Untitled

Colt and Microsoft optimize enterprise pricing

• Colt Technology Services and Microsoft have completed a proof of concept of an agentic AI engine to speed up and clarify pricing in complex deals.
• The tool reduces the time to generate and share pricing from days to about 10 minutes.
• The agent was trained in three days to cover most of Colt’s markets, with a reported accuracy of 99%.
• Quotes are generated with AI, but go through review by human teams before being sent to the customer.
• Colt plans to extend the use of agentic AI to more stages of the customer journey, and expects to make the pricing engine available later this year.
• The announcement refers to a proof of concept; general availability is positioned as a next step.

Speed and control in pricing
In global B2B pricing, the “what changed” is not just using AI, but compressing a flow that typically depends on country-by-country rules, commercial exceptions, and internal validations. In this proof of concept, Colt and Microsoft report two guiding metrics (time and accuracy) and a trust mechanism (human review) so that the leap in speed does not mean losing control.
As a public reference, the figures “10 minutes,” “3 days,” and “99%” are communicated in Colt’s announcement and in industry coverage that repeats it (e.g., Finextra/Telecompaper). As this is a pilot, final performance may vary when it is rolled out to more products, markets, or commercial conditions.

Development of the agentic AI engine by Colt and Microsoft

Colt Technology Services, a global provider of digital infrastructure, and Microsoft have jointly developed an agentic AI engine aimed at one of the most friction-filled points in the B2B telecommunications business: pricing complex deals across multiple geographies.

According to the company, the goal of the project is to “accelerate, simplify and clarify” pricing for enterprise customers, especially when international growth introduces market variables, commercial conditions, and service structures that often make the process opaque. The initiative was presented as a proof of concept already validated, and is part of a broader Colt program to apply agentic AI throughout the customer lifecycle.

Colt–Microsoft collaborative quoting flow
High-level workflow describing the Colt–Microsoft collaboration in this pricing use case:
1) Input of the

Agreement: customer requirements (sites/countries, capacity, timelines, SLAs) and commercial constraints.
2) Normalization: mapping to catalog/services and rules by market (currency, availability, local conditions).
3) Generation: the agent proposes an initial quote and its components (service lines, assumptions, and dependencies).
4) Checks: consistency validations (across countries, discounts, compatibilities) and exception detection.
5) Human review: specialized teams review and adjust before issuing.
6) Issuance and communication: delivery to the customer faster and, ideally, with more clarity about what is included.
Points where failures tend to appear in the real world: incomplete agreement data, exceptions not encoded in rules, or last-minute changes in availability/conditions by country.

Reduction of pricing time

In this context, “pricing” refers mainly to the generation of quotes and their communication to the customer in complex enterprise agreements.

The most striking data point from the pilot is the jump in speed: the engine reduces the time spent developing and sharing prices from days to 10 minutes. In an environment where global infrastructure projects can take weeks to obtain an accurate commercial proposal, this compression of the quoting cycle aims to eliminate a classic bottleneck between presales, engineering, and commercial teams. (The 10-minute figure is reported in public communications of the announcement and its industry coverage, such as Finextra/Telecompaper.)

The operational promise is twofold: on the one hand, to speed up customer decision-making; on the other, to allow Colt to respond more agilely in competitive processes where response time can tip the balance.

Accuracy in price generation

Colt claims that the agent delivers complex pricing with 99% accuracy (as reported by the company itself in the announcement). In practical terms, this aims to reduce inconsistencies between markets, minimize manual errors, and improve the comparability of proposals when a customer requests connectivity and services in multiple countries.

Even so, the workflow design maintains human control: specialized Colt teams review them before issuing. In practice, this positions AI as the initial generator and the teams as the final validation before sharing the quote with the customer. This “human-in-the-loop” approach aims to sustain output quality and strengthen customer trust in a process that, by its nature, has a direct impact on budgeting, planning, and cost control.

The real scope of the “99%”
What the “99%” contributes (and what it doesn’t) in a pricing case like this:
– Quality signal: it suggests that, within the scope of the pilot, most generated quotes match the expected result under the available rules and data.
– Operational control: the text already indicates that the quote is not sent “as is”; it goes through review by human teams before being issued, which reduces the risk of an error reaching the customer.
– Important limitation: the percentage depends on the scope (products/markets/rules) and on how “accuracy” is measured (for example, matching line items, amounts, discounts, or assumptions). That’s why it’s relevant that this is a proof of concept and that human control is maintained.

Training the AI agent

The project also stands out for its rapid rollout: Colt and Microsoft trained the agent in three days to deliver pricing for complex deals in most of Colt’s markets.

The collaboration combines Colt’s domain knowledge—pricing, commercial terms, and telecom-specific nuances—with Microsoft’s cloud AI capabilities. The result, according to both companies, is a system capable of automating highly complex tasks and scaling process consistency without relying exclusively on manual cycles.

Fast track to operational pricing
How a “trained in 3 days” milestone is usually achieved in a pricing agent (stage-by-stage view):
1) Data and rules (preparation): catalog, country-by-country rules, discount policies, technical dependencies, and permitted assumptions.
2) Configuration/training: tuning the agent to interpret requirements of the

agreement and apply rules consistently (often more “orchestration + rules + prompts” than training from scratch).
3) Validation: testing with real or historical cases, detecting exceptions, and adjusting rules/templates.
4) Controlled deployment: use in a pilot with mandatory human review and tracking of recurring errors.
This sequence helps explain why the time can be so short in a PoC: it speeds up when well-defined rules and catalogs already exist, and when deployment is limited and supervised.

Colt’s ‘people first’ AI strategy

The pricing engine is one of the first developments within Colt’s AI strategy called “people first”. The company describes it as an approach to create safe, scalable, and responsible AI ecosystems for employees and customers, with the goal that both can operate with confidence in a future defined by AI.

In practice, the pricing use case serves as a showcase: automation where it brings speed and consistency, with human oversight to ensure the final standard. Colt also notes that it is exploring applying agentic AI from pricing to onboarding and beyond.

Key trade-offs in enterprise pricing
Trade-offs that this “people first” approach tries to resolve in enterprise pricing:
– Speed vs control: compressing timelines to minutes requires maintaining a final validation point to avoid costly errors.
– Consistency vs commercial flexibility: standardizing rules improves comparability, but exceptions (discounts, bundles, local conditions) must be managed without breaking the model.
– Automation vs expert knowledge: the agent speeds up what’s repeatable; human teams are reserved for edge cases, negotiation, and verification.
– Scale vs trust: the more markets and products are covered, the more important the traceability of assumptions and review before issuing.

Impact on the customer experience

The customer experience is the narrative محور of the announcement. Colt argues that, as its customers grow and expand globally, pricing complexity and lack of transparency can become “confusing and costly.” Reducing quote time to minutes and improving process clarity aims to tackle two fronts: speed (fast responses) and understanding (less opacity in how the price is built).

In business terms, a shorter quoting cycle can translate into less friction in IT purchases, greater planning capacity, and a perception of a provider that is more “easy to do business with,” especially in multinational projects where the customer’s internal coordination already is

complex.

Visible improvements in customer experience
Concrete signs of CX improvement that a customer usually notices when pricing is accelerated and standardized:
– Initial response in minutes (not days) to move internal decisions forward.
– Fewer back-and-forths due to inconsistencies between countries or service lines.
– More comparable proposals across alternatives (same structure, same assumptions).
– Greater clarity on what the price includes (components, dependencies, conditions).
– Fewer surprises in later reviews thanks to human validation before issuing.

Views from leaders at Colt and Microsoft

Frank Miller, Colt’s chief AI and platforms officer, framed the problem as structural: enterprise IT purchasing “has always been complex,” and global projects can take weeks to obtain accurate pricing, which “can slow delivery.” His thesis is that AI can break that pattern: by accelerating and clarifying pricing, customers can focus sooner on their objectives, while Colt maintains connectivity.

Miller added that Colt will continue innovating with Microsoft to build an agentic application “factory” in telecommunications, focused on the best possible experience: “simple, accurate, secure and reliable.”

From Microsoft, Rick Lievano, Worldwide CTO, Telco, Media & Gaming, underscored the potential of agentic AI to transform complex enterprise workflows in telecom, where “speed, accuracy and scale” are critical. In his view, the combination of Colt’s industry expertise with Microsoft’s cloud AI capabilities helps automate complex processes, improve consistency, and give customers faster access to the information they need to decide.

Digital Transformation in Telecommunications: The Future of the Customer Experience

Colt and Microsoft’s proof of concept illustrates a broader trend in telecommunications: shifting automation from internal operational tasks to high-impact moments in the commercial relationship, such as quoting, onboarding, and service management. In a market where connectivity has become critical and global, differentiation shifts toward the simplicity and speed with which a provider enables buying, deploying, and operating.

Continuous Innovation in Quote Generation

Enterprise quoting has historically been a handcrafted process: multiple systems, country-by-country rules, commercial exceptions, and internal validations. The bet on an agent that condenses that work into minutes suggests a new standard of “time of

response” in pre-sales, especially for multinational deals.

If the pilot’s performance holds up in production, the impact will not be only internal efficiency: it could also redefine customer expectations about how long a provider should take to deliver a complete, comparable proposal.

The Role of Artificial Intelligence in Operational Efficiency

Beyond the pricing use case, Colt is already aiming to extend agentic AI to other stages of the customer journey. In telecom, where processes often cut across silos (sales, engineering, provisioning, support), goal-oriented automation—capable of executing chained tasks—promises to reduce idle time and improve consistency.

The key, as this deployment shows, will be balancing automation with expert control: using AI to speed up and standardize, without giving up human review in decisions that affect cost, feasibility, and service commitment.

Transformation of customer service in telecommunications with Suricata Cx

The need for an optimized customer experience

In telecommunications, the customer experience is defined by how quickly they get answers, the clarity of the information, and consistency across channels. The pressure to reduce timelines—from quotation to incident resolution—has made service optimization a competitive factor.

Specific solutions for industry problems

specific to industry challenges

Operators and infrastructure providers often face fragmented processes, multiple tools, and dependencies between teams. Specialized customer experience solutions aim to unify workflows, reduce rework, and improve traceability of requests, orders, and cases.

Use cases that make the difference

Among the most common use cases are: quote acceleration, onboarding automation, proactive order tracking, incident management with intelligent prioritization, and consistent customer communication at every service milestone.

Functional capabilities that boost efficiency

The capabilities that typically add the most value include: process orchestration, automation of repetitive tasks, real-time status dashboards, integration with existing systems, and analytics to identify bottlenecks and opportunities for improvement.

Why choose Suricata Cx for your business

Suricata Cx is positioned as an option for organizations looking to improve operational efficiency and customer experience through simpler workflows, end-to-end visibility, and tools that reduce response times in service operations.

The ideal customer profile for Suricata Cx

The typical fit is in telecommunications and B2B services companies with a high volume of requests, multiple teams involved, and a need to standardize processes without losing control, especially in multinational environments or with complex catalogs.

Strategic value and the future of customer experience

The industry’s direction points to “frictionless” experiences: less waiting, more self-service, greater transparency, and faster decisions. In that context, platforms and automation engines —including AI applied to critical processes— are consolidating as strategic levers to compete in a market where connectivity, by itself, is no longer enough.

Colt and Microsoft: AI Engine for Pricing in Enterprises shows how agentic AI can compress critical cycles from days to minutes without giving up human control. At Suricata Cx we share that same “human-in-the-loop” logic applied to telecom operations: automate what is repeatable to gain speed and consistency, and maintain expert oversight where the impact on customer and business demands maximum precision.

From Request to Resolution CX
Typical “request-to-resolution” journey in telecom CX operations (where a platform like Suricata Cx often adds value):
1) Intake

omnichannel: the customer opens a request (quote, activation, order, incident) and the minimum context is captured.
2) Classification and routing: it is assigned to a queue/team according to type, priority, SLA, and dependencies.
3) Orchestration: chained tasks across teams/systems (sales ↔ provisioning ↔ support) with visible statuses.
4) Checkpoints: data validation, milestone confirmation (order accepted, provisioning started, testing, closure).
5) Communication: consistent updates to the customer at each milestone and in case of blockers.
6) Closure and learning: root cause, time per stage, and feedback to reduce rework.

This analysis is approached from the perspective of an omnichannel CX platform for ISPs and telecom, where goal-oriented automation and human oversight are often the most reliable operational balance to scale without degrading the experience.

This article is based on publicly available information at the time of its publication about a proof of concept announced by Colt and Microsoft. The cited metrics (time, accuracy, and training timeline) reflect only the scope of the pilot communicated at that time. With the move to general availability, performance and implementation details may vary by market, product, and operating conditions.