, and those KPIs can be added to an existing dashboard or used to build a new one. It also supports forecasting: Zoho invites you to "ask Zia anything about future trends, predict scenarios, and plan ahead of time". If your team is new to the idea, our glossary entry on sales forecasting explains the basics.
A real example from Zoho's own page shows the pattern. A Business Intelligence Manager at Renu Energy Solutions says: "I can ask for a specific agent name for their quarterly or monthly performance and get to see if their sales numbers are trending up or down." That is the kind of question a sales manager asks every week.
Typical questions that work well
- A named rep's deal value this month compared with last month.
- Deals in a given stage with no activity this week.
- Accounts created last quarter in a given region.
- Where total pipeline is heading for the rest of the quarter.
Each of these depends on a field that is filled in consistently. That is the theme of the rest of this guide.
The suggestions Zia offers without being asked
Besides answering questions, Zia in Zoho CRM puts suggestions in front of salespeople. These matter for daily use because they appear on records whether or not anyone opens a chat window.
- Zia scores point to the prospects most likely to convert, so reps can prioritise them.
- Next best experience identifies the next step for open deals and leads, such as sending an email or scheduling a call or meeting.
- Similarity recommender refers to previously closed records with similar attributes, showing how those deals went through each stage.
- Field prediction lets you build custom predictions, such as the likelihood of winning a deal or its expected revenue, based on a record field you choose.
- Churn scores flag customers at risk. For subscription-based records, Zia also indicates which product or service a customer might churn from.
- Anomaly detection notices events that deviate sharply from expected trends and notifies the right people.
- Data enrichment gathers extra lead and customer details from the internet and extracts information from email signatures.
- AI for calls transcribes call recordings into plain text and draws out customer sentiment, intent, emotion and a summary. A dashboard shows your monthly minute limit, minutes spent and minutes remaining.
Most of these features support lead management, and every one of them reads from your records. A prediction built on a field that half the team leaves blank will be weak, however clever the model behind it.
Where the assistant lives: the surfaces your team will use
Your team will meet Zia in more than one place. Knowing which surface answers which question saves frustration.
Inside Zoho CRM
This is where reps work records all day. Use it for retrieval, quick reports and the record-level suggestions described above.
Inside Zoho Analytics
Zoho's instruction is direct: "click the 'Ask Zia' button on the left navigation bar and ask your questions to Zia." Zoho Analytics reads data from Zoho CRM, Zoho Books, Zoho Desk, Zoho SalesIQ, QuickBooks and Google Ads, so it suits questions that cross systems. It is also a good place for AI for reporting. Zoho notes that Zia "runs inside the platform, so there's no separate AI tool to license."
On mobile
Zoho says Zia in Zoho Analytics is available on mobile devices as a chat-like interface. That suits a manager checking numbers between meetings.
Embedded in your own site or application
Ask Zia can be embedded into your own application or site. With Access with Login, users must log in to Zoho Analytics, and the tables they need must be shared with them before Zia can build reports for them. Access without Login embeds Ask Zia publicly and works only in public workspaces, so it is not the option for internal sales figures.
Zia Skills: what to expect and where the limits are
Zia Skills is used to build the answers a Zia chatbot gives in Zoho SalesIQ. For a sales team, that chatbot is often the first contact a website visitor has with you. It can answer simple questions and qualify interest before a rep picks up the conversation.
The pack's evidence on Zia Skills is a single thread in Zoho's own community, so treat it as one user's experience. The poster says they "managed to program the single answers" in their SalesIQ bot, offering five answer options, but could not work out whether conditional answers were possible. Conditional routing means sending a visitor down a different path depending on what they choose.
The poster reports that Zoho said this can be done but did not provide sufficient documentation. According to the poster, Zoho told them to write the routing themselves or pay a Zoho partner to build it. They add that a partner estimated 15 hours to code one conditional routing, and that Zoho could not tell them which programming language Zia Skills uses.
The practical lesson is to start with single answers, which the poster did build without help. If you need branching, get a written scope and estimate before committing. Also note that the same community page lists a separate developer post on the "Zoho CRM Skill", a file used with GitHub Copilot to generate complex COQL queries. That is a different thing from Zia Skills, despite the similar name.
Why clean data and field setup decide the answers
Ask Zia does not know what your business calls things. If a rep asks for "deal value" and the column is called "Amount", Zia has to guess. Zoho Analytics gives you three settings to remove the guessing, grouped under Ask Zia's Data Settings.
- Synonyms map natural-language terms to the column that holds the data. You can specify any number of synonyms for a column.
- Column Priority tells Ask Zia which column to rank first when several have similar names. You set the priority from high to low.
- Default Function sets the summary function, such as Average, Minimum or Maximum, applied to a column by default. It overrides the function Ask Zia would otherwise choose, and a user can still override it by naming another function in the question.
You open these settings from Table Edit Design or from the Column Settings of a column.
At Svennis we find that most wrong answers from an AI assistant trace back to a vaguely named field or a picklist nobody maintains. So we fix field names and clear out unused fields before anyone is trained, then add synonyms for the words the sales team actually uses.
Settings only help if the underlying records are sound. Duplicates inflate counts, blank owners hide deals from a rep's totals, and stale stages distort forecasts. Our guide on how to customise your CRM covers field design in more depth.
A worked example: teaching Ask Zia your sales language
Suppose your sales director wants reps to ask Ask Zia in Zoho Analytics for their deal value by month. The Deals data from Zoho CRM is synced into Zoho Analytics as a table. It has a column called Amount, another called Expected Revenue, and a Discount column.
Before any setup, a rep types "my deal value last month". Ask Zia may not link "deal value" to Amount, or it may pick Expected Revenue because the name looks related. The rep gets a number, but not the one the director means.
- Open the Deals table in Table Edit Design and open Ask Zia's settings.
- On Amount, add Synonyms your team uses, such as "deal value", "order value" and "sale size".
- Set Column Priority on Amount to high and on Expected Revenue to low, so questions about value go to the booked figure first.
- On Discount, set the Default Function to Average, so "discount last quarter" returns the typical discount, not a single extreme.
- Test with the phrases reps actually use. A rep who wants the largest discount can still ask for the "maximum discount", which overrides the default in that question.
The outcome is that the same question returns the same figure for every rep and for the director. That consistency is what earns the assistant a place in the weekly routine.
Setup checklist: what to configure and why
Use this table as the order of work. Each row is a setting or task named in Zoho's own pages, with the reason it affects the answers your team gets.
| Task | Where | Why it matters |
|---|---|---|
| Clean duplicates and fill key fields | Zoho CRM records | Every answer and prediction reads from these fields |
| Add synonyms for business terms | Zoho Analytics, Ask Zia settings | Links the words reps use to the right column |
| Set column priority | Zoho Analytics, Ask Zia settings | Stops Zia picking a similarly named column |
| Set default functions | Zoho Analytics, Ask Zia settings | Controls whether a figure is averaged, a minimum or a maximum |
| Share tables with users | Zoho Analytics, embedding with login | Zia cannot build reports from tables a user cannot see |
| Check the language setting | Zoho Account | Zia uses the account language by default |
| Scope Zia Skills answers | Zoho SalesIQ | Single answers first; branching may need custom work |
Work top to bottom. Tuning synonyms on a table full of duplicates only makes wrong numbers easier to reach.
What to tell your team so they actually use it
A sales team ignores a tool that gives one wrong answer in its first week. So introduce Ask Zia after the setup work, not before, and set expectations plainly.
Give them a starter list of questions
Write down a short set of questions you have already tested, using the phrases you added as synonyms. Reps copy what works. A tested list removes the blank-page problem of a chat box.
Explain what it reads from
Tell reps that Zia answers from their records. If they leave the amount or the stage blank, their own totals will be wrong. This links data entry to something they care about, which works better than a reminder from management.
Say which surface to use for what
Record suggestions and quick lookups sit in the CRM. Cross-system figures and trends sit in Zoho Analytics. Website visitor questions belong to the SalesIQ chatbot built with Zia Skills.
Keep a feedback route
Ask reps to report any question that returned an odd answer. Each report usually points to a missing synonym or a field problem, and fixing it improves the answer for everyone. Review these at a regular team meeting until the reports dry up.
What this means for a UK company
For a UK sales team, language support is straightforward. Zoho says Zia works in English, with Spanish in beta, and by default it uses the language set in the Zoho Account. Ask Zia's Conversation mode does not yet support Spanish, which matters only if part of your team works in Spanish.
The bigger UK point is vocabulary. Your team may say "turnover", "order book" or "renewals", while the columns carry other names. Synonyms are the setting that bridges that gap, so build your synonym list from the words your own reps and directors use in meetings.
Two Zia features touch personal data directly. Data enrichment gathers details about leads and customers from the internet, and AI for calls transcribes recorded conversations. Before switching either on, check with whoever handles data protection how they fit your UK GDPR obligations and your privacy notices.
Finally, remember that Zoho Analytics runs Zia inside the platform, so you do not need to license a separate AI tool for these questions. The cost that matters most is the time to clean data and configure fields, not an extra subscription.
Practical next steps
This closes the three-part series. If you work through it in order, you will have Zia writing and checking code, scoring leads and answering your team's daily questions from data you trust.
- List the questions your sales team and directors ask most often, in their own words.
- Check the fields those questions depend on for blanks, duplicates and stale values, and fix them in Zoho CRM.
- In Zoho Analytics, add synonyms, column priority and default functions for those fields, then test each question.
- Decide which Zia suggestions to switch on, and review data enrichment and call transcription against your data protection position.
- For the website chatbot, build single Zia Skills answers first and get a written estimate before any conditional routing.
- Give the team a tested question list and a way to report odd answers.
If your field setup needs more work than you can do in-house, our Zoho CRM implementation and consulting page explains how we approach a CRM build and clean-up.