Gabriel de Chazournes, Sales Director at Ankorstore, explains how the company uses Claude to make part of its stack accessible from a single interface. Modjo and BigQuery feed workflows for meeting preparation, follow-ups, data analysis and sales coaching, among others.
At Ankorstore, using AI in the sales teams is not about replacing existing software. Rather, the company is looking to make better use of the data that software contains, without forcing Sales to constantly jump from one tool to another.
This approach relies in particular on Modjo, BigQuery and Claude.
Today, 100% of calls are recorded in 
Modjo. The data from these conversations can then be combined with data from 
BigQuery and used from 
Claude.
In practice, every morning, Sales receive a brief on their meetings for the day. After a call lasting more than 10 minutes, a follow-up draft is automatically prepared within ten minutes. And every week, a personalized coaching email for each salesperson is automatically generated for the managers. This email is reviewed before being sent to the ICs and used as the basis for coaching in 1-1s. It makes it possible to review 100% of the week's calls and bring out the IC's strengths and weaknesses, rather than focusing on one particular call with a 100% manual review by the manager.
Gabriel sums up the place Claude has taken in his team as follows:
“Today, I'd say Claude centralizes 80% of the tools we use on the sales side.”
Why did Ankorstore centralize its conversations in Modjo?
At first, the problem was fairly simple: conversations with partners were scattered across several tools.
Sales used Modjo, while some Product Managers worked with 
Fellow. It was therefore hard to find all the exchanges with an account in one place.
Product Managers now have Modjo accounts too, with different licenses. But above all, Modjo data can be queried from Claude via MCP.
Before a meeting, Gabriel can for example ask who last spoke with an account, how often it has been contacted, or how many different people have talked with it over the last three months. All information that previously had to be pieced together by navigating between several sources.
Ankorstore has also long used 
Aircall for telephony. Conversations flow into Modjo, and the information can then be used in Claude.
The idea, then, is not to replace specialized software. Aircall remains the telephony tool, Modjo the conversation intelligence tool. Claude instead adds an interface for querying and cross-referencing their data.
How does Ankorstore connect Claude to BigQuery?
Data from conversations is only one part of the setup. Gabriel has also built a local connection for querying BigQuery data from Claude.
He explains that he installed a local MCP giving access to the database on his sales team's machines. A skill knows the structure of the tables and helps Sales find the information they need.
The benefit shows up especially during meetings. A brand or a retailer may ask a very specific question that no dashboard answers directly. Rather than looking for the data manually or calling on the Data team, the salesperson can phrase the question in natural language.
Ankorstore also uses this connection to generate personalized decks for its partners, with fixed slides (Business Model, Ankorstore market data) and slides tailored to the pain points identified during discovery and to the partner's market.
A connector has since been launched, but its use still has to be approved internally, notably with the Data team.
Can Claude replace some of Sigma's uses?
Ankorstore currently uses 
Sigma for data visualization. Gabriel estimates that the company has several hundred dashboards.
For him, the limit of the dashboard shows up mainly when a question was not anticipated.
“If you don't have the data in the table and you don't know where to look for it in the Sigma dashboards, it immediately gets complicated.”
Conversational access changes this logic. The salesperson no longer necessarily needs to know which dashboard to look in: they can just ask their question directly.
The team already uses Claude's Artifacts to produce certain visualizations. Gabriel therefore expects that some traditional BI uses could gradually move to this type of interface, provided that data access is sufficiently governed and no longer depends on a local installation on the machines.
How does Ankorstore automatically prepare its salespeople's meetings?
One of the most advanced workflows combines Modjo, BigQuery and Claude.
Every morning, Sales receive an email brief on their meetings for the day. The system pulls both the data available in BigQuery and the transcripts of the latest calls recorded in Modjo. It can thus surface the next steps discussed in previous exchanges, while also providing information on the account's performance.
The workflow currently runs on the Sales Ops machine:
BigQuery + Modjo → Claude skill → meeting brief → email to the salesperson
The benefit is fairly simple: bring together in a single brief what the partner said and what the data shows about their account.
A follow-up draft ready in ten minutes
The same principle is used after meetings.
Based on the call transcript, a skill automatically prepares a draft of the follow-up email.
“Within ten minutes of your meeting ending, your email is ready.”
Ankorstore has nonetheless chosen to keep human approval. No email is sent automatically. The salesperson rereads and adapts the draft before sending it. Gabriel explains that salespeople have two time slots a day set aside for these reviews.
Gabriel himself wrote some of the first skills. As certain uses were industrialized, Sales Ops took over.
How do you automate sales coaching with Modjo?
Coaching is probably where the change in method is most visible.
Previously, Gabriel regularly held call reviews with his salespeople. He would pick a call, listen to it again, then work on it with the salesperson. According to him, the exercise could take about one hour per salesperson, with an obvious limit: the analysis was based on a single call.
Modjo agents now make it possible to work on all conversations.
“You review all the calls and you bring out real trends that are super interesting.”
Every Friday, Team Leads thus receive a brief for each salesperson. Gabriel describes in particular three positive points, three pieces of advice, the things that could have been handled differently, a recommendation and a scorecard with a grade for the week.
The main change is therefore not just the time saved. It is also the number of conversations analyzed: coaching can draw on an entire week of calls rather than on a single exchange picked occasionally, and bring out the areas for improvement that truly need to be prioritized.
Is MCP becoming a software selection criterion at Ankorstore?
This way of working is also starting to influence software choices.
Ankorstore uses 
Mixmax for its email sequences, for example. Gabriel welcomes the arrival of its MCP which, in his experience, helps design a sequence, create it and then enroll contacts in it before review and launch.
When choosing the new Sales CRM, Gabriel also cites the availability of an MCP among the arguments in favor of 
Attio. He attributes similar reasoning to the choice of 
Pylon on the CX side.
That said, Ankorstore has not yet made a final choice on the AI platform around which to organize these uses. The company has a contract with 
Dust and uses Claude Pro licenses in parallel. In the sales team, Claude is now the go-to tool, but Gabriel explains that Ankorstore still wants to experiment with different use cases and measure their ROI before deciding between the solutions.
What is emerging, then, is not really a stack reduced to a single piece of software. Specialized tools continue to fulfill their role: Aircall for telephony, Modjo for conversations, BigQuery for data and Sigma for BI.
What changes is the way they are accessed.
At Ankorstore, rationalizing the stack is no longer just about removing software. It is also about reducing the number of interfaces teams have to open and master every day, by gradually making Claude a gateway to part of the stack.
Summary of the software this article mentions in Ankorstore's stack
| Status | Workflow | Tool | Usage |
|---|---|---|---|
| In use | AI / interface | Claude | Skills, MCP, briefs, follow-ups, data access and Artifacts |
| In use | Conversation intelligence | Modjo | Call recording, transcripts, coaching and MCP |
| In use | Data warehouse |  Google BigQuery | Data source used in Claude workflows |
| In use | BI | Sigma | Dashboards and data visualization |
| In use | Telephony | Aircall | Calls flowing into Modjo |
| In use | Sales engagement | Mixmax | Sales sequences and MCP |
| In use | AI | Dust | Current contract and tests alongside Claude |
| Replaced | Meetings / transcription | FellowReplaced | Former tool of some Product Managers |
| In use | CRM | Attio | MCP cited among the arguments in favor of the new CRM |
| In use | Support / CX | Pylon | CX tool chosen on a rationale that includes access via MCP |
