Chief Executive Officer / India
The promise changed what we funded
- Company colleagues
- 170
- Six functional heads report to the CEO
- Annual revenue
- ₹54 crore
- FY2025–26; ₹6.8 crore operating profit
- Organisational customers
- 96
- Accessible document and workplace software in India
Decisions & consequences
01Rebuilding an export that preserved appearance but lost meaning
In 2023, customers using our document-authoring product reported that exported files looked correct but did not retain the structure needed for some assistive-technology workflows. Internal demonstrations had concentrated on the document inside our application. The gap appeared when users moved the output into another part of their work. Thirty-seven customers used the affected export route. Product colleagues could reproduce the problem, but the initial proposal was to provide instructions for manual repair while continuing the planned paid upgrade campaign.
I asked the team to show the full task, including the effort transferred to a user by manual repair. That made the commercial choice clearer. We could continue selling the upgrade with a prominent limitation, narrow the supported use, or rebuild the export before promoting the broader workflow. I recommended the rebuild and obtained board approval for ₹63 lakh over nine months. We deferred ₹1.1 crore of planned upgrade revenue rather than count it as achieved. Engineers owned the implementation; research colleagues designed task checks; commercial teams corrected the description given to customers. I owned the investment recommendation and the change to the revenue plan.
The revised export was checked with customer documents and the assistive-technology configurations included in the agreed test scope. We offered the correction to affected customers without charging for the repair itself. Some planned upgrades moved into a later period; they were not all recovered immediately. We also documented unsupported combinations instead of extending the result into a claim that every possible workflow had been solved.
The important change was in how we approved a product promise. We now ask where the output goes next and which parts of its meaning must survive that handover. A visually satisfactory file is not sufficient evidence of task completion. The case cost us time and near-term revenue, but it gave product, sales and directors a more precise account of what the software could support.
Paying for research conditions that changed the roadmap
Our 2024 research plan relied heavily on customers who could join a standard online session during office hours. The team had useful findings, but participation requirements favoured people already comfortable with our process. When researchers proposed widening the group, the first budget covered recruitment and interviews but little support for participants' different tools, schedules or communication preferences. I asked for a revised plan before treating a larger participant count as evidence of better coverage.
We approved ₹18 lakh for a programme involving eighteen participants with varied access needs and workplace contexts. It covered preparation, agreed assistance, suitable session formats and payment for time. Consent was separate from any publicity request, and payment did not depend on a favourable response. Researchers retained responsibility for the method. I asked them to report where findings differed and what the group could not establish, rather than produce a single score for accessibility across all users.
The programme changed our priorities. Several participants could complete a document task only by remembering a sequence that our training materials assumed would be obvious. The product team had planned to add another collaboration feature; we moved part of that capacity into clearer task navigation and recovery from an interrupted session. That delayed the collaboration feature by ten weeks. A follow-up with eight additional participants examined the revised workflow and exposed two further problems, which remained open in the release record until corrected. It was a focused check, not proof that every user would have the same experience.
My decision was to fund the conditions for credible participation and accept the consequence when the findings challenged the roadmap. The research team did not owe me an encouraging result in return. We now include participation support in the initial product-research budget and identify the decision the study is meant to inform. That makes its value easier to assess without reducing people's experience to a promotional statistic.
Declining a reseller promise that had no end to support
A reseller proposed a ₹4.2 crore agreement in 2025 for a broad workplace deployment. The headline amount was attractive relative to our annual business. During review, commercial and support colleagues found that the proposed terms combined a perpetual licence with support for an unspecified number of future versions and users. The partner described this as a simple procurement arrangement. For us, it created an obligation whose duration and demand were not sufficiently bounded to price responsibly.
I asked finance and the operating teams to model the service rather than apply a standard margin percentage to the licence price. Their estimates depended heavily on assumptions about future versions, deployment size and the partner's own support capability. Legal advisers helped distinguish the licence rights from service obligations. We could accept the original wording, propose a higher but still uncertain price, or offer a defined scope with a route to renew support. I chose the third option and told the board that we might lose the entire opportunity.
The eventual agreement was ₹1.7 crore. It specified the licensed deployment, supported versions and three years of support, with further service requiring a new agreement. Estimated support expenditure for that defined period was ₹52 lakh; that figure was not the total cost of delivering the deal or a claim of net profit. The partner reduced its initial deployment and accepted responsibility for agreed first-line support. Some prospective users were therefore outside the first phase, which was a real commercial sacrifice rather than an invisible success.
I do not present the difference between the two contract values as money saved. We gave up a larger potential sale to avoid selling an obligation we could not describe convincingly. The case led us to review how reseller proposals separate product rights, support duration and demand assumptions. A simpler contract is useful only if both parties can understand the work it actually commits them to perform.