Think Talent Services

AI Assessments: Six Design Dilemmas

Himanshu Nayak — Wed, 11 Nov 2026 10:04:49 +0000

Building AI Assessments?
The Hard Part Isn’t Just the AI.

Six design dilemmas we navigate as we build & refine AI-led assessments.

AI has made it possible to create assessment experiences that are more conversational, adaptive and scalable.

But building these experiences raises questions that go well beyond technology.

How do we ensure rigour when AI conversations are dynamic?
How much should an assessment adapt?
Where should human judgement come in?
And how do we maintain reliability as AI models evolve?

At Think Talent, these aren't just theoretical questions. They're decisions we navigate as we design, test and refine AI-led assessments.

Here are six dilemmas that continue to shape our thinking.

The Rigour Dilemma:
A Good Conversation Isn't Necessarily a Good Assessment

A realistic conversation doesn't automatically make a rigorous assessment.

We start by defining the effective and ineffective behaviours relevant to the assessment. We then use real-life human conversations to train the AI on how different behaviours may emerge and how it should respond within the scenario.

This helps the AI move beyond simply maintaining a conversation to creating meaningful opportunities for participants to demonstrate the behaviours being assessed.

Our Thinking

Rigour starts with behavioural design. The AI must be trained to recognise and respond to relevant behaviours, not just sustain a natural conversation.

The Standardisation Dilemma:
Consistency Without a Script

Standardisation supports fairness, but overly scripted interactions can compromise realism.

Our approach is to standardise the scenario, behaviour measurement opportunities and assessment criteria, while allowing conversations to adapt to individual responses within defined guardrails.

We rigorously test different participant response patterns to check whether the AI remains in character, respects conversational boundaries and provides comparable opportunities to demonstrate the target behaviours.

Our Thinking

Standardisation shouldn't mean identical conversations. It should mean comparable assessment conditions, with enough flexibility to reflect real-world interactions.

The Human Judgement Dilemma:
Where Should the AI's Role End?

Human involvement in AI-led assessments isn't an either-or decision.

Depending on the assessment's purpose and decision impact, we use different approaches: human-led evaluation, AI-assisted evaluation, risk-based human oversight or AI/logic-based insights.

For some assessments, trained assessors evaluate the behavioural evidence captured through AI interactions. In others, AI supports evaluation or generates structured insights, with oversight designed around the significance of the decision.

Our Thinking

There is no single model of human involvement. The objective is to determine where human judgement is necessary and where technology can reliably support the process.

The Software vs. Model Dilemma:
When Is an AI Assessment Finished?

Unlike conventional rule-based software, AI models can produce variable responses. Even small changes to models, prompts or configurations may influence how an assessment unfolds.

Our approach includes version control, regression testing and exception monitoring to identify unexpected behaviour, inconsistencies and deviations from the intended assessment experience.

This makes testing an ongoing responsibility rather than a one-time activity before launch.

Our Thinking

AI assessments cannot be treated as static software. Continuous testing and monitoring are essential to maintaining reliability as the underlying technology evolves.

The Build vs. Buy Dilemma:
What Do You Really Need to Own?

Powerful AI tools are readily available, but a conversational model alone doesn't make a scientifically grounded assessment.

Our approach combines existing AI capabilities with our own assessment design expertise and technology layers. This allows us to retain control over behavioural logic, interaction design, assessment workflows and reliability.

Rather than building everything from scratch, we focus on developing the layers where assessment expertise, customisation and control matter most.

Our Thinking

The choice isn't simply between building and buying. It's about using existing technology while retaining ownership of the assessment science and processes that make the solution credible.

The Data & Control Dilemma:
Trust Beyond the Conversation

AI assessments involve sensitive participant information, making privacy and data governance fundamental design requirements.

Our architecture separates personally identifiable information from AI interactions, ensuring it is never shared with AI. Defined access controls and audit trails support accountability, while strict data-purging practices ensure information is retained only as long as necessary.

These controls are embedded into the system from the outset rather than added as a separate compliance layer.

Our Thinking

Data privacy and governance are product design decisions. Protecting participant information requires deliberate choices about architecture, access, auditability and retention.

Beyond the Technology

Our recent pilot involving 112 AI-led role-plays with first-line managers provided encouraging early signals about participant engagement and the AI's ability to maintain consistent character behaviour.

But it also reinforced an important point.

A successful AI assessment depends on much more than how naturally the model responds.

It depends on the decisions made behind the scenes: how behaviours are defined, how consistency is maintained, how human judgement is incorporated, how reliability is monitored, and how participant information is protected.

As AI capabilities continue to evolve, these design decisions will become even more important.

Good AI assessments need more than good AI.
They need good assessment design.

Because building a conversational AI experience is one challenge. Building an assessment that organisations can trust is another.

The post AI Assessments: Six Design Dilemmas appeared first on Think Talent Services.

Seeing Behaviour at Scale: What 112 AI Role-Plays Reveal About First-Line Managers

vivek mohapatra — Mon, 26 Oct 2026 10:49:12 +0000

Seeing Behaviour at Scale:
What 112 AI Role-Plays Reveal About First-Line Managers

Building future-ready talent pipelines means seeing how leaders actually behave. This paper shows how AI-based role-plays make that visible, reliably and at scale.

Context: what we set out to test

Assessment centres rely on trained role-players and assessors to see how leaders behave in difficult conversations. That approach is reliable but hard to scale, and it varies with the person playing the role. This pilot asked whether an AI-based role-play could do the same job across a whole cohort of first-line managers.

We went in with three questions:

Question	What a "yes" would look like	Answer from the data
1. Can we observe key behaviours clearly?	Behaviours such as ownership, questioning and use of evidence show up as distinct, countable moments in the transcripts	Yes. Ownership split the cohort 46% / 54%, and questioning fell from 25% to 2% of messages
2. Does the AI perform well and consistently?	The character stays in role, replies at conversational speed and reacts to what each manager says	Yes. Zero persona breaks in 1,353 messages, 7-second median reply, and sentiment that tracks the manager's approach
3. Do participants engage effectively?	Managers stay long enough and write enough for their behaviour to be judged, without being made to	Yes. 96% held a real conversation, for a median of 23 minutes and 458 words

Executive summary

An AI character ran 112 unsupervised feedback conversations with first-line managers over 52 days without once breaking role, and the transcripts show clear, measurable differences in how managers handled them.

MeasureResult

Managers who held a real conversation (3+ exchanges): 96%
Median conversation length: 23 minutes
Median words written per manager: 458
AI persona breaks across 1,353 AI messages: 0
Median AI response time: 7 seconds
Managers who did not acknowledge their own failure: 54%

The last row is the point. An assessor panel sees a handful of managers a day. Here, a behaviour that matters (owning a mistake) was observed across an entire cohort, in the managers' own words.

The exercise

First-line managers at an IT services company each held one written conversation with an AI character: a team member disputing an "average" annual rating after receiving no critical feedback all year. The scenario is built around a failure on the manager's side, so it tests ownership, listening and resolution under mild pressure.

The deployment ran from 28 July to 17 September 2026. It was optional, unsupervised and available at any hour. The dataset holds 112 sessions and 2,706 messages, split evenly between manager and character, with zero structural defects (no missing, duplicated or out-of-order messages).

Reliable conditions for every participant

A behavioural measure is only as good as the conditions it is taken under. On every check that needs no human validation, the AI held those conditions steady.

Same character, every time. Zero breaks of persona in 1,353 AI messages, confirmed by a keyword screen and a hand-checked contradiction test. Every manager met the same person in the same emotional starting state: no rater fatigue, no order effects, no rapport bias.
Fast enough to feel like conversation. Median reply 7 seconds; 90% inside 12 seconds; only 14 of 1,353 replies over 30 seconds.
Available when managers were. Sessions ran on 43 of 52 days; one in eight ran at the weekend and 14% started between midnight and 6am.
Deep enough to produce evidence. Managers wrote 0.8 words for every word the character wrote, so the transcript captures their behaviour, not a script.

Four in five managers were still in the conversation at 10 minutes Share of all 112 participants still talking at each point; the decline is gradual, with no cliff

Session metrics, 112 sessions, Jul to Sep 2026. Cumulative: each point counts everyone at or beyond that threshold.

Only 4 of 112 sessions ended before three exchanges, and 29% ran past half an hour. People chose to continue rather than endured it.

What the transcripts revealed

The transcripts surfaced three observable behaviours, each measured across all 112 managers rather than a sampled few.

1. Ownership.

54% of managers did not acknowledge a failure of their own, in a scenario where the failure (feedback withheld all year) was objectively theirs.

The 46% who did used phrasing such as "I should have shared this earlier" or "I take responsibility for that". This measures accountability, not politeness: 83% of managers said thank you.

2. Listening gives way to solving.

Managers open by asking, then switch to defending and fixing. Concrete examples peak at 44% of messages in the second fifth, mostly after the rating had already been defended.

Questions fall from 25% of manager messages to 2% by the close Share of manager messages showing each behaviour, by stage of the conversation

Language-marker analysis of 1,353 manager messages across 112 sessions. Each session split into fifths so short and long conversations compare.

3. The character responds to what managers do.

Her negative sentiment rises early and falls to its lowest at the close, a shape four separate model measures agree on. Her sense of feeling unheard peaks at 19% of messages in the second fifth and drops to 4% by the end. In 53 sessions her sentiment shows a clear, lasting upturn at a median of exchange 7; in 15 it turns down, at a median of exchange 4. Conversations that go wrong, go wrong early.

Managers also reuse more of her vocabulary as the conversation goes on (10% overlap in the opening fifth, 20% in the fourth), a supporting signal of active listening.

Where her sentiment turned Sessions with a clear, lasting turn, placed at the median exchange; bubble size shows the number of sessions

15sessions turned down 53sessions turned up

12345678910

Exchange

Feeling unheard19% 4%of her messages, second fifth to the end

Vocabulary overlap10% 20%opening fifth to the fourth

What this means for assessment

AI role-play turns live interaction into evidence at cohort scale. Three properties make that evidence usable:

Standardisation.
Every participant faces an identical stimulus, which human role-players cannot guarantee across a cohort.
Observability.
Behaviours such as ownership, questioning and use of evidence appear as countable moments in a full transcript, traceable to the exact exchange.
Coaching precision.
Because the turning point of each conversation can be located, feedback can point to the specific exchange where it changed, and what the manager said there.

What this data does not yet show.

Behaviour markers here come from keyword and language-model analysis, not trained assessor ratings. Agreement with human assessors, inter-rater reliability and links to on-the-job performance have not yet been tested. Each participant took part once, so behaviour change cannot be measured.

AI in Assessments

See the innovation behind these 112 conversations.

See how Dialogue and AI Interaction work, and the rigour behind them.

Explore AI in Assessments

The post Seeing Behaviour at Scale: What 112 AI Role-Plays Reveal About First-Line Managers appeared first on Think Talent Services.

From Assessment to Action: Learning Needs Analysis at Scale

vivek mohapatra — Mon, 26 Oct 2026 09:08:58 +0000

From Assessment to Action:
Learning Needs Analysis at Scale

A light development center (DC) for 92 Store Managers at a large jewelry retailer enabled a more targeted brief for an already planned learning intervention.

8 indicators · scale shown from 1 (centre) to 3 of 5

Why the competency profile comes first

Effective learning starts with an accurate picture of current capability. Most learning needs analysis (LNA) for large cohorts relies on nominations, self-ratings or performance scores. These reflect visibility, confidence or outcomes rather than the behaviors behind them, so programs end up generic.

A competency profile built from structured assessment adds three things:

Precision:

each competency is broken into observable behaviors, so a workshop can target the exact point where capability drops off.
Prioritization:

group data separates cohort-wide gaps (core program) from gaps in a subset (targeted support).
Design:

comparing what people know with what they do shows whether the need is knowledge, practice or on-the-job reinforcement.

The engagement

A leading jewellery retailer asked us to assess 92 store managers across four zones, ahead of planned workshops on its four role-specific competencies: Business Acumen, Customer Centricity, Operational Excellence and People Leadership. Each competency had two indicators.

Participants completed three exercises in two hours: a situational judgement test (what they know), a one-to-one roleplay with a team member (what they do), and a scenario-based video presentation (how they reason about the business). Assessors scored 24 behaviors on a 1–5 scale, with 2 or below treated as a development area. The methods measured genuinely different things: roleplay and presentation scores barely correlated (r = 0.06).

What the data revealed

The cohort averaged 2.39 out of 5. The value lay in where scores fell.

Indicator	Mean	% scoring 2 or below	Judgement (SJT)	Roleplay
Process discipline & compliance	2.12	67%	2.74	1.64
Entrepreneurial Mindset	2.16	64%	—	1.89
Data & Digital Savvy	2.27	50%	—	1.97
Ownership & Accountability	2.46	36%	2.89	2.70
Consultative Selling	2.47	32%	2.62	2.87
Delegation	2.53	38%	2.48	2.46
Feedback & Coaching	2.54	40%	3.26	2.41
Service Excellence	2.57	27%	2.85	2.87

What the data revealed

Knowing outruns doing.

Judgement scores were often well above live behavior, by over a point on Process discipline (2.74 vs 1.64).

Judgement (SJT) vs Roleplay

Process discipline & compliance 2.741.64

Ownership & Accountability 2.892.70

Consultative Selling 2.622.87

Delegation 2.482.46

Feedback & Coaching 3.262.41

Service Excellence 2.852.87

Follow-through and business framing are the sharpest gaps. Setting a review mechanism for agreed commitments scored 1.64; explaining why a new opportunity matters to the store scored 1.74.

Managers take over rather than delegate. Only 3% chose to brief a newer team member, let her lead a high-value sale and review afterwards; 52% chose to handle it themselves.

Need is uneven. 20% of managers had no development area at competency level, while 14% had one on all four. The South zone averaged 2.20 against 2.56 in the North.

From insight to learning needs analysis

Workshop	Recommended focus
Business Acumen	Framing the store business case; turning numbers into a short pitch to the team
Operational Excellence	Less rule explanation; drills on setting review points, owners and dates
People Leadership	Briefing, stepping back and reviewing, so delegation becomes development
Customer Centricity	Shorter; questioning needs before recommending a piece

The Program design was modified further basis the DC inputs. Workshop design was adapted to be more practice-heavy than concept-heavy with more on-the-job scenarios built in. Follow-up assignments were created post each workshop to ensure sufficient skill practice for participants. Content was simplified with an “advanced” cohort planned for high-performing participants.

Lessons for large cohorts

Measure behaviors, not just competencies: the data showed the issue was follow-through, not rule knowledge.
Use more than one method: judgement tests alone would have overstated capability.
Look at the spread: an average of 2.39 hid both ready-made peer coaches and managers needing intensive support.

The result is a learning plan built on evidence, targeted to real gaps, and measurable when the same behaviours are assessed again.

The post From Assessment to Action: Learning Needs Analysis at Scale appeared first on Think Talent Services.

What Really Moves 85k+ Employees to Respond

vivek mohapatra — Thu, 22 Oct 2026 11:44:11 +0000

What Really Moves 85k+ Employees to Respond

Lessons from surveys run for 85k+ employees across 10+ organizations, in 12 languages, over 3 months.

Most organizations judge an employee survey by its response rate. Across these 10+ survey-based engagements, we covered 85k+ employees, from corporate teams to on-the-ground workforces with 91% overall completion.

But participation is not the goal in itself. It is what makes the data credible: results leaders can act on because they reflect the whole organisation, not just the people who answer every email. So we looked at the behaviour behind the number. What moved people to respond, when they did it, which language they chose, and where they hesitated. Four lessons stood out.

91% of invited employees completed the survey Employees at each stage, with share of everyone invited

Invited 85,800 · 100.0%

Started the survey 79,322 · 92.4%

Completed all statements 78,250 · 91.2%

Of those who started, 98.6% went on to finish. Only 1,072 dropped out along the way.

Reminders build awareness. People drive action.

All the surveys we administered stayed open for 3-5 weeks, and the follow-up came in layers.

One-click access. Every employee got a personal single sign-on link by email. There were no IDs or passwords to find, copy-paste or forget.
Two voices at launch. Each client's internal communications explained why the survey mattered. A separate email from us, as an independent partner, made the promise of anonymity credible.
Paced reminders. Every other day for the first 2 weeks, then twice a day near the close.
People on the ground. In the last 2 weeks, HR business partners followed up with their business units, one by one.

A client example Starts jumped 5× the day HRBP follow-ups began

Reminders only · 10,786 starts (19%) From 6 Oct: + HRBP follow-ups · 47,144 starts (81%)

1,863

10,187

Dips fall on Sundays and the 2 October holiday. Excludes the pilot group, who responded in mid-September.

The reminders worked: Each reminder day stands out against the quiet days in between. But on their own, they brought in less than a fifth of all responses. The real shift came when HR business partners stepped in. Daily starts went from under 2,000 to over 10,000 overnight, and 4 in 5 responses came after that.

What this means: Your HRBPs and business leaders are the survey's real distribution network. Brief them before launch, give them a clear role, and build their follow-up into the plan from week one, not as a last-week rescue.

Timing shapes the response

People answer surveys in the same hours they do their work. 2 in 3 started between 11 am and 5 pm, and the busiest hour was 12 to 1 pm.

2 in 3 started between 11 am and 5 pm Employees starting the survey, by hour of the day · 11 am–5 pm: 66% of all starts

Reminder send times

9 am 11 am 3 pm

Early on, large sends sat in the email queue for a few hours before they reached inboxes.

The send time mattered. Reminders went out at 9 or 11 am, with a second at 3 pm on two-reminder days. Morning sends kept starts high through lunch. The 3 pm sends brought a quick rise that faded once the working day ended.

What this means for reminders:

Send early, around 9 am, so emails land before the late-morning peak. For very large groups, send in batches so nothing waits in a server queue.
Use an afternoon send as a short nudge, not the main ask.
Follow the organisation's working week: include working Saturdays.

Language decides whose voice is counted

In one client example, the survey was offered in English and 11 Indian languages. Nearly 15,000 employees, 1 in 4, chose a language other than English. Without those versions, a quarter of the workforce would have answered in a second language, or not at all.

A client example 1 in 4 employees chose a language other than English Respondents by survey language

English · 44,415 (75%)11 other languages · 14,794 (25%)

The 11 other languages

Hindi8,672

Marathi1,975

Tamil857

Telugu845

Gujarati778

Kannada547

Malayalam362

Bengali333

Punjabi196

Odia153

Assamese76

They finished just as often. Respondents who chose a regional language completed at 98.7%, against 98.2% for English. Every Malayalam respondent who started went on to finish.

They did take longer, though. Most regional-language respondents spent 1 to 3 minutes more on the same 33 statements.

A client example Most regional languages took longer to complete than English Average minutes to complete, by language · 8 of 11 languages took longer than English

English 12.7 min

Assamese15.8 min (+3.1)

Hindi14.9 min (+2.2)

Bengali14.8 min (+2.2)

Malayalam14.4 min (+1.7)

Marathi14.1 min (+1.4)

Tamil14.0 min (+1.3)

Punjabi13.5 min (+0.8)

Telugu13.1 min (+0.4)

Odia12.7 min (+0.0)

Kannada12.2 min (−0.5)

Gujarati12.0 min (−0.7)

11 min 12 min 13 min 14 min 15 min 16 min

What this means: Language is not a translation task. It decides whose experience shows up in the results. Each version here was adapted for its audience, not translated word for word, and reviewed by native speakers from both the client's team and ours. A work scenario that sounds natural in English can read stiffly, or mean something different, in Tamil or Assamese. And when you plan the survey window and the time you ask people for, plan for the slowest language, not the fastest. (We wrote more on this in Beyond English: Designing Talent Interventions for Bharat.)

Remove every bit of friction

At this scale, small frictions add up to thousands of missing voices. This is where the survey platform earns its place.

Getting in. In one client example, of the 59,209 employees who opened the survey, 58,210 answered all statements, a completion rate of 98.3%. Most of the other 999 stopped right at the start.

A client example 2 in 3 drop-outs happened before page 2 Where the 999 people who didn't finish stopped, by survey screen

667 people · 67% of drop-outs

322

345

104

LanguageLang. Page 1 Page 2 Page 3 Page 4 Page 5 Page 6 Page 7 Page 8 Page 9

Page 9 includes 13 people who answered every statement but never pressed Submit.

A third of all drop-outs (322 people) chose a language and left before answering a single statement. Another 345 left on the first page. After page 2, drop-off falls to a handful of people per screen. The first minute decides participation: open with the most familiar situations, keep the instructions short, and tell people upfront how long the survey will take.

On any device. Most people at this firm work at a desk, and 88% answered on a laptop or desktop. But almost 7,000 employees, 1 in 9, took the survey on a phone, mostly Android.

A client example 1 in 9 employees answered on a phone Laptop or desktop: 88.2% · Phone: 6,935 people, 11.7%

Chrome (55%) and Edge (44%) covered almost everyone, and Safari was used by fewer than 1 in 500. Test on the browsers your people actually use, and treat the phone as a must-have: a scenario with four answer options has to read comfortably on a small screen, without zooming or sideways scrolling.

Finishing. About 1 in 14 employees paused and came back on another day to finish, so saved progress matters. And 13 people answered every statement but never pressed Submit. A clear final button, and a nudge to anyone who stalls on the last page, closes that gap.

What this means: None of this is glamorous, but each friction point costs responses. Look for a survey platform that offers single sign-on, saves progress, works on any phone and browser, and shows you exactly where people stop.

5 questions to ask before your next survey

Who will follow up on the ground, and from which day?
When do your people actually work, and when will your reminders land?
Can every employee answer in the language they think in?
What stands between an employee and the first question?
Is every question worth the time it asks for?

Beyond the response rate

A response rate tells you that people showed up. Behaviour data tells you how they showed up: what moved them, when they had time, which language they trusted, and where they hesitated. Read together, these signals make the next survey shorter, simpler and easier to say yes to, and make the results more reliable.

If you are planning an employee survey at scale, across languages, locations and roles, we would be glad to share what we have learnt.

Talk to a Consultant

Your talent transformation starts with a conversation.

The post What Really Moves 85k+ Employees to Respond appeared first on Think Talent Services.

Demo

Himanshu Nayak — Thu, 03 Sep 2026 10:40:58 +0000

From Framework to Field:
Stories from the Ground

Name	Age	City
Rahul Sharma	25	BBSR
Priya Das	28	Katak
Ravi	31	Puri

The post Demo appeared first on Think Talent Services.

From Framework to Field: Stories from the Ground

Himanshu Nayak — Fri, 03 Apr 2026 11:46:37 +0000

From Framework to Field:
Stories from the Ground

Across organizations, the real test is not in frameworks, but in how work gets done on the ground. This piece brings together a set of client stories that reflect how we are working with clients to redesign talent interventions to stay close to business realities.

From large-scale sales assessments to plant-level potential mapping, the focus is consistent: making talent processes more industry-, function-, and role-specific, more integrated, and more practical for leaders and teams. These examples show how organizations are moving beyond generic models to solutions that reflect their operating context and drive execution outcomes.

Designing Competency Framework for a FMCG firm

Mid-size FMCG firm looking to build a structured competency framework for the sales function across channels (GT, MT, Ecom, etc.) and customer marketing, to enable talent development and career management.

Integrated
Functional Framework

Single framework with 14 functional competencies across all roles

Behavioural competencies expected to be covered separately

Client Feedback:
Very comprehensive but competency indicators still feel generic & not sharp enough. How do we bring out the unique context of each channel & role more clearly?

Channel-wise Functional Frameworks

8 separate frameworks with differentiated competency sets for each channel/group (GT, MT, B2B, Ecom, CSD, Customer Marketing, Commercial Excellence)

Focused solely on Functional competencies; indicators contextualized for each channel

Client Feedback:
Much sharper, but feels incomplete – need to integrate behavioural elements.

Channel-wise Integrated Frameworks

8 channel-specific frameworks combining functional and behavioural competencies

Streamlined to 6 competencies per group

Client Feedback:
Rolled out to sales leadership teams for further fine-tuning of language. Final framework rolled-out across the entire sales function.

Competency Profiling & Gap Analysis using ADCs

ADCs have been long considered one of the most scientific methodologies for competency profiling. The following cases show how scale, agility, and integrated frameworks are shaping how ADCs are designed and delivered.

Competency profiling for the entire on-ground Sales team at an Auto major

Differentiated frameworks designed for Sales & Service teams and for each Business Unit, in consultation with the business stakeholders

Combination of behavioural & functional competencies (eg: Influencing, Sales planning)

Assessment deployed for 1500+ participants in IC & First-level Manager roles in sales across all verticals

Skill profiling for India Sales team at an FMCG major

Skill profiling process for 450+ participants in IC & ASM roles

Skills identified from global skills library & contextualized to India & division-wise operating context through market visits & other interactions

8 differentiated exercise sets across divisions & levels

Assessment conducted for all participants virtually & completed within a 3-week window

Potential Mapping for manufacturing & supply chain function at a leading Tyres manufacturer

Org-level competency framework already existed, and senior manufacturing leaders expected it to be translated to the plant-level context

Framework, exercises, and delivery aligned to plant operating rhythm

In-person DCs covered 200+ participants, across multiple plants; workshops & 1-1 sessions conducted to create IDPs & group data leveraged to identify hi-potential employees

Designing and Delivery of Sales Excellence Learning Program

Leading DTH provider building a customized program for 400+ ASMs across India to drive future readiness in a market shaped by OTT growth & changing customer segments.

Grounded,
Context-led Content

Ground realities faced by sales & service ASMs captured through FGDs & market visits across urban & semi-urban markets

Content built around industry-specific challenges, linking daily operational issues to broader business outcomes

Simulations & reflection-based activities focused on consultative selling, customer-first decisions, balancing acquisition targets with service quality, & identifying opportunities for differentiated client activations at the market level

Facilitator Panel with a Sales Background

Senior facilitators with direct experience in channel-driven sales environments

Practical experience across partner management, activations, and frontline challenges

Outside-in perspective combining operational depth with industry benchmarks; examples closely aligned to participants’ daily realities

Experiential
Workshop Delivery

Simulation-based workshops anchored in real scenarios

Case-based learning linked to actual workplace challenges

Hands-on skill practice with guided self-reflection

Structured action planning to drive on-the-job execution

The post From Framework to Field: Stories from the Ground appeared first on Think Talent Services.

The Shifting Ideal: Workplace Personality Traits in the Age of AI

Himanshu Nayak — Thu, 26 Mar 2026 04:58:32 +0000

The Shifting Ideal:

Workplace Personality Traits in the Age of AI

For decades, organizations have relied on psychometric inventories to define, measure, and select for an “ideal” workplace personality. Traits like conscientiousness, extraversion, and emotional stability have been repeatedly linked to job performance, leadership emergence, and career success.

But the rise of artificial intelligence (AI)—especially systems that automate analytical, repetitive, and even creative tasks—is beginning to reshape not just what work gets done, but how humans add value at work.

This raises a critical question for employers and talent leaders:
as AI changes the nature of work, will the personality traits we value most also change?

What AI Changes: From Execution to Judgment

AI adoption does not eliminate roles overnight. Instead, it reconfigures task bundles within roles. Tasks that are repetitive, rule-based, data-heavy, and optimisation-focused are increasingly automated or augmented by AI systems.

As a result, human effort shifts toward areas where judgment, interpretation, and coordination matter more than execution alone.

This redistribution of work has important implications for personality–performance relationships. Traits that were valuable because they supported execution and consistency may decline in relative importance, while traits linked to adaptability, learning, and collaboration may rise.

How Personality Value
Is Shifting

Beyond the Big Five: Expanding the Lens

Traditional personality frameworks may underweight traits that are becoming more relevant in AImediated work. These include learning orientation, metacognitive awareness, ethical sensitivity, and systems thinking.

In other words, understanding personality at work may need to move beyond static trait definitions toward a more dynamic view of how individuals learn, adapt, and make decisions in evolving environments.

Decoding Behavioural Readiness for AI

We reviewed personality data of managers from 15 BSE100 organisations to review broader readiness for driving the AI agenda.

Interpretation

Today, managers are well-equipped to drive outcomes when it is well-def ined. But only few are inclined to push beyond norms and fully leverage a new technology to lead transformation. Identifying and developing these leaders as AI champions becomes a critical priority for organizations. As the AI agenda trickles downward, organizations also need to equip managers with skills to lead their teams navigate through the change cycle.

With targeted behavioural interventions to bridge these gaps, India inc. can move towards a more sustained and scalable AI-led transformation.

Closing Note

AI ambition in Indian organisations is supported by capable and driven managers, but as the nature of work evolves, so must the behavioural pro f ile required to lead it.

Instruments like the Think Talent Personality Inventory (TTPI) help organisations decode these behavioural patterns, enabling them to build leadership capability aligned to the demands of an AI Age

The post The Shifting Ideal: Workplace Personality Traits in the Age of AI appeared first on Think Talent Services.

The New Thinking Behind Talent Management

Himanshu Nayak — Wed, 18 Mar 2026 10:13:58 +0000

The New Thinking Behind

Talent Management

Organisations today must balance stability with constant change in a rapidly evolving world. The shift in work, accelerated by the pandemic, has redefined how individuals engage with careers and organisations.

Leadership is moving beyond hierarchy, with learnability and skills becoming the true drivers of influence. In this new era, human capabilities and adaptability will define long-term organisational success.

Stability vs Change: The Core Tension

“Organisations are built to create stability in an unstable world.” And “Hierarchy will melt, skills will determine leadership.”

These two quotes capture one of the deepest tensions in talent management today.

At first glance, these observations appear simple. Yet they reflect a fundamental reality. Organisations rely on systems, structures, and processes to create continuity and predictability. At the same time, markets, technologies, and customer expectations are changing at an unprecedented pace.

For talent leaders and HR professionals, the challenge is neither to preserve stability alone nor to drive constant disruption. The real question is how to balance this tension, this paradox, between stability and change.

The Changing Relationship Between People and Workt

The pandemic reshaped not only how work is organised but also how individuals relate to organisations.

The past fifteen years have brought profound shifts in how organisations think about talent, leadership, and capability. One of the most significant turning points came during the COVID-19 pandemic. What initially appeared to be a temporary disruption proved far more consequential, accelerating changes already underway.

Remote work, flexibility, and deeper reflection on personal priorities forced both employees and companies to rethink long-standing assumptions. The world of talent has become more self-focused, not in a selfish sense, but as a form of healthy, enlightened self-interest. Individuals are more deliberate about their careers, the meaning they derive from work, and the environments in which they choose to contribute.

This growing agency is reshaping attraction, development, and retention. The employment relationship is increasingly becoming a dialogue rather than a one-sided expectation of loyalty.

The Future of Leadership and Human Capability

Today, it is about learnability.

At the same time, the foundations of credibility and leadership are shifting. For decades, experience and tenure signalled authority and expertise. That equation is changing. The value of tenure or experience has declined. Today, it is about learnability.

In a world where knowledge cycles are shrinking and industries are evolving rapidly, the ability to learn, adapt, and reinvent oneself is becoming more valuable than accumulated experience alone. Leadership, therefore, may increasingly follow capability rather than title.

Technology is also reshaping the nature of work. As automation and artificial intelligence take over routine tasks, the differentiator for organisations will lie in what humans uniquely contribute, the ability to think differently and build capabilities that machines cannot replicate.

Capabilities such as creativity, empathy, judgement, collaboration, and complex problem-solving will increasingly define organisational advantage. Coaching, too, is becoming a core leadership capability. It has to be embedded in the way of managing and leading.

Looking ahead, even coaching will evolve. The coach of the future will be an augmented coach, partnering with AI while bringing emotions, compassion, and human connection.

Taken together, these shifts suggest that talent management is not just evolving in practice but is being reconsidered at a more fundamental level, with new assumptions about work, capability, and the role of human potential in organisations.

The post The New Thinking Behind Talent Management appeared first on Think Talent Services.

The Quiet Reinvention of Talent Management

Himanshu Nayak — Thu, 05 Mar 2026 06:53:04 +0000

The Quiet Reinvention of

Talent Management

The Snapshots team asked Mr. Satish Pradhan, Former Group HR Head, Tata Sons and a thought leader in HR, for his views on Talent Management. The questions were focussed on where the practice of Talent Management is headed.

The central question for leadership is no longer ‘Who are our best people?’ but ‘How effectively do people move, grow, and take on increasing complexity?

Satish Pradhan

HR Advisor & Former Chief Group HR

Tata Sons

From Inventory to Flow:

The End of Talent Stockpiling

For decades, talent management was treated as a matter of industrial stockpiling: identify “high potentials” early and insulate them within carefully designed programs.

However, this model has proven fragile. Static labels age poorly in a world defined by rapid role change and rising complexity. We are finally recognizing that talent is not inventory—it is a flow system. The central question for leadership has shifted from “Who are our best people?” to “How effectively do people move, grow, and take on increasing complexity over time?”.

Beyond Competencies:

The Rise of Situational Capability

The exhaustive competency frameworks of the past— which described behaviours in reassuring detail—often failed to predict actual performance under pressure. Today, we view capabilities as composite, situational, and perishable. It is no longer enough to display a behaviour in isolation; what matters is the real-time integration of judgment, learning velocity, and systems thinking.

As AI and people analytics expose the weak correlation between traditional models and actual effectiveness, the language of competencies is losing its former authority.

The Real Crucible:

Development Through Consequence

Leadership development is moving away from episodic, event-based interventions. While residential programs and inspirational speakers haven’t disappeared, we now better understand their limits. True development is a function of role design, decision rights, and exposure to real consequences.

Assignment planning, stretch assignments, crisis roles, and enterprise-wide projects have regained legitimacy as the primary drivers of growth. Ultimately, development is constrained less by an individual’s motivation than by the structure of the work they are actually asked to do.

Talent as Enterprise Risk

Talent management has moved out of the “safe custody” of HR and into contested organisational space.

Driven by succession failures and technological disruption, CEOs and Boards now view talent as a form of enterprise risk, alongside financial and cyber threats. There is a new, healthy pressure to demonstrate how talent architecture genuinely supports strategic direction rather than merely reflecting current hierarchies.

Recalibrating Potential and the Necessity of Unlearning

The idea of “potential” needs a sober recalibration; lately, it has become more of a compliment than a commitment. Properly understood, potential is an obligation to take on future accountability at higher levels of ambiguity.

Labelling people as “HiPo” without offering them the authority to make consequential decisions only breeds entitlement and disengagement. Furthermore, we must reward “unlearning”.

The inability to let go of once-successful practices is now a material risk; leaders must be judged on their ability to retire old skills and question legacy wisdom.

The Future: Fractured Careers and Cultural Infrastructure

Looking ahead, talent systems must accommodate “zig-zag” trajectories— trajectories that include pauses, re-entries, and multiple reinventions over longer working lives. Success will be evaluated through the health of talent portfolios rather than individual “bets”.

Finally, culture will be redefined. It will be assessed not through lyrical value statements, but through concrete signals: decision latency, error recovery, and the psychological cost of speaking up. Increasingly, the true measure of a leader will be how safe it is to think in their presence.

The post The Quiet Reinvention of Talent Management appeared first on Think Talent Services.

Leadership in the Age of AI

Himanshu Nayak — Tue, 03 Mar 2026 09:42:24 +0000

Leadership in the

Age of AI

Insights from conversations with 50+ CXOs

Artificial Intelligence is no longer a distant possibility or a specialist topic—it is rapidly reshaping how organizations think, decide, and compete. Algorithms now outperform humans at pattern recognition, predictive analysis, and optimization at scale. Yet when we spoke to over 50 CXOs across industries, one message came through clearly: the most critical leadership capabilities in the age of AI are deeply human. While the conversations ranged widely-from data maturity to organization and talent transformation-four broad themes consistently stood out. Together, they form a blueprint for leadership in an AI-enabled world: Judgment, Balance, Resilience, and Trust.

01. Judgement

Making Sense When the Signal Is Loud

In an age of AI, leaders are surrounded by more data, more dashboards, and more recommendations than ever before. Paradoxically, this has made judgment—not certainty—more valuable.

CXOs spoke about the need to sense patterns rather than chase precision. AI can surface correlations and probabilities, but it cannot fully grasp context, ethics, or long-term consequence. Leaders are increasingly required to decide under conditions where answers are incomplete, timelines are compressed, and outcomes matter deeply.

Judgment in this environment is not about having all the facts; it is about asking the right questions:

What is the decision we are really making?

Where should we trust the model—and where should we challenge it?

What are the second- and third-order implications?

How do we create moats for our businesses beyond technology and AI?

Several leaders emphasized that thoughtful decision making under pressure—especially when AI outputs conflict with intuition and real world meaning—is becoming a defining leadership skill..

02. Balance

Managing Opposing Forces

Almost every CXO described leadership today as a constant act of balance.

On one side is speed—AI enables rapid experimentation, automation, and execution. On the other is deliberation—ensuring that decisions are ethical, inclusive, and sustainable. Leaders must balance technology adoption with human impact, efficiency with empathy, and short-term performance with long-term capability building and sustainability.This balance is not static.

It shifts across moments:

Centralization vs. empowerment

Automation vs. augmentation

Innovation vs. operational stability

Core value creation engines vs. new ones

The leaders who seem most effective are those who are comfortable holding opposing pulls at once— without rushing to oversimplify. They see balance not as compromise, but as conscious calibration

03. Resilience

Bouncing Back, Again and Again

If AI is accelerating change, it is also increasing volatility.

Strategies evolve faster. Skills become obsolete sooner. Mistakes are more visible—and sometimes more costly.

In this context, resilience emerged as a critical mindset rather than a one-time response to crisis. CXOs spoke about the ability to recover quickly from setbacks, unlearn outdated assumptions, and keep moving without becoming defensive or cynical.

Importantly, resilience was not framed as individual grit alone. Leaders highlighted the need to model vulnerability-acknowledging uncertainty, admitting when something has not worked, and signaling that it is safe for others to do the same.

The leader’s personal ability to experiment and fail, and learn and be public about the failures, can be a key strength. In organizations learning to work alongside AI, failure is inevitable. Resilient leaders normalize learning loops rather than blame cycles.

04. Trust

The Leadership Currency of the AI Era

Perhaps the most repeated theme was trust.

As AI systems influence hiring, performance evaluation, and strategic decisions, trust becomes both more fragile and more essential. Leaders must navigate multiple layers of vulnerability: their own dependence on technology, employees’ fears of replacement, and stakeholders’ concerns around fairness and transparency.

Trust, the CXOs noted, is built through:

Clear communication about how and why AI is being used

Honest conversations about limitations and risks

Willingness to share decision logic rather than hide behind systems

Leaders who communicate openly, listen deeply, and treat uncertainty with respect are better positioned to carry their organizations through transition. In the age of AI, trust is not a “soft” skill-it is a strategic one.

The Human Advantage

Across all four themes, a common insight emerged:

AI will continue to evolve rapidly, but leadership success will hinge on human strengths that technology cannot replicate.

Judgment over reaction.

Balance over extremism.

Resilience over rigidity.

Trust over control.

The age of AI does not diminish the role of leaders—it raises the bar. The leaders who will succeed are those who combine technological fluency with emotional intelligence, clarity of thought with humility, and speed with wisdom.

In the end, the future may be powered by machines but it will still be shaped by people.

The post Leadership in the Age of AI appeared first on Think Talent Services.

Think Talent Services

AI Assessments: Six Design Dilemmas

Building AI Assessments? The Hard Part Isn’t Just the AI.

AI has made it possible to create assessment experiences that are more conversational, adaptive and scalable.

The Rigour Dilemma:A Good Conversation Isn't Necessarily a Good Assessment

The Standardisation Dilemma:Consistency Without a Script

The Human Judgement Dilemma:Where Should the AI's Role End?

The Software vs. Model Dilemma:When Is an AI Assessment Finished?

The Build vs. Buy Dilemma:What Do You Really Need to Own?

The Data & Control Dilemma:Trust Beyond the Conversation

Beyond the Technology

Seeing Behaviour at Scale: What 112 AI Role-Plays Reveal About First-Line Managers

Seeing Behaviour at Scale: What 112 AI Role-Plays Reveal About First-Line Managers

Context: what we set out to test

Executive summary

The exercise

Reliable conditions for every participant

What the transcripts revealed

1. Ownership.

2. Listening gives way to solving.

3. The character responds to what managers do.

What this means for assessment

See the innovation behind these 112 conversations.

From Assessment to Action: Learning Needs Analysis at Scale

From Assessment to Action: Learning Needs Analysis at Scale

Why the competency profile comes first

The engagement

What the data revealed

Knowing outruns doing.

From insight to learning needs analysis

Lessons for large cohorts

What Really Moves 85k+ Employees to Respond

What Really Moves 85k+ Employees to Respond

Reminders build awareness. People drive action.

Timing shapes the response

Language decides whose voice is counted

Remove every bit of friction

5 questions to ask before your next survey

Beyond the response rate

Demo

From Framework to Field: Stories from the Ground

From Framework to Field: Stories from the Ground

From Framework to Field: Stories from the Ground

Designing Competency Framework ​ for a FMCG firm

Competency Profiling & Gap Analysis ​ using ADCs

Designing and Delivery of Sales ​Excellence Learning Program

The Shifting Ideal: Workplace Personality Traits in the Age of AI

The Shifting Ideal:

Workplace Personality Traits in the Age of AI

What AI Changes: From Execution to Judgment

How Personality Value Is Shifting

Beyond the Big Five: Expanding the Lens

Decoding Behavioural Readiness for AI

Interpretation

Closing Note

The New Thinking Behind Talent Management

The New Thinking Behind

Talent Management

Stability vs Change: The Core Tension

The Changing Relationship Between People and Workt

The Future of Leadership and Human Capability

The Quiet Reinvention of Talent Management

The Quiet Reinvention of

Talent Management

Satish Pradhan

HR Advisor & Former Chief Group HR

Tata Sons

From Inventory to Flow:

The End of Talent Stockpiling

Beyond Competencies:

The Rise of Situational Capability

The Real Crucible:

Development Through Consequence

Talent as Enterprise Risk

Recalibrating Potential and the Necessity of Unlearning

The Future: Fractured Careers and Cultural Infrastructure

Leadership in the Age of AI

Leadership in the

Age of AI

01. Judgement

Building AI Assessments?
The Hard Part Isn’t Just the AI.

The Rigour Dilemma:
A Good Conversation Isn't Necessarily a Good Assessment

The Standardisation Dilemma:
Consistency Without a Script

The Human Judgement Dilemma:
Where Should the AI's Role End?

The Software vs. Model Dilemma:
When Is an AI Assessment Finished?

The Build vs. Buy Dilemma:
What Do You Really Need to Own?

The Data & Control Dilemma:
Trust Beyond the Conversation

Seeing Behaviour at Scale:
What 112 AI Role-Plays Reveal About First-Line Managers

From Assessment to Action:
Learning Needs Analysis at Scale

From Framework to Field:
Stories from the Ground

From Framework to Field:
Stories from the Ground

Designing Competency Framework for a FMCG firm

Competency Profiling & Gap Analysis using ADCs

Designing and Delivery of Sales Excellence Learning Program

How Personality Value
Is Shifting